Game description generation device, game description generation method, and computer program

Abstract

To express good points of a side losing a game when automatically generating an article relating to a match-type game based on game data.SOLUTION: A game description generation device 10 when receiving game data representing a content of a match-type game from a game data distribution system 2, compares scores of one side and of the other side to play based on the received game data and determines whether the one wins or loses the game. When the one loses the game, the game description generation device 10 determines whether or not there is a period when the score of the one is more than the score of the other during the game based on the game data, and when there is the period when the score of the one is more than the score of the other, automatically generates a fight-review article containing such a description that although the one loses the game, there is a period when the one has been leading during the match.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a game description generation device, a game description generation method, and a computer program that automatically generate descriptions of a match in a competitive format, and in particular, to a game description generation device, a game description generation method, and a computer program that automatically generate descriptions of a match from a perspective that focuses on one side, even when one side of the opponent loses or loses during the match. [Background technology]

[0002] There have been several conventional technologies for automatically generating articles about competitive matches. As a specific example, Patent Document 1 below discloses a technology for generating sentences that explain the results and interim status of a match in a way that is easy for people to understand in a sport where points are scored, such as baseball. Furthermore, Patent Document 2 below discloses a technology for acquiring event information, such as the actions of players during a match, from video content related to a sport (for example, tennis), and generating sentences that summarize the content based on the acquired event information. Technologies for generating descriptions of content related to a match (for example, baseball) from such video are also disclosed in Patent Documents 3 and 4 below.

[0003] Furthermore, Patent Document 5 below discloses a technology for automatically creating sports commentary based on result data showing the content of a sports (for example, soccer) game provided in real time. Furthermore, Non-Patent Document 1 below discloses a technology for automatically creating a progress review based on live at-bat reports of high school baseball games. The automatic generation of high school baseball game reviews is also disclosed in Non-Patent Document 2 below. Regarding baseball, Patent Document 6 below discloses a technology for generating live inning reports.

[0004] Furthermore, the following Non-Patent Document 3 discloses a technique for evaluating the batting and other performance of a specific player in professional baseball using index values and outputting highly objective game reviews regarding batting. Similarly, the following Non-Patent Document 4 discloses a technique for scoring each at-bat in a baseball game and creating summaries of high-scoring at-bats. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2001-92815 [Patent Document 2] International Publication No. 2004 / 012100 (Patent Application No. 2004-524084) [Patent Document 3] Japanese Patent Application Laid-Open No. 2000-261754 [Patent Document 4] Japanese Patent Application Laid-Open No. 2001-275058 [Patent Document 5] Japanese Patent Application Laid-Open No. 2005-173651 [Patent Document 6] Japanese Patent Application Laid-Open No. 2017-151864 [Non-Patent Document 1] Kobe Shimbun NEXT, "Progressive Game Review Robot, Using AI, Automatically Creates High School Baseball Game Reviews," [online], [Retrieved March 4, 2021], Internet<URL:https: / / www.kobe-np.co.jp / info / nextguide / koya2018summer / robot / > [Non-patent document 2] Asahi Shimbun Digital, "High School Baseball, Otori's AI Review," [online], [Retrieved March 4, 2021], Internet<URL:https: / / www.asahi.com / sports / baseball / ai-news / > [Non-patent document 3] J-STAGE, "Automatic Generation of Baseball Game Reviews Using WPA," Journal of Information and Knowledge Society, Kentaro Kudo, Kyohei Okawa, Kei Kanazawa, Gen Murai, Vol. 29, No. 2, pp. 181-186, 2019, [online], [Retrieved March 4, 2021], Internet<URL:https: / / www.jstage.jst.go.jp / article / jsik / 29 / 2 / 29_2019_033 / _article / -char / ja / > [Non-patent document 4] J-STAGE, "Proposal and Comparison of Template-Based and Neural-Based Methods for Natural Language Processing and Sports Summary Generation," Hiroki Tagawa, Kazutaka Shimada, Vol. 25, No. 4, 2018, pp. 357-391, [online], [Retrieved March 4, 2021], Internet<URL:https: / / www.jstage.jst.go.jp / article / jnlp / 25 / 4 / 25_357 / _article / -char / ja / > Summary of the Invention [Problem to be solved by the invention]

[0006] As mentioned above, there have been several conventional technologies for generating descriptions of content related to competitive matches, but all of them have focused mainly on the winning side or the winning side of the match, and no mention has been made of technology that makes it possible to generate descriptions of content related to the match from a perspective that focuses on the losing side or the side that is losing during the match.

[0007] For example, the above-mentioned Patent Document 1 includes a description in paragraph 0052 that "there are two ways to set extraction conditions: (A) specifying conditions one by one, and (B) using conditions that are already prepared. Furthermore, there are two types of extraction conditions: (a) a section that specifies the length of the generated sentence and the user's favorite team and players, and (b) a section that specifies the situation to be extracted." It also includes a description in paragraph 0084 that states, "The above sentence generation process provides explanatory text that focuses on interesting parts of the game, favorite plays, and favorite players, in a way that is easy for readers and listeners to understand." These descriptions disclose that if a favorite team or player is specified, explanatory text that focuses on the favorite team or player will be provided, but do not go so far as to say that explanatory text that focuses on the losing side (the side that is losing) will be provided.

[0008] Similarly, the republication publication relating to the above-mentioned Patent Document 2 states from the third line from the bottom of page 6 to the second line of page 7 that "The surface sentence information generation unit 105 generates surface sentence information that expresses content in a sentence summarizing the content based on the summary text elements and user information. Furthermore, when generating the surface sentence information, expressions centered on the target player are used based on the user input information. As a result, the surface sentence information is formed by expressing it from the perspective of the designated player or team." This statement discloses that sentences are formed that are expressed from the perspective of the designated player or team, but does not go so far as to say that sentences are formed that can reliably express noteworthy points during the game for the losing side (the losing side).

[0009] In the above-mentioned Patent Documents 1 and 2, if a team or player is specified, explanatory text focusing on the specified team or player or a sentence expressed from the perspective of the specified team or player is generated, but if such a team or player is not specified, there is a problem in that only sentences from a normal perspective are generated. For this reason, for example, among users who enjoy reading descriptions of competitive matches, there are a certain number of users who emotionally root for the losing side (the losing side), and such users are interested in content related to the losing side (the losing side), but the technologies disclosed in the above-mentioned patent documents and non-patent documents cannot meet the needs of such users.

[0010] The present invention has been made in consideration of the above circumstances, and aims to provide a game description generation device, a game description generation method, and a computer program that, even without a user's specification, focuses on the side that lost the game or the side that is losing during the game, and, if the side that lost the game or the side that is losing during the game was leading at some point during the game, can automatically generate a description that includes content about the side that was leading as a notable feature of the losing side or the losing side. Another object of the present invention is to provide a game description generation device, a game description generation method, and a computer program that can automatically generate a description that includes content about the losing side or the losing side being in the lead for most of the game as a notable feature of the losing side or the losing side, when the losing side or the losing side was in the lead for most of the game.

[0011] Furthermore, the present invention aims to provide a game description generation device, a game description generation method, and a computer program that can automatically generate a description that includes content about the seesaw-like development of the game as a notable feature for the losing side or the losing side, if the losing side of the game or the side that is losing during the game makes several comebacks during the game. Furthermore, the present invention aims to provide a game description generation device, a game description generation method, and a computer program that, when the game to be described includes rules that allow multiple points to be scored in a single play, can automatically generate a description that includes content about the losing side or the losing side once taking a large lead during the game as a notable feature of the losing side or the losing side, if the losing side or the losing side has ever taken a lead by such a large margin that it is impossible for them to make a comeback with a single play. [Means for solving the problem]

[0012] In order to solve the above problems, the present invention provides a game description generation device that acquires game data from the outside indicating the details of a game in which one side competes against another side in a scoring format, and generates a description of the game result based on the acquired game data, and is characterized by comprising: means for making a judgment about the outcome of the game based on the game data acquired from the outside; if the result of said judgment shows that one side lost the game, means for making a judgment based on the game data acquired from the outside, whether there was a period during the game when one side's score exceeded the other side's score; and if the result of said judgment shows that there was a period when one side's score exceeded the other side's score, means for generating a description including content related to one side being ahead of the other side during the game.

[0013] Furthermore, the present invention is characterized by comprising means for determining whether, if the result of said determination is that there was a period during the match when one team's score exceeded the other team's score, that period lasted more than half of the entire match period from the start to the end of the match, and means for generating a description including content relating to one team being ahead of the other team for most of the match, if the result of said determination is that the period during which one team's score exceeded the other team's score exceeded more than half of the entire match period.

[0014] The present invention is a game description generation device that acquires game data from the outside that indicates the details of a game in which one team competes against another team in a scoring format, and generates a description relating to the game in progress based on the acquired game data, and is characterized by comprising: means for determining the magnitude of the scores of one team and the other team at a point in time in the game based on the game data acquired from the outside; means for determining, if the result of said determination shows that the score of one team is lower than the score of the other team, whether there was a period up to said point in time in the game when the score of one team exceeded the score of the other team, based on the game data acquired from the outside; and means for generating a description including content relating to one team being ahead of the other team during the game, if the result of said determination shows that there was a period when the score of one team exceeded the score of the other team.

[0015] Furthermore, the present invention is characterized by comprising means for determining whether, if the result of said determination is that there was a period in which one team's score was higher than the other team's score, that period exceeded half of the period of the game up to said mid-game point, and means for generating a description including content relating to the fact that one team was leading the other team for most of the period up to said mid-game point, if the result of said determination is that the period in which one team's score was higher than the other team's score exceeded half of the period of the game.

[0016] Furthermore, the present invention is characterized by comprising means for determining whether, if the result of said determination is that there was a period in which one team's score exceeded the other team's score, that period was interrupted multiple times, and means for generating a description including content relating to one team making a comeback or winning again during the match, if the result of said determination is that there was a period in which one team's score exceeded the other team's score but was interrupted multiple times.

[0017] Furthermore, the present invention is characterized in that the match includes rules that allow multiple points to be obtained in one play, and if the result of the judgment shows that there was a period when one team's score exceeded the other team's score, the present invention is provided with a means for judging whether the score difference at that time was greater than the maximum score that can be obtained in one play, and if the result of the judgment shows that the score difference is greater than the maximum score that can be obtained in one play, a means for generating a description including content related to the fact that one team was leading the other team by a score difference that could not be reversed in one play during the match.

[0018] Furthermore, the present invention is characterized in that the game description generation device acquires game data from the outside indicating the details of a game in which one team competes against another in a scoring format, and generates a description of the game result based on the acquired game data, and is equipped with means for determining, based on the game data acquired from the outside, whether the case applies to one of the following cases: one team lost the game but had a lead during the game, one team lost the game but had a lead during the game and that lead lasted for more than half of the entire game period from the start to the end of the game, or one team lost the game but had a lead during the game and that lead was interrupted multiple times, and when it is determined that any of the cases applies, it generates a description including details related to the corresponding case.

[0019] Furthermore, the present invention is a match description generation device that acquires match data from the outside, which indicates the details of a match in which one team competes against another team in a scoring format, and generates a description relating to the middle of the match based on the acquired match data, and is characterized by comprising: means for determining, based on the match data acquired from the outside, whether the situation corresponds to one of the following cases: one team is behind the other team at the middle of the match, but one team was leading up to the middle of the match; one team is behind the other team at the middle of the match, but one team was leading up to the middle of the match, and that lead lasted for more than half the time that the match had elapsed up to the middle of the match; or one team is behind the other team at the middle of the match, but one team was leading up to the middle of the match, and that lead was interrupted multiple times; and means for generating a description including details relating to the corresponding case when it is determined that any of the cases applies.

[0020] The present invention is a match description generation method in which a match description generation device externally acquires match data indicating details of a match between one team and another team in a score-based format, and generates a description of the match result based on the acquired match data, characterized in that the match description generation device executes the steps of: determining the outcome of the match based on the match data acquired from the outside; if the result of said determination shows that one team lost the match, determining based on the match data acquired from the outside whether there was a period during the match when one team's score exceeded the other team's score; and if the result of said determination shows that there was a period when one team's score exceeded the other team's score, generating a description including details about one team being ahead of the other team during the match.

[0021] The present invention is a match description generation method in which a match description generation device externally acquires match data indicating details of a match in which one team competes against another team in a score-based format, and generates a description relating to the middle of the match based on the acquired match data, characterized in that the match description generation device executes the steps of: determining, based on the match data acquired from the outside, the magnitude of the scores of one team and the other team at a point in time during the match; if the result of said determination shows that the score of one team is lower than the score of the other team, determining, based on the match data acquired from the outside, whether there was a period up to said point in time when the score of one team exceeded the score of the other team; and, if the result of said determination shows that there was a period when the score of one team exceeded the score of the other team, generating a description including content relating to one team being ahead of the other team during the match.

[0022] Furthermore, the present invention is a match description generation method in which a match description generation device externally acquires match data indicating details of a match between one team and another team in a points-based format, and generates a description of the match result based on the acquired match data, wherein the match description generation device executes the steps of: determining, based on the match data acquired from the outside, whether the case applies to one team losing the match but having a lead during the match; one team losing the match but having a lead during the match that lasted for more than half of the entire match period from the start to the end of the match; or one team losing the match but having a lead during the match that was interrupted multiple times; and, when it is determined that the case applies to any of the cases, generating a description including details relating to the applicable case.

[0023] Furthermore, the present invention is a match description generation method in which a match description generation device externally acquires match data indicating details of a match in which one team competes against another team in a score-based format, and generates a description relating to the middle of the match based on the acquired match data, wherein the match description generation device executes the steps of: determining, based on the match data acquired from the outside, whether the situation corresponds to one of the following cases: one team is behind the other team at a time during the match but one team was leading up to the time during the match; one team is behind the other team at a time during the match but one team was leading up to the time during the match and that lead lasted for more than half the time of the match up to the time during the match; or one team is behind the other team at a time during the match but one team was leading up to the time during the match and that lead was interrupted multiple times; and, when it is determined that the situation corresponds to any of the cases, generating a description including details relating to the corresponding case.

[0024] The present invention is a computer program for causing a computer to externally acquire game data indicating details of a game in which one player competes against another player in a scoring format, and to execute a process of generating a description of the game result based on the acquired game data, characterized in that the computer is caused to execute the following steps: determining the outcome of the game based on the game data acquired from the outside; if the result of said determination shows that one player lost the game, determining based on the game data acquired from the outside whether there was a period during the game when one player's score exceeded the other player's score; and if the result of said determination shows that there was a period when one player's score exceeded the other player's score, generating a description including details about one player being ahead of the other player during the game.

[0025] Furthermore, the present invention provides a computer program for causing a computer to externally acquire game data indicating the details of a match in which one team competes against another in a scoring format, and for generating a description relating to the middle of the match based on the acquired game data, the computer being characterized by having the computer execute the following steps: determining the magnitude of the scores of one team and the other team at a point in time during the match based on the game data acquired from the outside; if the result of said determination shows that the score of one team is lower than the score of the other team, determining based on the game data acquired from the outside whether there was a period up to the point in time during the match when the score of one team exceeded the score of the other team; and if the result of said determination shows that there was a period when the score of one team exceeded the score of the other team, generating a description including content relating to one team being ahead of the other team during the match.

[0026] Furthermore, the present invention is a computer program for causing a computer to externally acquire game data indicating the details of a match in which one team competes against another in a scoring format, and to execute a process of generating a description of the match result based on the acquired game data, characterized in that the computer is caused to execute the following steps: determine, based on the game data acquired from externally, whether the case applies to one team losing the match but having one team leading during the match, one team losing the match but having one team leading during the match and that leading period lasting more than half of the entire match period from the start to the end, or one team losing the match but having one team leading during the match and that leading period being interrupted multiple times; and, when it is determined that the case applies to one of the cases, generate a description including the details of the applicable case.

[0027] Furthermore, the present invention provides a computer program for causing a computer to externally acquire game data indicating the details of a game in which one team competes against another team in a scoring format, and for generating a description relating to the middle of the game based on the acquired game data, the computer being characterized by the steps of: determining, based on the game data acquired from externally, whether the situation corresponds to one of the following cases: one team is behind the other team at a time during the game, but one team was leading up to the time during the game; one team is behind the other team at a time during the game, but one team was leading up to the time during the game, and that lead lasted for more than half the time of the game up to the time during the game; or one team is behind the other team at a time during the game, but one team was leading up to the time during the game, and that lead was interrupted multiple times; and generating a description including the details relating to the corresponding case when it is determined that the situation corresponds to one of the cases.

[0028] In this invention, the loser of the match is determined based on match data obtained from an external source, and whether there was a time during the match when the score of the losing side exceeded that of the other side is determined, and if there was such a time, a description including content relating to the time when one side exceeded the other side's score is automatically generated, thereby making it possible to provide content informing the user that the losing side was leading during the match. Note that the score-based matches of this invention include matches between individuals or matches between multiple people (for example, matches between pairs, matches between teams, etc.).

[0029] Furthermore, in the present invention, if there is a period during the game when the losing team's score exceeds that of the other team, it is determined whether such a period lasted more than half of the entire game period. If it is determined that such a period has lasted more than half of the entire game period, a description including content relating to the period during which one team's score exceeded the other team's score for more than half of the entire game period is automatically generated, thereby making it possible to provide content informing the user that the losing team was leading for most of the game. Note that, in the present invention, the term "game period" refers to the length of the game when the progress and end of the game are determined by time; however, in the case of baseball, when the progress and end of the game are determined by concepts other than time, such as innings in baseball, or games or sets in tennis, such concepts (innings, games, sets, etc.) refer to the period of the game. Therefore, for example, in the case of baseball, if a game ends after the ninth inning, if one team has been leading for more than five innings, this would be considered to have exceeded half of the entire game period in the present invention.

[0030] In this invention, based on match data obtained from an external source, one of the two players is determined to be losing at some point in the match, and whether there was a period from the start of the match until that point in time when the score of that player exceeded that of the other player is determined. If such a period exists, a description including content relating to the period when one player's score exceeded the other player's score is automatically generated, so that the player who is losing at some point in the match can be provided with content informing the user that they had been leading up to that point. Match-related descriptions are provided at any time even during the match, and this invention is suitable for providing such match descriptions that focus on the player who is losing at some point in the match.

[0031] Furthermore, in the present invention, if there is a period during the match in which the score of one losing team has exceeded the score of the other team up to that point, a determination is made as to whether such a period lasted for more than half the duration of the match, and if it has, a description is automatically generated including content relating to the period in which one team's score exceeded the other team's score for more than half the duration of the match, thereby making it possible to provide content informing the user that the team losing during the match had been leading for most of the match up to that point.

[0032] In the present invention, if there is a period during the match when the score of one team that has lost or is losing in the middle of the match exceeds the score of the other team, it is determined whether there were multiple uninterrupted periods when the score of one team exceeded the score of the other team, and if there were multiple uninterrupted periods when the score of one team exceeded the score of the other team, a description is automatically generated including content relating to the multiple uninterrupted periods when the score of one team exceeded the score of the other team, thereby making it possible to provide content informing the user that one team has made a comeback or won several times during the match.

[0033] In the present invention, when the match in question includes a rule that allows multiple points to be obtained in one play, if there is a time during the match when the score of one side that has lost the match or is losing during the match exceeds the score of the other side, a determination is made as to whether the score difference at such a time exceeds the maximum score that can be obtained in one play, and if the score difference exceeds the maximum score that can be obtained in one play, a description is automatically generated that includes content relating to the fact that the score difference when one side was leading during the match exceeds the maximum score that can be obtained in one play, thereby making it possible to provide content that informs the user that one side was leading the other side by such a score difference that it was impossible to reverse the situation in one play.

[0034] In the present invention, it is possible to determine in parallel which of a number of situations applies when one team lost to the other but was in the lead during the match, and to automatically generate a description including content relating to which of the cases is found to apply, so that a description of the match results that focuses on the losing side can be smoothly provided, even if a user who supports the losing side does not have to set the losing side in advance. Similarly, in the present invention, when one side is losing to the other side during the match, it is determined in parallel which of multiple situations in which one side was leading during the match applies, and a description including content related to which of the cases that is found to apply is automatically generated. Therefore, even when providing a description of the match during the match, it is possible to smoothly provide a description of the match during the match that focuses on the losing side, and this makes it easier for users who are rooting for the losing side to understand the progress of the situation when their side is losing, without having to make any special settings for the side they are rooting for in advance. [Effects of the Invention]

[0035] In the present invention, based on match data obtained from an external source, a description is automatically generated that includes information about the period during the match when the score of the losing side exceeded the score of the winning side, so that a description that conveys the good points of the losing side during the match can be provided without having to perform any operation in advance such as specifying one side. Furthermore, the present invention automatically generates a description that includes content relating to the fact that the losing side's score exceeded the winning side's score for more than half of the entire match period, making it possible to provide a description that includes content relating to the fact that although one side lost the match, they were in the lead for most of the match, and conveying the nuance that although one side lost the match, they were in the lead for many periods throughout the match.

[0036] In this invention, based on match data obtained from an external source, a description is automatically generated that includes content relating to the period during the match when the score of one side that is losing during the match exceeded the score of the opposing side (the other side).Therefore, even without performing any prior operation such as specifying one side, a description can be provided that conveys the good points of the side that is losing at the time during the match. Furthermore, the present invention automatically generates a description that includes content relating to the fact that the score of one side that is losing the match exceeded the score of the opposing side (the other side) for more than half of the period that has elapsed since the start of the match until the point in time of the match. This makes it possible to provide a description that includes content relating to the fact that although the side is currently losing, it was leading for most of the time before that, and can convey the nuance that the side that is losing the match during the course of the match had often been in the lead up to that point.

[0037] In the present invention, a description is automatically generated that includes content relating to the fact that there were multiple, uninterrupted periods during the match in which the score of one side that lost the match or was losing during the match exceeded the score of the opposing side (the other side). This makes it possible to provide a description that includes content relating to the fact that the side that lost the match or was losing during the match made a comeback or won again during the match, and it is possible to convey the nuance that the match was like a seesaw game, with several occasions in which one side was in the lead, even though the other side had lost the match or was losing the match.

[0038] Furthermore, the present invention automatically generates a description including content relating to a side that has lost the match or is losing during the match taking an irreversible lead in one play during the match, so it is possible to provide a description including content relating to a side that has lost the match or is losing during the match building up a lead that could be considered safe at one point during the match, and it is possible to convey the nuance that at one point the side that has lost the match or is losing during the match was in the lead and advancing the match with an advantage.

[0039] Furthermore, in the present invention, a concurrent determination is made at the same stage as to whether the losing side of the match falls into one of several situations related to when the losing side was in the lead during the match, and a description including the relevant content is automatically generated if it is determined that this applies, so that a description of the match content can be smoothly generated according to the needs of users who are rooting for the losing side, etc. Furthermore, in the present invention, in the same manner as described above, for a side that is losing during the match, a parallel determination is made at the same stage as to whether it falls into one of several situations that relate to when the side is in the lead during the match, and if it is determined that it falls into that situation, a description including the relevant content is automatically generated. Therefore, even when providing a description during the match, a description that meets the needs of users who are supporting the losing side can be smoothly provided. [Brief explanation of the drawings]

[0040] [Figure 1] 1 is a schematic diagram showing the overall configuration of a game description generation system including a game description generation device according to an embodiment of the present invention. [Figure 2] 1 is a block diagram showing the main internal configuration of a match description generation device. [Figure 3] 10 is a diagram showing an example of the contents of a judgment table. [Figure 4] 10 is a diagram showing an example of the contents of a condition content list. [Figure 5] 10 is a diagram showing an example of the contents of a pattern content list. [Figure 6] 10 is a diagram showing an example of the contents of a point table. [Figure 7] 1 is a flowchart showing a series of processing steps relating to a game description generating method according to a first embodiment (Example 1). [Figure 8] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 9] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 10] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 11] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 12] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 13] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 14] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 15] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 16] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 17] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 18] 10A is a chart showing scores for each inning, showing an example of the results of a baseball game, and FIG. 10B is an example of a game review article of the first embodiment, which was automatically generated for the example of the results of the game in FIG. [Figure 19] 10 is a flowchart showing the first half of a series of processing steps relating to a game description generating method according to a modified example of the first embodiment. [Figure 20] 10 is a flowchart showing the second half of a series of processing steps relating to a game description generating method according to a modified example of the first embodiment. [Figure 21] 10 is a table showing an example of the contents of a condition content list according to the second embodiment (Example 2). [Figure 22] 10 is a table showing an example of the contents of a pattern content list according to the second embodiment. [Figure 23] 10 is a flowchart showing a series of processing steps relating to a game description generating method according to a second embodiment (Example 2). [Figure 24] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 25] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 26] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 27] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 28] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 29] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 30] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 31] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 32]10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 33] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 34] 10A is a score chart showing an example of the score situation up to a midpoint in a baseball game, and FIG. 10B is an example of a progress report of Example 2 that was automatically generated midpoint in FIG. [Figure 35] 10 is a flowchart showing the first half of a series of processing steps relating to a game description generating method according to a modified example of the second embodiment. [Figure 36] 10 is a flowchart showing the second half of a series of processing steps relating to a game description generating method according to a modified example of the second embodiment. DETAILED DESCRIPTION OF THE INVENTION [Example]

[0041] 1 is a schematic diagram showing the overall configuration of a game description generation system 1 including a game description generation device 10 according to a first embodiment (Example 1) of the present invention. The game description generation system 1 according to the present invention is configured such that a game data distribution system 2 that distributes (transmits) game data and a game description generation device 10 that automatically generates descriptions related to game results are communicably connected via a network. Furthermore, the description data generated by the game description generation device 10 can be provided to user terminals T1, T2, T3, T4, etc. via the communication network.

[0042] In this first embodiment (Example 1), the target for generation is a description of a baseball game (e.g., professional baseball, high school baseball, etc.) in which one team competes against another in a scoring format.When the baseball game ends, the game data distribution system 2 transmits game data related to the game to the game description generation device 10, and the game description generation device 10 automatically generates a game review article including a description of the result of the game based on the received game data.

[0043] The game data distribution system 2 includes a server device 2a and a DB (database) device 2b, and performs processing to distribute game data related to baseball games (for example, professional baseball in Japan or overseas, high school baseball in Japan, etc.). Note that the game data distribution system 2 is an external system (device) to the game description creation device 10.

[0044] The game data distributed by the game data distribution system 2 includes electronic scorebook data that electronically represents the contents written in a baseball scorebook, as well as data on the play of the game (such as the movements of each player and the trajectory of the ball's movement) acquired by a tracking system, etc., and data on records, etc. The electronic scorebook data included in the game data contains information indicating the results of each pitch, the score status of each inning of baseball, the strike and ball status of the pitcher's pitch, the batting results of the batter against the pitch (foul, hit, etc.), the outcome of the game, etc.

[0045] The game data includes data related to game plays, such as pitching data and batting data. The pitching data includes data indicating the number of spins of the pitched ball, the magnitude of its trajectory, its position relative to the strike zone, and the location where the ball was released (release point). The batting data includes data indicating the speed, angle, and distance of the batted ball. The server device 2a reads data related to records that may occur in the game (e.g., in the case of professional baseball, achieving 2,000 hits or the 100th win) from the past electronic scorebook data stored in the DB device 2b, and data related to the continuation of records (e.g., consecutive games played or consecutive on-base runs), from the DB device 2b, and includes the data in the game data to be distributed. Since the game data distributed in this manner is generally well known (see, for example, the aforementioned Patent Document 5 and Non-Patent Document 4), further details will be omitted.

[0046] 2 is a block diagram showing the main internal configuration of a match description generator 10, which is the central device of the present invention. The match description generator 10 of Example 1 is constructed using a general server computer (corresponding to the computer according to the present invention), but it is of course possible to construct the match description generator 10 by combining multiple server devices by performing distributed processing or the like. The match description generator 10 automatically generates descriptions for match review articles based on match data distributed from the above-mentioned match data distribution system 2 using a unique program (AI: artificial intelligence).

[0047] The match record generation device 10 is composed of an MPU 10a that performs overall control and various processing, and various devices etc. that are connected to it via internal connection lines 10h. The various devices etc. include a communication module 10b, RAM 10c, ROM 10d, input interface 10e, output interface 10f, and memory unit 10g.

[0048] The communication module 10b is a communication device equivalent to a network connection module and conforms to a required communication standard (e.g., a LAN module). The communication module 10b is connected to a network (including internal, external, or other communication networks) via required communication equipment (not shown, such as a router), enabling communication with an external game data distribution system 2 (the game data distribution system 2 corresponds to an external system for the game description generation device 10) and user terminals T1, T2, T3, etc. After a baseball game ends, the game data distribution system 2 distributes game data related to that game. The game data is received by the communication module 10b and acquired by the game description generation device 10. The acquired game data is stored in RAM 10c or memory unit 10g, etc.

[0049] The RAM 10c temporarily stores the contents and files associated with the processing of the MPU 10a, and the ROM 10d stores in advance programs and the like that define the basic processing contents of the MPU 10a. The input interface 10e is connected to a keyboard 11, a mouse, etc. that accept operation instructions from an administrator or the like of the game record generator 10. The output interface 10f is connected to a display 12 (display output device), and outputs the contents associated with the processing of the MPU 10a to the display 12 so that an administrator or the like of the game record generator 10 can check the current processing contents, etc.

[0050] The storage unit 10g stores programs, tables, databases, etc., and stores the OS program P1 and description generation program P2 as programs, the judgment table 13, point table 16, template table 17, etc. as tables, and stores description DB 18 (DB that stores data of generated battle review articles) as a database. Note that, in order to install the description generation program P2 into the storage unit 10g, it is conceivable to store the description generation program P2 on a storage medium such as an optical disk and install it into the storage unit 10g via that storage medium.

[0051] Among the programs stored in the memory unit 10g, the OS program P1 specifies various processes according to the operating system for the server computer, and the MPU 10a performs processes based on these specifications, allowing the match description generation device 10 to perform each function as a server computer (server device).

[0052] 3 shows an overview (one example) of the judgment table 13 stored in the storage unit 10g. The judgment table 13 is a table used by the MPU 10a to determine what content to include in a review of a baseball game (such as professional baseball) based on the specifications of the review generation program P2, as will be described later. In particular, the present invention is characterized in that the review is generated with a focus on the losing team. In order to enable the losing team to write about notable events during the game or the success of players on the losing team, the judgment table 13 includes conditions (corresponding to cases) for determining these cases and patterns of content to be written when each condition is met.

[0053] In the judgment table 13, judgment conditions 1 to 7 are arranged in the leftmost vertical column 13a, and patterns A to K indicating the type of content to be entered are arranged in the bottom row 13b. In addition, in the grid section 13c surrounded on both sides by the leftmost vertical column 13a and the bottom row 13b, a circle (○) is arranged in the grid location where the condition applies. In the grid section 13c, when a condition does not apply or when it does not affect the judgment regardless of whether it applies or not (for example, the grid location in the row of condition 7 excluding pattern J), that grid location is left blank.

[0054] 4 shows a condition content list 14 that lists the specific contents of each of the cases of conditions 1 to 7 arranged in the leftmost vertical column 13a in the above-mentioned judgment table 13. In judgment table 13, one of the two teams competing in a baseball (for example, professional baseball) game is designated as one team, and the other team is designated as the opposing team, and then the contents of each condition (each case) are specified with a focus on that one team.

[0055] Specifically, condition 1 specifies "if one team wins the match," and condition 2 specifies "if one team loses the match and has zero points or no lead during the match." Below, condition 3 specifies "if one team loses a game, but there is a period during the game when one team's score exceeds that of the other team", condition 4 specifies "if one team loses a game, but there is a period during the game when one team's score exceeds that of the other team, and the difference in score during that period is more than four points", condition 5 specifies "if one team loses a game, but there is a period during the game when one team's score exceeds that of the other team, and that period lasts more than half the number of games", condition 6 specifies "if one team loses a game, but there is a period during the game when one team's score exceeds that of the other team, and there are multiple periods where that period is interrupted", and condition 7 specifies "if there is a player (batter or pitcher) whose points calculated based on the point table exceed the threshold value".

[0056] Condition 1, "If one team wins the match," corresponds to the generation of content equivalent to conventional general match review articles, and is a condition appropriate for generating match review articles centered on the winning team. Condition 2, "If one team loses the match and scores zero or does not even have a lead during the match," corresponds to the case where one team loses the match and the losing team has nothing noteworthy in terms of scores (if they have no good points).

[0057] Condition 3, "If one team loses the match, but there is a period during the match when the score of that team exceeds that of the other team," corresponds to the condition that the losing team has a period during the match when the score of that team exceeds that of the other team (when they are in the lead), which is a highlight of the match.

[0058] Condition 4, "If one team loses the game, but there is a period during the game when the score of one team is greater than the score of the other team, and the difference in score during that period is greater than four points," corresponds to a situation in which, as a highlight of the game for the losing team, one team was leading the other team for a period of time by a difference in score that is greater than the maximum number of points that can be scored in a single play in baseball (four points from a grand slam home run).

[0059] Condition 5, "When one team loses the game, but there is a period during the game when one team's score exceeds the other team's score, and this period lasts for more than half the number of innings of the game," corresponds to the case when, as a highlight of the losing team, if a baseball game ends in nine innings (corresponding to the entire period of the game from start to finish), one team has been in the lead for more than five innings (one team has been leading the other team for most of the game).

[0060] If a baseball game goes overtime, all innings, including the extra innings, count as the entire game period; for example, if the game goes overtime to 13 innings, condition 5 is met if one team is leading by more than seven innings. Also, if a baseball game does not go to the ninth inning due to a rainout or other reason (the game ends in fewer than nine innings), the entire game period is from the start of the game to the innings that have progressed; for example, if the game ends in the bottom of the seventh inning due to a rainout, condition 5 is met if one team is leading by more than four innings.

[0061] Condition 6, "One team loses the match, but there is a period during the match when one team's score is greater than the other team's, and there are multiple periods where this occurs," corresponds to a highlight of the match for the losing team, where the team makes multiple comebacks during the match (they make multiple comebacks from a losing position), or where the team makes multiple winning runs (they are tied or overtaken from a winning position, but then return to a winning position at least once).

[0062] Condition 7, "When there is a player (batter or pitcher) whose points exceed the threshold value," is a condition for highlighting a player (batter or pitcher) who has performed above a certain level in a game and introducing the details of his / her performance. The determination of whether a player has performed above a certain level is based on the cumulative points awarded based on the specified details (calculation formula for adding points) stored in point table 16 (details will be described later).

[0063] 5 shows a pattern content list 15 listing the specific contents of patterns A to K arranged in the bottom row 13b of the judgment table 13. The contents of these patterns A to K define guidelines for the content to be included in the generated game review. Specifically, pattern A defines the guideline of "content related to one team winning the game," and since pattern A is a guideline for including content related to winning the game, it basically defines content equivalent to that of a conventional general game review.

[0064] Pattern B specifies the guideline "One team lost the game, but the content related to one team being in the lead during the game." This pattern B is a guideline for generating descriptions that include content related to the fact that one team lost the game, but that one team was in the lead over the other team during the game, as a good point of the losing team, and the content is written from the perspective of the losing team, specifically listing noteworthy aspects of the losing team's game.

[0065] Pattern C specifies the situation in which one team was leading in more detail than Pattern B above, and prescribes the guideline, "One team lost the game, but at one point during the game they were leading by a score difference that was not easily recaptured." This Pattern C also serves as a guideline for generating descriptions that include content related to the fact that, although they lost the game, the losing team had a good point, leading the other team by a score difference that was so large that the opposing team could not come back, even if they had allowed a grand slam home run (a score difference greater than four points, the maximum score that can be obtained in one play in a baseball game). From the perspective of focusing on the losing team, the content specifically explains the losing team's lead during the game in terms of the score difference.

[0066] Pattern D also specifies the lead situation in more detail than Pattern B, and prescribes the guideline that "one team lost the game, but was ahead of the other team for most of the game." This Pattern D also serves as a guideline for including in the game review content that, although the losing team lost the game, they were ahead of the other team for more than half of the nine innings that make up a baseball game (more than five innings) (for most of the game), and the content focuses on the losing team and explains the losing team's lead during the game based on the length of their lead.

[0067] Pattern E also specifies the situation of the lead in more detail than Pattern B, and prescribes the guideline "content related to one team losing the game, but having made multiple comebacks or winning streaks during the game." This pattern E also serves as a guideline for generating descriptions that include content related to the fact that, although the losing team lost the game, they made multiple comebacks during the game from behind (having made a re-competition), or that they were leading but were tied or overtaken, but had at least one re-winning streak (having a re-win), and it specifically describes the situation in which the losing team was fighting a see-saw game from the perspective of focusing on the losing side of the game.

[0068] Pattern F specifies a situation that combines the above-mentioned Patterns C and D, and prescribes the guideline that "one team lost the game, but led by a score that was difficult to catch up with, and that this lead lasted for most of the game." Pattern G specifies a situation that combines the above-mentioned Patterns C and E, and prescribes the guideline that "one team lost the game, but led by a score that was difficult to catch up with, and that there were multiple comebacks or wins during the game."

[0069] Furthermore, Pattern H specifies a situation that combines the above-mentioned Patterns D and E, and prescribes the guideline that "one team lost the game, but was leading for most of the game, and had multiple comebacks or wins during the game." Furthermore, Pattern I specifies a situation that combines the above-mentioned three patterns, Pattern C, Pattern D, and Pattern E, and prescribes the guideline that "one team lost the game, but was leading by a score that was difficult to catch up with, and that lead lasted for most of the game, and had multiple comebacks or wins during the game."

[0070] Pattern J specifies the guideline "Content related to one team losing the game, but having a player (batter or pitcher) who performed well on that team during the game." This pattern J is a guideline for generating descriptions that include the fact that, even though the losing team lost the game, there was a player (batter or pitcher) on that losing team who performed better than the other players (batters or pitchers) and describes their performance from the perspective of the losing team.

[0071] Pattern K specifies the guideline that "one team loses the game and the content is related to praising the opposing team (the other team)." This pattern K is a guideline for including in the game review article the fact that the opposing team (the other team) was excellent, because when a team loses a game and there is nothing particularly noteworthy about that team in the game, or when none of the players on that team performed well, it is difficult to mention the losing team even from the perspective of the losing team.

[0072] 6 shows an overview (one example) of the points table 16 stored in the storage unit 10g. The points table 16 is a table used to determine condition 7 (when there is a player (batter or pitcher) whose points exceed the threshold) in the judgment table 13 described above, and includes a batter table section 15a and a pitcher table section 15b, and defines the value of points to be added depending on the content of the play (play details) of the batter or pitcher.

[0073] Examples of point conditions included in the batter's table section 15a shown in Figure 6 include "100 points added (plus 100 points) if the winning run is hit (the player who batted to score the winning run)", "50 points added (plus 50 points) if the game is won (the player who batted to score the winning run)", "30 points added (plus 30 points) if the game is tied (the player who batted to score the tying run)", "50 points added (plus 50 points) if the game is won (the player who batted to score the winning run)", "30 points added (plus 30 points) if the game is hit to take the lead (the player who batted to take the lead)", "20 points added (plus 20 points) if the game is hit to triple", "10 points added (plus 10 points) if the game is hit to double", "5 points added (plus 5 points)" and so on.

[0074] Examples of additional points included in the pitcher's table section 15b shown in Figure 6 include "adding points equal to the number of innings in which the batter was kept scoreless multiplied by 20" and "adding points equal to the number of strikeouts multiplied by 5."

[0075] The template table 17 stored in the storage unit 10g stores templates corresponding to various expressions used when generating game reviews, and includes a variety of expressions related to standard content according to various parts of speech such as nouns, particles, conjunctions, adverbs, and verbs used in expressing baseball game reviews, as well as units such as phrases and clauses. The templates corresponding to these various expressions have been appropriately extracted from expressions frequently used in many baseball game reviews created by humans in the past.

[0076] Examples of templates for nouns include "game," "team," "score," "lead," "batter," "pitcher," "substitution," "pinch hitter," "home run," "double," "hit to left field," "grounder to first base," "fly to center field," "strikeout," "fly to third base," "steal," and "timely hit." Examples of templates for particles include "is," "but," "from," "of," "even," "with," and "while."

[0077] Examples of conjunction templates include "here," "and," and "also." Examples of adverb templates include "very," "intense," "highly," and "eye-opening." Examples of verb templates include "won," "lost," "turned the tables," "gained the lead," "failed to defend," "supported," and "held back." These various templates are stored in an organized state in the template table 17 in association with the conditions under which they should be used, and are read out by the MPU 10a as appropriate and used to generate descriptions according to the specifications of the description generation program P2.

[0078] The description generation program P2 stored in the storage unit 10g corresponds to a computer program according to the present invention and defines various processes for the MPU 10a to generate descriptions for baseball game reviews based on game data distributed from the game data distribution system 2. The MPU 10a functions as various means in accordance with the definitions of the description generation program P2. The description generation program P2 of this embodiment defines various processes for the MPU 10a to automatically generate articles as an artificial intelligence and includes a text creation engine corresponding to the various processes for automatic article generation by the artificial intelligence. This text creation engine also allows the MPU 10a to function as an AI (artificial intelligence) for writing sentences and automatically generates descriptions based on identified patterns using a template table 17.

[0079] Furthermore, this description generation program P2 stipulates that the MPU10a generates multiple types (at least three types in this embodiment) of descriptions for game review articles for one baseball game.The first is to generate a description for a game review article that focuses on the winning team, as in the past (when the description is generated assuming that the winning team corresponds to one of the teams), the second is to generate a description for a game review article that includes content that focuses on the losing team, which is a feature of the present invention (when the description is generated assuming that the losing team corresponds to one of the teams), and the third is to generate a description for a game review article that includes content that focuses on players (batters or pitchers) who performed particularly well in the game or achieved records, etc. (when a player has earned points exceeding a threshold, a description is generated that includes content related to the performance of that player in accordance with the points earned, regardless of the win or loss of the team to which that player belongs).

[0080] The processing defined by such a description generation program P2 can be divided into two main stages in the processing flow. The first stage is when the MPU 10a uses the above-mentioned judgment table 13 (see Figure 3) to analyze the received match data and identify the content (pattern) of the description to be generated. The second stage is when the MPU 10a uses the text creation engine to actually generate a match review article including a description corresponding to the content identified in the first stage, using various information contained in the received match data and the template table 17.

[0081] The processing contents of the description generation program P2 will be explained in more detail based on a specific example. Assume that the game data distribution system 2 distributes game data about a baseball game between team XXX (batting first) and team YYY (batting second). When the game data is received (acquired by reception) by the game description generation device 10, the MPU 10a of the game description generation device 10 starts the first stage of processing described above, triggered by the reception of the game data. When the processing starts, in order to generate a game review article including a description focusing on team XXX (batting first), which is one of the opposing teams included in the game data, the MPU 10a first provisionally applies team XXX to the judgment table 13 as one of the teams, and determines which of conditions 1 to 7 defined in the judgment table 13 applies to team XXX. Note that, in this state where the MPU 10a assumes team XXX to be one of the teams, it will recognize team YYY, the opposing team, as the other team.

[0082] In a certain game, when the game data indicates that team XXX's score (game result score) exceeds team YYY's score (game result score) and team XXX has won, the MPU 10a determines that team XXX falls under condition 1 in this game (determines that team XXX has won). Also, when the game data indicates that team XXX's score is lower than team YYY's score and team XXX has lost, the MPU 10a determines that team XXX has lost, and in this case, the MPU 10a analyzes the acquired game data and compares the score of team XXX with the score of team YYY for each time (each inning) in this game, and determines which of conditions 2 to 6 team XXX falls under.

[0083] The processing results for determining (detecting) whether each condition is met include cases where only one condition is met, as well as cases where multiple conditions are met, as shown in the vertical columns of the judgment table 13 in Figure 3.

[0084] Specifically, condition 3, which relates to when one team loses a match, is "when one team loses the match, but there is a period during the match when one team's score is greater than the other team's," and therefore condition 3 is included in condition 4 (when one team loses the match, but there is a period during the match when one team's score is greater than the other team's, and the difference in score during that period is greater than four points), condition 5 (when one team loses the match, but there is a period during the match when one team's score is greater than the other team's, and that period lasts for more than half the number of matches), and condition 6 (when one team loses the match, but there is a period during the match when one team's score is greater than the other team's, and there are multiple periods where that period is interrupted).Therefore, there will be cases where multiple conditions are met, such as when conditions 3 and 4 are met, when conditions 3 and 5 are met, when conditions 3 and 6 are met, or when conditions 3 to 5 are met.

[0085] Therefore, in the process of determining whether any of the conditions applies, if it is determined that condition 1 (one team wins the match) does not apply, that is, if it is determined that one team loses the match, the MPU 10a determines which of conditions 2 to 6 relating to a loss applies. Furthermore, as described above, condition 3 is included in conditions 4 to 6, so if it is determined that condition 3 applies, the MPU 10a also determines whether any of conditions 4 to 6 applies.

[0086] Next, to generate a game review article that includes a description focusing on team YYY, the opposing team of team XXX, the MPU 10a reverses the above and sets team YYY as one team and team XXX as the other team, and determines which of the conditions defined in the judgment table 13 applies to the team. Here, if the MPU 10a determines that team XXX won when it is provisionally assumed that team XXX, who batted first, was one of the teams as described above, it can be automatically determined that team YYY lost the game, and in this case the MPU 10a simultaneously determines (detects) which of conditions 2 to 7 related to losses in the judgment table 13 applies to team YYY as one of the teams.

[0087] Furthermore, as described above, if the MPU 10a determines that team XXX, who bats first, has lost when the MPU 10a has provisionally assumed that team XXX is one of the teams, then it can be automatically determined that team YYY has won, and in this case the MPU 10a determines that team YYY, as one of the teams, meets condition 1 in the judgment table 13. Note that in the above explanation, the MPU 10a provisionally assumes that team XXX, who bats first, is one of the teams and applies the judgment table 13 to determine which of the conditions the team meets, but it is of course also possible to proceed with the process by assuming that team YYY, who bats second, is one of the teams first (in short, it is sufficient to assume that one of the opponents is one of the teams and apply the judgment table 13 to determine and identify the applicable condition).

[0088] Furthermore, in order to generate a description of the players who performed well in the game, the MPU 10a performs the condition specification process of the judgment table 13 based on the win / loss results of the opposing teams described above, and also performs a process of calculating the points of each player (batter) and each pitcher included in the game data using the points table 16 shown in FIG. 6, and detecting whether or not any player (batter or pitcher) has calculated points that exceed a threshold. In this embodiment, two thresholds are prepared: a threshold for the winning team and a threshold for the losing team. The threshold for the winning team is 80 points, and the threshold for the losing team is 50 points, which is lower than the threshold for the winning team. Players (batters or pitchers) who have earned points equal to or greater than these thresholds are identified. Note that these threshold point numbers are merely examples, and other point numbers can of course be applied to the thresholds. For example, depending on whether you want to narrow down the players to be extracted or extract a wide range of players, it is of course possible to change the threshold for the winning side to be greater or less than 80 points, and similarly, the threshold for the losing side can be greater or less than 50 points, but to make it easier to extract players on the losing side, it is preferable to make the number of points in the threshold for the losing side smaller than the number of points in the threshold for the winning side. Also, if multiple players exceed the threshold, a narrowing process is performed so that the top two players or the player with the highest points are detected.

[0089] The game description generation device 10 processes articles including descriptions about players whose points exceed the threshold in two ways: when the article is generated in association with the team's win or loss and condition 7 is met, and when the article is generated including descriptions about the player who meets condition 7 regardless of the team's win or loss.

[0090] Furthermore, the MPU 10a also detects from the received game data whether or not each player has achieved any kind of record. The types of records to be detected include records achieved in a single game related to the received game data (e.g., a cycle hit or a no-hitter) and records achieved in a series of games including the game related to the received game data (e.g., reaching base in 20 consecutive at-bats, achieving 1,000 hits, allowing no runs in 20 consecutive innings, achieving 1,000 strikeouts, etc.). The game description generation device 10 also processes information related to such record achievements in the article it generates, regardless of whether the team won or lost.

[0091] After the above-described process has been completed to determine (detect) the conditions (conditions 1 to 7) that apply to both opposing teams and to determine (detect) whether or not there are any outstanding players (batters or pitchers) or players (batters or pitchers) who have achieved records, the MPU 10a then performs a process to identify the content (pattern) of the conditions that apply. This process also uses the decision table 13 shown in Figure 3, and the MPU 10a, based on the provisions of the entry generation program P2, identifies a pattern (any of patterns A to K) that applies to the conditions of the first-batting team XXX, and also identifies a pattern (any of patterns A to K) that applies to the conditions of the second-batting team YYY, the opposing team.

[0092] In the above-mentioned process, the MPU 10a determines that a player (batter or pitcher) exceeding the threshold satisfies condition 7, and if condition 7 is met and condition 2 is also met, it identifies pattern J as the content of the description based on the provisions of the text creation engine included in the description generation program P2 (see the judgment table in Figure 3).

[0093] After the first stage of processing for identifying the content of the description is completed, the MPU 10a then performs a second stage of processing to automatically generate a match review article relating to the results of the match, including the identified content of the description (patterns A to K), based on the specifications of the text creation engine included in the description generation program P2. The text creation engine of the description generation program P2 defines various processes for the MPU 10a to automatically generate articles as an AI (artificial intelligence) for text creation, and this text creation engine automatically generates a description based on the identified pattern using the template table 17.

[0094] Specifically, when generating a description that corresponds to a specified pattern for team XXX, team XXX is used as the subject item, and numerical and lexical information that corresponds to the specified pattern (such as the number of runs and the number of innings in which runs were scored) and the scoring situation (such as a home run by player XX or a timely hit by player X△) are analyzed and extracted from the received game data by the text creation engine, and the extracted items (information) are combined as subjects, objects, verbs, phrases, clauses, etc. with various templates included in template table 17 as appropriate to create expressions, thereby generating a description with content that matches the specified pattern.

[0095] Note that the game review article generated in this embodiment is not composed only of descriptions according to the specified pattern, but is configured to partially include descriptions according to the specified pattern. Therefore, assuming that Team XXX is one of the teams, the MPU 10a generates a game review article that includes descriptions focusing on Team XXX, which is the other team, but also includes overall content other than that and the scoring situation of the opposing team. The game review article generated in this way not only conveys to the user (reader) the good points of Team XXX in the game, but also contains content that allows the user (reader) to understand the overall game and flow.

[0096] As for team YYY, the opponent team of team XXX, in the same manner as in the case of team XXX described above, the MPU 10a analyzes and extracts information such as numerical values, phrases, and score status corresponding to the pattern specified by team YYY from the received match data in accordance with the provisions of the text creation engine, and automatically generates an article description for team YYY using the extracted information and each template included in the template table 17. The MPU 10a also automatically generates descriptions related to the overall content other than team YYY in parallel, thereby automatically generating a match review article including descriptions focusing on team YYY.

[0097] Furthermore, in order to generate descriptions of content according to pattern J for players who have achieved points above the threshold or who have set records, the MPU 10a analyzes and extracts information about the target players (player names, play details corresponding to points above the threshold, record details achieved, etc.) from the received match data in accordance with the provisions of the text creation engine of the description generation program P2, and automatically generates descriptions focusing on players who performed well in the match or players who have set records using the extracted information and each template included in the template table 17. Note that in this case as well, the MPU 10a automatically generates descriptions related to overall content other than the focused player in parallel, and as a result, a match review article including a description about the focused player is automatically generated by the MPU 10a.

[0098] The order of each description that makes up the content of the automatically generated game review article is in chronological order, just like the content of a baseball game, and the MPU 10a generates the game review article by combining each description automatically created based on the received game data in the order in which the game progressed in the game data. When the MPU 10a generates a game review article, it issues an identification ID indicating the game date and time, opposing team, etc., and associates the issued identification code with the generated game review article data and stores it in the description DB 18 of the storage unit 10g, so that it can be quickly read and provided in response to an external request. The issued identification code may also include information indicating the opposing team and information indicating the game date and time, etc., or the information indicating the opposing team and the game date and time, etc., may be associated with the identification code and stored in the description DB 18, making it easier to search for the generated game review article data.

[0099] 7 is a flowchart showing the procedure for the game description generation device 10 (MPU 10a) to automatically generate a match review article in accordance with the description generation program P2 described above (showing an example of the processing content of the game description generation method according to the present invention, and this processing content is executed by the game description generation device 10 (MPU 10a) by the description generation program P2). Below, the processing content of the game description generation device 10 (MPU 10a) in accordance with the description generation program P2 will be organized and explained in accordance with this flowchart.

[0100] First, the game record generation device 10 (MPU 10a) determines whether or not it has received and acquired game data distributed from the external game data distribution system 2 as the game ends (S1). If it has not acquired game data (S1: NO), it waits for the game data to be distributed. If it has received game data (S1: YES), the game record generation device 10 (MPU 10a) determines (detects) whether each of the opposing teams falls under any of the conditions 1 to 7 defined in the judgment table 13, based on the received game data (S2). For example, in the case of Team XXX and Team YYY in the above example, it is simultaneously determined at this stage (S2) whether Team XXX falls under any of the conditions 1 to 7 as one team, and whether Team YYY falls under any of the conditions 1 to 7 as one team.

[0101] At this stage (S2), the game description generation device 10 (MPU 10a) also performs the process of calculating the points of each player based on the game data, detecting whether or not there is a player who exceeds the threshold, and determining that condition 7 is met if there is a player who exceeds the threshold. At this stage (S2), the game description generation device 10 (MPU 10a) also detects whether or not there is a player who has achieved a record based on the game data, and determines (specifies) that condition 7 is met if there is a player who has achieved a record. As mentioned above, in this determination process, there may be cases where multiple conditions are met.

[0102] Next, the match record generation device 10 (MPU 10a) performs a process (S3) to identify a pattern of content corresponding to the corresponding condition from the judgment table 13. At this stage (S3), a pattern is identified for each of the opposing teams, and if there are players who have exceeded a threshold or achieved a record, a pattern corresponding to the identified condition 7 for those players is also identified.

[0103] The match record generation device 10 (MPU 10a) then performs a process of automatically generating a match review article based on the acquired match data and using the template table 17, with the review article including content related to the pattern identified in step S3 (S4). As described above, a pattern has been identified for each of the opposing teams, so at this step (S4), at least two types of match review article are generated: one including a review article focusing on the winning team, and one including a review article focusing on the losing team. Furthermore, at this step (S4), if there was a player who exceeded the threshold or achieved a record, and the losing team met condition 2, a match review article including a review article focusing on the achievements of such player or record achievement will be generated in addition to the match review article generated based on pattern J corresponding to condition 7.

[0104] Finally, the match description generating device 10 (MPU 10a) issues an identification code for each generated match review article, and performs a process of storing data related to the generated match review article in association with the issued identification code in the description DB 18 (S5). By storing match reviews (data related to match review articles) in association with identification codes in the description DB 18 in this way, match review articles can be provided promptly in response to external requests, etc.

[0105] Note that various situations can be envisioned regarding the specific manner in which the generated game review articles are provided, depending on the content of the service, business policies, etc. For example, if the game review generation device 10 provides a game review article that includes a description focusing on a specific baseball team to a website related to a specific baseball team (e.g., the website of Team XXX), even if the specific baseball team loses a game, the game review article including a description focusing on the specific baseball team will be posted on the website related to the specific baseball team. In this way, fans of the specific baseball team can enjoy a compelling game review article that lets them understand the performance of their favorite baseball team even in a losing game, simply by accessing the website related to the specific baseball team using a communication terminal (see user terminals T1, T2, T3, T4, etc. shown in FIG. 1 ), without performing any special operations.

[0106] Furthermore, if the game description generation device 10 provides a game review article that includes a description focusing on a particular player to a website related to that player, fans of that particular player can enjoy a well-read game review article that allows them to understand the performance of their favorite baseball team even in a losing game, without having to perform any special operations, simply by accessing the website related to that particular player's baseball team using a communication terminal (see user terminals T1, T2, T3, T4, etc. shown in Figure 1).

[0107] Furthermore, if the game description generation device 10 provides all generated game review articles to a website aimed at a wide range of baseball fans, and the website posts multiple patterns of game review articles for each game, game review articles that suit the preferences of a wide range of baseball fans will be included, and the needs of users who enjoy baseball from various perspectives will be met. Next, we will show a specific example of a game review article that was automatically generated by the game description generation device 10 (MPU 10a) using the present invention based on acquired game data (baseball game data in which Team XXX, batting first, faced Team YYY, batting second).

[0108] FIG. 8(a) shows a score 20 for each inning illustrating Example 1 of the results of a baseball game between one team, Team XXX, and another team, Team YYY, and information corresponding to this score 20 is included in the game data. FIG. 8(b) shows an example game review article 30 automatically generated for this example game result of score 20. As shown in the score 20 in FIG. 8(a), Team XXX won against Team YYY in this game, and therefore, one team, Team XXX, falls under Condition 1 (when one team wins the game) specified in the judgment table 13, and the example game review article 30 in FIG. 8(b) is automatically generated to include a description of Team XXX based on Pattern A (content related to one team winning the game).

[0109] Specifically, in the example game review article 30 shown in FIG. 8(b), the description 30a, "Team XXX easily defeated Team YYY." corresponds mainly to the description corresponding to Pattern A. Note that the underlines in the example game review article 30 are added to make the main relevant description easier to understand, and are not underlined in the original game review article (the same applies to other example articles). Note that in the example game review article 30, the description 30b, "Team XXX got off to a good start in the first inning with a two-run home run by O○ with two outs and a runner on first base," the description 30c, "In the third inning, O○ added another run with a double, extending the lead to three runs," and the description 30d, "Team XXX added two runs in the eighth inning, and in the ninth inning, Fuji○ comfortably held off Team YYY's batting line-up for a 5-0 victory," also correspond to descriptions related to Pattern A. In this way, the example game review article 30 in FIG. 8(b) contains content detailing the points that Team XXX won in the match, making it a game review article that will be popular with fans of Team XXX.

[0110] FIG. 9(a) shows an example 2 of the results of a baseball game between one team, Team XXX, and another team, Team YYY, with the score 21 for each inning (the score 21 is information included in the game data). FIG. 9(b) shows an example game review article 31 automatically generated for this example game result of the score 21. As shown in the score 21 in FIG. 9(a), Team XXX lost to Team YYY in this game, but there were times during the game when Team XXX was leading. Therefore, Team XXX, one of the teams, falls under Condition 3 specified in the decision table 13 (when one team lost the game but there was a period during the game when the score of that team exceeded the score of the other team). Therefore, the example game review article 31 in FIG. 9(b) was automatically generated to include a description of Team XXX based on Pattern B (content indicating that one team lost the game but was leading during the game).

[0111] Specifically, in the example game review article 31 shown in Figure 9(b), the following statements correspond mainly to Pattern B: "In the first inning, with two outs and a runner on first base, Team XXX got off to a good start with a two-run home run by O○, and in the third inning, they added another run with a double by O○, extending their lead to three runs."; and "In the ninth inning, with one out and a runner on second base, Team XXX had a chance to tie the game with one hit, but O○ hit an infield grounder and O○ struck out, so they were unable to get the last hit and suffered a close come-from-behind loss." Thus, the example game review article 31 in Figure 9(b) contains content that describes the good points of Team XXX during the game, such as the fact that Team XXX lost the game but was leading during the game and how the offense performed when they gained that lead, making it a suitable game review article for fans of Team XXX.

[0112] Figure 10(a) shows the score 22 for each inning (the score 22 is information included in the game data) showing Example 3 of the results of a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 10(b) shows an example of a game review article 32 that was automatically generated for this example game result of score 22. As shown in score 22 in Figure 10(a), in this game, Team XXX lost to Team YYY, but at one point during the game, Team XXX was leading by a score difference of 5 to 0. Therefore, one of the teams, Team XXX, falls under Condition 4 defined in Decision Table 13 (when one team lost the match, but there was a period during the match when one team's score exceeded the other team's score, and the difference in score during that period was greater than 4 points), and the example match review article 32 in Figure 10(b) was automatically generated to include a description of Team XXX based on Pattern C (content related to the fact that one team lost the match, but at one point during the match they were leading by a score difference that was difficult to catch up with).

[0113] Specifically, in the example game review article 32 shown in Figure 10(b), the following passage 32a, "Team XXX got off to a good start in the first inning with a two-run home run by Oh○ with two outs and a runner on first base, and in the third inning, Oh○ hit a three-run home run in his second consecutive at-bat, putting them in the lead at 5-0 early on," and the passage 32b, "In the sixth inning, Team XXX's second pitcher Fuji○ allowed a solo home run to △, allowing the team to take the lead. The third and subsequent pitchers also allowed runs, resulting in a 5-10 loss. The team was unable to take advantage of their early lead, resulting in a disappointing loss," are mainly descriptions that correspond to pattern C. In this way, the example game review article 32 in Figure 10(b) describes how Team XXX lost the game, but at one point during the game they took a huge lead with a score difference that the opposing team could not catch up with even if they hit a grand slam home run, and what it was like when they took such a large lead, so it is a game review article that is suitable for fans of Team XXX, etc., as it contains content that explains the good points of Team XXX during the game.

[0114] Figure 11(a) shows the score 23 for each inning (the score 23 is information included in the game data) that shows Example 4 of the results of a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 11(b) shows an example game review article 33 that was automatically generated for this example game result of score 23. As shown by the score 23 in Figure 11(a), in this game, Team XXX lost to Team YYY, but Team XXX was in the lead during the game, and this lead lasted for more than half of the entire game period (all innings) (Team XXX was in the lead from the top of the first inning to the top of the ninth inning). Therefore, one of the teams, Team XXX, falls under Condition 5 (when one team lost the match, but there was a period during the match when one team's score exceeded the other team's score, and this period lasted for more than half of the matches) specified in Decision Table 13, and the example match review article 33 in Figure 11(b) was automatically generated to include a description of Team XXX based on Pattern D (content related to one team losing the match, but leading the other team for most of the match).

[0115] Specifically, in the example game review article 33 shown in Figure 11(b), the following passage 33a, "Team XXX got off to a good start in the first inning with a two-run home run by Oh○ with two outs and a runner on first base. In the third inning, Oh○ hit two consecutive two-run home runs, putting them in the lead at 4-0 early on," and the passage 33b, "In the ninth inning, when a shutout win was finally in sight, Iwa○ lost control of his pitching, walked three batters to load the bases, and gave up a double to Saka○ that cleared the bases, and he was taken out of the game with the score just one point behind. However, Fuji○, who was entrusted with the team's place, was unable to stop the momentum of Team YYY's batting line, and gave up a double to Moto○ and a timely hit to Oki, turning the game around at the last minute. Despite leading for most of the game, they suffered an unexpected reversal defeat in the end." mainly correspond to the passage 33b, which reads: In this way, the example game review article 33 in Figure 11(b) contains content that explains the good points of Team XXX during the game, such as that although Team XXX lost the game, they were leading for most of the game, and how they performed offensively in relation to that lead, making it a suitable game review article for fans of Team XXX, etc.

[0116] FIG. 12(a) shows an example 5 of a baseball game result between team XXX and team YYY, with inning scores 24 (the score 24 is information included in the game data). FIG. 12(b) shows an example game review article 34 automatically generated for the example game result of the score 24. As shown in the score 24 in FIG. 12(a), team XXX lost to team YYY in this game, but made two comebacks during the game (in the top of the second inning and the top of the seventh inning). Therefore, team XXX, one of the teams, meets condition 6 specified in decision table 13 (when one team lost the game, but there was a period during the game when the team's score exceeded the other team's, and such periods occurred multiple times). Therefore, the example game review article 34 in FIG. 12(b) was automatically generated to include a description of team XXX based on pattern E (content related to one team losing the game, but making multiple comebacks or winning streaks during the game).

[0117] Specifically, in the example game review article 34 shown in Figure 12(b), the following statements correspond mainly to Pattern E: "In the first inning, Team XXX's starting pitcher Iwa○ was hit with consecutive hits by Team YYY's batting line, allowing them to take a two-run lead, but immediately afterwards in the second inning, Oh○ hit a three-run home run and they succeeded in turning the game around."; "In the seventh inning, Team XXX, desperate to catch up, managed to turn the game around again with consecutive hits from Ito○ and Oh○, making the score 6 to 5."; and "In the ninth inning, however, Team XXX, who was sent to the mound in an attempt to escape, walked Saka○ with two outs, and then allowed ○ to hit a two-run home run to turn the game around, resulting in an unexpected walk-off loss. They were unable to protect their two leads, resulting in a bitter defeat." In this way, the match review article example 34 in Figure 12(b) is suitable for fans of Team XXX because it contains content that explains what Team XXX did well during the match, such as the fact that Team XXX lost the match but made multiple comebacks during the match, and the state of their attacks during those comebacks. Note that this example is for a case where Team XXX made two comebacks, but a match review article similar to the above will be automatically generated even if Team XXX has multiple wins.

[0118] Figure 13(a) shows the score 25 for each inning (score 25 is information included in the game data) that shows Example 6 of the results of a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 13(b) shows Example 35 of a game review article that was automatically generated for this example game result of score 25. As shown by score 25 in Figure 13(a), in this game, Team XXX lost to Team YYY, but there was a period during the game when Team XXX was leading, and this lead lasted for more than half of the entire game period (all innings) (leading from the top of the first inning to the top of the ninth inning), and the maximum score difference by Team XXX during that lead was 5 points (leading 5-0).

[0119] Therefore, one of the teams, Team XXX, falls under Condition 4 (where one team lost the game, but took a large lead during the game that could not be caught up even if the other team hit a grand slam (a play that can score the maximum number of points)) and Condition 5 (where one team lost the game, but the number of times they were ahead during the game exceeded half of the total number of innings in the game) specified in Decision Table 13, and Example Game Review 35 in Figure 13(b) was automatically generated to include a description for Team XXX based on Pattern F (where one team lost the game, but took a lead during the game by a large lead that could not be easily caught up, and that lead lasted for most of the game).

[0120] Specifically, in the example game review article 35 shown in Figure 13(b), the following passage 35a, "Team XXX got off to a good start in the first inning with a two-run home run by Oh○ with two outs and a runner on first base, and in the third inning, Oh○ hit a three-run home run in his second consecutive at-bat, putting them in a large lead of 5-0 early on and in the lead." and the passage 35b, "In the ninth inning, when a shutout win was finally in sight, Iwa○ lost control of his pitching, walked three batters to load the bases, and gave up a double to Saka○ that cleared the bases, and he was taken out of the game with the score just two points behind. However, Fuji○, who was entrusted with the team's place, was unable to stop the momentum of Team YYY's batting line, and gave up a double to Moto○ and a home run to Oki, turning the game around at the last minute. Despite having an overwhelming lead for most of the game, they suffered an unexpected reversal defeat." mainly correspond to the passage 35a. In this way, the game review example 35 in Figure 13(b) is about Team XXX, and although they lost the game, at one point during the game they took a huge lead with a score difference that the opposing team could not catch up even if they hit a grand slam home run, they were in the lead for most of the game, and the state of their offense in relation to the lead. This content includes descriptions that explain the good points of Team XXX during the game, making it a suitable game review article for fans of Team XXX, etc.

[0121] Figure 14(a) shows the score 26 for each inning showing an example 7 of the result of a baseball game between one team, Team XXX, and another team, Team YYY, and Figure 14(b) shows an example game review article 36 that was automatically generated for the example game result of this score 26. As shown in the score 26 in Figure 14(a), in this game, Team XXX lost to Team YYY, but Team XXX came back twice during the game (in the top of the second inning and the top of the eighth inning), and the largest score difference when Team XXX was initially in the lead was 5 points (a 7-2 lead).

[0122] Therefore, one of the teams, Team XXX, falls under Condition 4 (where one team lost the game, but took a large lead during the game with a score difference that could not be easily caught up even if the other team hit a grand slam (a play that can score the maximum number of points)) and Condition 6 (where one team lost the game, but there were multiple innings during the game in which the other team was in the lead), as specified in Decision Table 13. The example game review article 36 in Figure 14(b) was automatically generated to include a description for Team XXX based on Pattern G (where one team lost the game, but took a lead during the game with a score difference that could not be easily caught up, and also had multiple comebacks or wins during the game).

[0123] Specifically, in the example game review article 36 shown in Figure 14(b), there is a section 36a that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take a two-run lead, but immediately after in the second inning, O○ hit a three-run home run and successfully turned the game around.", a section 36b that reads, "In the third inning, Ogawa hit a grand slam, giving Team XXX a large lead of 7 to 2 in the early stages and gaining an advantage in the game." (content corresponding to condition 4), and a section 36c that reads, "In the eighth inning, Team XXX The statements in 36c (corresponding to condition 6) that "With one out and a runner on second base, O○ hit his second home run of the day and turned the game around again," and the statement in 36d that "However, in the ninth inning, Team XXX's Fuji○, who was sent to the mound in an attempt to pull away, walked Saka○ with two outs, and then allowed ○ to hit a two-run home run to take the lead, resulting in an unexpected walk-off loss. The team was unable to hold on to their two leads and suffered a bitter defeat," are mainly statements that correspond to pattern G.

[0124] In this way, the example game review article 36 in Figure 14(b) contains descriptions that explain the good points of Team XXX during the game, such as how, although they lost the game, they made multiple comebacks during the game, how they took a large lead with a score difference that the opposing team could not catch up even if they hit a grand slam home run, and how their offense performed when they made the comeback or took a large lead, making it a suitable game review article for fans of Team XXX, etc.

[0125] Figure 15(a) shows the score 27 for each inning showing an example 8 of the result of a baseball game between one team, Team XXX, and another team, Team YYY, and Figure 15(b) shows an example game review article 37 that was automatically generated for this example game result of score 27. As shown in score 27 in Figure 15(a), in this game, Team XXX lost to Team YYY, but they made two comebacks during the game (in the top of the second inning and the top of the eighth inning), and there were also times when they were in the lead for more innings than the majority of the game (leading from the top of the third to the top of the seventh inning, and from the top of the eighth to the top of the ninth inning).

[0126] Therefore, one of the teams, Team XXX, falls under Condition 5 (when one team lost the game, but the number of times they were ahead during the game exceeded half of the total number of innings in the game) and Condition 6 (when one team lost the game, but there were multiple innings during the game when they were ahead), as specified in Decision Table 13. Example game review article 37 in Figure 15(b) was automatically generated to include a description of Team XXX based on Pattern H (content related to one team losing the game, but being ahead for most of the game, and having multiple comebacks or winning attempts during the game).

[0127] Specifically, in the example game review article 37 shown in Figure 15(b), there is a section 37a (related to conditions 5 and 6) that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take a two-run lead, but immediately after that, in the second inning, Dai○ hit a three-run home run and they succeeded in turning the game around."; there is a section 37b (related to conditions 5 and 6) that reads, "However, in the seventh inning, after a long lead, the second pitcher, Fuji○, and the third pitcher, Naka○, were hit by consecutive hits and lost three runs, allowing Team YYY to take the lead."; The following statements correspond mainly to pattern H: "In the 8th inning, with one out and a runner on second base, Dai○ hit his second home run of the day, turning the game around again." (Statement relating to conditions 5 and 6); and "However, in the 9th inning, Team XXX's Fuji○, who was sent to the mound in an attempt to pull away, walked Saka○ with two outs, and then allowed ○ to hit a two-run home run to take the lead, resulting in an unexpected walk-off loss. They were unable to hold on to their two leads, resulting in a bitter defeat." (Statement relating to conditions 5 and 6).

[0128] In this way, the example game review article 37 in Figure 15(b) contains content that explains the good points of Team XXX during the game, such as that although Team XXX lost the game, they made multiple comebacks during the game, that they were leading for most of the game, and the offensive situation when they made comebacks or took the lead, making it a game review article that is suitable for fans of Team XXX, etc.

[0129] Figure 16(a) shows the score 28 for each inning showing an example 9 of the result of a baseball game between one team, Team XXX, and another team, Team YYY, and Figure 16(b) shows an example game review article 38 that was automatically generated for this example game result of score 28. As shown in score 28 in Figure 16(a), in this game, Team XXX lost to Team YYY, but Team XXX came back twice during the game (in the top of the second inning and the top of the eighth inning), and the largest difference in points when Team XXX was initially in the lead was 5 points (a 7-2 lead), and there were times when Team XXX was in the lead for more innings than the majority of the game (leading from the top of the second to the top of the seventh inning and from the top of the eighth to the top of the ninth inning).

[0130] Therefore, one of the teams, Team XXX, falls under Condition 4 (where one team lost the game, but during the game the other team took a large lead by such a large margin that it could not be caught up even if the other team hit a grand slam (a play that can score the maximum number of points)), Condition 5 (where one team lost the game, but the number of times they were ahead during the game exceeded half of the total number of innings in the game), and Condition 6 (where one team lost the game, but there were multiple innings during the game in which they were ahead), all of which are specified in Decision Table 13. Example game review article 38 in Figure 16(b) was automatically generated to include a description for Team XXX based on Pattern I (where one team lost the game, but during the game they were ahead by such a large margin that it was not easy to catch up, and this lead lasted for most of the game, and there were multiple comebacks or wins during the game).

[0131] Specifically, in the example game review article 38 shown in Figure 16(b), there is a section 38a (related to conditions 5 and 6) that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take a two-run lead. However, immediately after that, in the second inning, Dai○ hit a three-run home run and successfully turned the game around."; a section 38b (related to condition 4) that reads, "In the fifth inning, Ito○ hit a grand slam, extending the lead to 7 to 2, a one-sided score difference."; and a section 38c (related to condition 4) that reads, "However, in the seventh inning, after a long lead, the second pitcher, Fuji○, and the third pitcher, Naka○, were hit by concentrated hits, causing Team XXX to lose six runs and allowing Team YYY to take the lead." The statements that correspond mainly to Pattern I are 38c (a statement related to conditions 5 and 6), 38d (a statement related to conditions 5 and 6) which states, "However, in the eighth inning, with one out and a runner on second base, Dai○ hit his second home run of the day and turned the game around again," and 38e (a statement related to conditions 4, 5, and 6) which states, "However, in the ninth inning, Team XXX, having sent Fuji○ to the mound in an attempt to pull away, walked Saka○ with two outs, and then allowed ○ to hit a two-run home run to take the lead, resulting in an unexpected walk-off loss. They were unable to hold on to their two leads and it was a bitter defeat."

[0132] In this way, the example game review article 38 in Figure 16(b) contains descriptions of Team XXX's good points during the game, such as that although they lost the game, they made multiple comebacks during the game, that they took a large lead with a score difference that the opposing team could not catch up even if they hit a grand slam home run, that they were in the lead for most of the game, and how their offense performed when they made comebacks or took the lead, making it a suitable game review article for fans of Team XXX, etc.

[0133] Figure 17(a) shows an example 10 of the results of a baseball game between one team, Team XXX, and another team, Team YYY, with a score 29 for each inning, and Figure 17(b) shows an example of a game review article 39 that was automatically generated for this example game result of score 29. As shown in score 29 in Figure 17(a), in this game, Team XXX lost to Team YYY, and although Team XXX scored points during the game, they never had a lead, and all of Team XXX's points were due to two home runs by player O○.

[0134] Therefore, one of the teams, Team XXX, meets Condition 2 (when one of the teams loses the game and has zero points or no lead during the game) and Condition 7 (when there is a player (batter or pitcher) whose points calculated using a certain formula exceed a threshold) specified in Decision Table 13, and Example Game Review Article 39 in Figure 17(b) was automatically generated to include a description of Team XXX based on Pattern J (content related to the fact that one of the teams lost the game, but that there was a player on one of the teams who performed well during the game).

[0135] Specifically, in the example game review article 39 shown in Figure 17(b), the following statements correspond primarily to pattern J: "No. 4 hitter O○ performed brilliantly in a tough game. He hit a solo home run in his third at-bat in the sixth inning, sparking a counterattack, and then in the eighth inning, with a runner on first base after a walk, he hit two consecutive home runs in two at-bats, scoring three RBIs that day." Also, the following statement 39b (related to condition 2) states, "Though the team lost 3-8 due to a large number of runs allowed early on, they showed tenacity that will help them do better tomorrow." Thus, the example game review article 39 in Figure 17(b) describes Team XXX's loss and there were no notable events for the team. However, since there was a player on Team XXX who performed well, the content includes a description of that player's performance, making it a suitable game review article for fans who want to know how Team XXX's players performed, even in a losing game.

[0136] Figure 18(a) shows an example 11 of the results of a baseball game between one team, Team XXX, and another team, Team YYY, with a score 40 for each inning, and Figure 18(b) shows an example of a game review article 41 that was automatically generated for this example of the game result with score 40. As shown by the score 40 in Figure 18(a), in this game, Team XXX was shut out by Team YYY, and the content of the game did not allow for any particular highlighting of Team XXX's strengths or outstanding players.

[0137] Therefore, one of the teams, Team XXX, only meets Condition 2 (when one team loses the match and has zero points or no lead during the match) specified in the decision table 13, and the example game review article 41 in Figure 18(b) is automatically generated to include a description for Team XXX based on Pattern K (content related to one team losing the match and praising the opposing team (the other team)).

[0138] Specifically, in the example game review article 41 shown in Figure 18(b), the section 41a (description related to condition 2) that reads, "Team YYY's batting line-up piled up strikeouts against Team YYY's starter, O-no, who continued to pitch impressively, and failed to reach second base, resulting in a shutout loss with only three hits," and the section 41b (description related to condition 2) that reads, "Team XXX allowed Team YYY, whose pitching and batting worked well together, to win 0-10," are descriptions that primarily correspond to pattern K. In this way, the example game review article 41 in Figure 18(b) contains descriptions that reluctantly praise the opposing team, since Team XXX suffered a crushing defeat in the game, there were no notable events in the game, and no players on Team XXX performed well, making it a game review article suitable for fans who appreciate a spirit of fair play, etc.

[0139] As described above, in this invention, even if a team loses a game, if the losing team was leading during the game, if the difference in points due to that lead was so large that it could not be caught up even with the maximum score in one play, if that lead lasted for most of the game, or if there were multiple comebacks or wins during the game, a game review article will be automatically generated that includes these details, making it possible to provide a description that fans of the losing team can enjoy.

[0140] The present invention is not limited to the above, and various modifications are possible. First, the game for which a description is generated is not limited to a baseball game. Any game in which one player competes against another in a scoring format can generate a description of the game based on game data indicating the details of the game. Specifically, the present invention can also be applied to various sports matches, such as soccer, rugby, basketball, American football, volleyball, softball, ice hockey, tennis, table tennis, badminton, handball, beach volleyball, curling, futsal, hockey, polo, lacrosse, indiaca, cricket, and water polo. These matches include those between teams (such as soccer and basketball), those between multiple players (such as doubles matches in tennis, table tennis, and badminton), and those between individuals (such as singles matches in tennis, table tennis, and badminton).

[0141] Furthermore, when the various sports mentioned above are used as the subject of description generation in the present invention, the numerical value of the score difference (a score difference greater than the maximum score that can be obtained in one play) in condition 4 (when one team loses the game but has a lead of more than four points during the game) in the decision table 13 shown in Figure 3 may differ from that in baseball (note that softball is the same as baseball).

[0142] For example, in rugby, the maximum number of points that can be obtained in one play is five points for a try, so the score difference pertaining to condition 4 in rugby is greater than five points. In basketball, the maximum number of points that can be obtained in one play is three points for a goal from outside the three-point line, so the score difference pertaining to condition 4 in basketball is greater than three points. In American football, the maximum number of points that can be obtained in one play is six points for a touchdown, so the score difference pertaining to condition 4 in American football is greater than six points.

[0143] Furthermore, when the various sports mentioned above are used as the subject of description generation in the present invention, the meaning of the number of innings in condition 5 (when one team lost the game but was ahead more than half the number of times in the game) in the judgment table 13 shown in Fig. 3 may differ from that in baseball (note that softball is the same as baseball). This condition 5 is intended to determine whether or not a team was ahead for most of the game, so for sports that do not use the concept of innings, a judgment will be made appropriately based on the meaning of condition 5.

[0144] For example, in sports where the game is decided by time, condition 5 is judged by whether the range of time during the game in which the team was leading (corresponding to the period) exceeded half of the total time of the game (corresponding to the entire period). Also, in sports where innings are not used but something with a similar meaning (for example, a set in tennis or an end in curling), the same judgment as for innings is made as to whether condition 5 is met.

[0145] Furthermore, when the various sports mentioned above are the subject of description generation in the present invention, the meaning of "when there are multiple breaks in the lead innings" in condition 6 (when one team loses the game but there are multiple breaks in the lead innings during the game) in the judgment table 13 shown in Fig. 3 (particularly the meaning of "innings") differs from that in baseball, as in the case of condition 5 mentioned above (note that softball is the same as baseball). Condition 6 is intended to determine whether a team has lost a game but has made multiple comebacks or won, so for sports that do not use the concept of innings, judgments are made appropriately based on the meaning of condition 6.

[0146] For example, in sports where the game is decided by time, condition 6 is determined by whether there were multiple periods during the game when one team was in the lead (periods when one team's score was higher than the other team's score) that were uninterrupted. Also, in sports that do not use innings but use something with a similar meaning to innings (for example, sets in tennis or ends in curling), whether condition 6 is met is determined in the same way as innings in baseball.

[0147] 3, conditions 2 to 6, which relate to when one team loses the match, do not all need to be used and any of them can be omitted depending on the specifications of the service to be provided. That is, conditions 4 to 6 all specify more detailed content for condition 3 (when one team loses the match but has a lead during the match), so when automatically generating descriptions using simple specifications, it is conceivable to omit any or all of these conditions 4 to 6.

[0148] Rather than omitting all of these conditions 4 to 6, depending on the specifications of the service to be provided and the type of sports match to be automatically generated, it is possible to omit only condition 4, omit conditions 4 and 5, omit conditions 4 and 6, omit only condition 5, omit conditions 5 and 6, or omit only condition 6. Furthermore, when the focus is on teams and the performance of individual players is not important, it is also possible to omit condition 7.

[0149] Furthermore, in the above-described processing, at the stage S2 of the flowchart shown in FIG. 7, the judgment table 13 shown in FIG. 3 is used to judge at the same stage (simultaneously judge) which of conditions 1 to 7 applies, but it is also possible to judge each condition in stages in accordance with the provisions of the description generation program P2 of the modified example.

[0150] 19 and 20 show an example of a processing procedure when each condition is judged step by step (showing an example of the processing contents of the game description generation method according to the present invention). The processing of such a modified example is defined in the description generation program P2 of the modified example, and is executed by the MPU 10a based on the definition.

[0151] As shown in the flowchart of the modified example in Fig. 19 (first half of the flowchart), in this modified example, the first step (S10) is the same as step S1 in the flowchart of Fig. 7, in which the game description generation device 10 (MPU 10a) determines whether or not it has received and acquired game data distributed from the external game data distribution system 2 as the game ends (S10). If it has not acquired game data (S10: NO), it enters a state of waiting for the game data to be distributed.

[0152] When game data is received (S10: YES), the MPU 10a identifies one of the opposing teams in the game related to the game data as one of the teams, and determines whether or not Condition 1 applies to that one of the teams, i.e., whether or not that one of the teams won the game, based on the received game data (S11). Note that, in this modified example, as in the case of the flowchart of Fig. 7 described above, a baseball game is targeted, and the team batting first in this baseball game is identified as one of the teams.

[0153] If one team did not win the match, i.e., if one team lost the match (S11: NO), the MPU 10a determines whether there was a period during the match when one team's score exceeded the other team's score, i.e., whether one team was in the lead during the match, based on the received match data (S12). Note that the processing at this S12 stage determines whether condition 3 (see FIG. 4) in the decision table 13 of FIG. 3 is met.

[0154] If one team is leading during the game (S12: YES), the MPU 10a determines whether the lead lasted for more than half of the entire game period based on the received game data (S13). Note that the processing at this S13 stage determines whether condition 5 (see FIG. 4) in the decision table 13 in FIG. 3 is met.

[0155] If one team was in the lead for more than half of the entire game (the period from the start to the end of the game) (S13: YES), that is, if one team meets condition 5 (see FIG. 4), the MPU 10a automatically generates a game review article (S24) including a description based on the fact that the one team lost the game but led the other team for most of the game. The description of the above content included in the game review article generated in this manner corresponds to pattern D (see FIG. 5) in the decision table 13 of FIG. 3. The automatic generation at this time is also performed using the template table 17 based on the acquired game data (the same applies below). After the game review article is automatically generated, the process proceeds to S17 of the flowchart of FIG. 20 (the latter half of the flowchart), where the MPU 10a issues an identification code to the generated game review article and stores data related to the generated game review article in the description DB 18 in association with the issued identification code.

[0156] Returning to step S13 of the flowchart in Fig. 19, if the period in which one team was leading did not exceed half of the entire period of the match (S13: NO), the process proceeds to step S14 of the flowchart in Fig. 20, where the MPU 10a determines whether or not there were multiple interruptions in the period in which one team was leading, i.e., whether or not one team made a comeback or won again during the match. Note that the process at step S14 determines whether or not condition 6 (see Fig. 4) in the decision table 13 in Fig. 3 is met.

[0157] If one team has had a lead multiple times (S14: YES), the MPU 10a automatically generates a match review article including a description based on the fact that one team lost the match but made a comeback or won again during the match (S25). The description of the above content included in the match review article generated in this way corresponds to pattern E (see FIG. 5) in the decision table 13 of FIG. 3. After the match review article is automatically generated, the process proceeds to S17, as in the case of the above-mentioned S24 stage, where data related to the generated match review article is stored in the description DB 18.

[0158] If, at step S14, one team has not been in the lead for a number of times (NO at step S14), the MPU 10a next determines whether one team has been in the lead by a score difference of more than four points (a score difference of five points or more) during the game based on the received game data (step S15).The processing at step S15 determines whether condition 4 (see FIG. 4) in the decision table 13 of FIG. 3 is met.

[0159] If one team is leading by a score difference of more than four points (S15: YES), the MPU 10a automatically generates a game review article (S26) including a description based on the fact that, although one team lost the game, it was leading the other team by a score difference that was difficult to catch up with (a one-sided score difference that could not be reversed with a single play). Note that the description of the above content included in the game review article generated in this way corresponds to pattern C (see FIG. 5) in the decision table 13 of FIG. 3. After the game review article is automatically generated, the process proceeds to step S17, as in steps S24 and S25 described above, where data related to the generated game review article is stored in the description DB 18.

[0160] Furthermore, if at the above-mentioned stage S15 one team is not leading by a score difference of more than four points (S15: NO), this stage of "S15: NO" is reached through "S11: NO (if one team loses)", "S12: YES (if one team was leading during the match)", "S13: NO (if one team was leading for less than half of the entire match)", and "S14: NO (if one team was leading for no more than half of the entire match)", so MPU 10a determines that this corresponds to the case that meets the conditions of "S11: NO" and "S12: YES" (if one team lost the match but was leading during the match), and MPU 10a automatically generates a match review article including a description according to the content of the determined case (S16).

[0161] In addition, when the result of the S15 step is NO, considering the above-mentioned process, it is when only condition 3 (see FIG. 4) in the judgment table 13 of FIG. 3 is met, and the above-mentioned description included in the battle review article generated in the above-mentioned S16 step corresponds to pattern B (see FIG. 5) in the judgment table 13 of FIG. 3. Furthermore, when the battle review article is automatically generated, the process proceeds to S17 step, as in the above-mentioned S24, S25, etc. step, and a process of storing data related to the generated battle review article in the description DB 18 is performed.

[0162] 19, if one team was not in the lead during the game (S12: NO), the process proceeds to step S21, where the MPU 10a calculates points for each player who played in the game for one team using the point table 16 of FIG. 6 based on the received game data, and determines whether there is a player (batter or pitcher) whose points exceed a threshold value (for example, the threshold value of 50 points for the losing team). The process of step S21 determines whether condition 7 (see FIG. 4) in the decision table 13 of FIG. 3 is met.

[0163] If there is a player whose calculated points exceed the threshold (S21: YES), the MPU 10a automatically generates a match review article (S22) including a description of the performance of the player who exceeded the threshold (the content of the play corresponding to the calculated points) (a description of the player who performed well on the losing team). Note that the description of the content included in the match review article generated in this way corresponds to pattern J (see FIG. 5) in the decision table 13 of FIG. 3. After the match review article is automatically generated, the process proceeds to step S17, as in the cases of steps S24, S25, etc. described above, where data related to the generated match review article is stored in the description DB 18.

[0164] Furthermore, if there is no player whose calculated points exceed the threshold at the S21 stage (S21: NO), since this "S21: NO" stage was reached via "S11: NO (if one team loses)" and "S12: NO (if one team did not have a lead during the match)," the MPU 10a determines that one team lost the match without any good points and that there were no players on that team who performed well during the match (determining that only condition 2 (see FIG. 4) in the decision table 13 of FIG. 3 is met), and the MPU 10a automatically generates a match review article including a description praising the winning team (opponent team) (S23). Note that the description of the above content included in the match review article generated in this way corresponds to pattern K (see FIG. 5) in the decision table 13 of FIG. 3. After the battle review article is automatically generated, the process proceeds to step S17, in the same manner as in steps S24, S25, etc., where data relating to the generated battle review article is stored in the description DB 18.

[0165] 19, if one team wins (S11: YES), this corresponds to condition 1, and the MPU 10a automatically generates a match review article including a description indicating that one team won (S20), as in the conventional case. After the match review article is automatically generated, the process proceeds to step S17, and data related to the generated match review article is stored in the description DB 18, as in the cases of steps S24, S25, etc. described above.

[0166] The contents of the modified example explained in the flowcharts shown in Figures 19 and 20 above are merely examples, and any of the paragraphs relating to the judgment processing such as S12, 13, 14, 15, and 21 can be omitted as appropriate depending on the specifications of the service to be provided, just as in the case of using the judgment table 13 described above. Note that the flowchart of the modified example shown in Figures 19 and 20 differs from the flowchart shown in Figure 7 in that it does not generate a battle review article with content corresponding to patterns F to I in the judgment table 13 of Figure 3 (because the flowchart of the modified example shown in Figures 19 and 20 does not make a judgment that satisfies three or more conditions).

[0167] Furthermore, if a sports match between two teams is a tournament or knockout match, the match data distributed from the match data distribution system 2 shown in Fig. 1 also includes information indicating whether the match is a tournament or knockout match, and the match description generation device 10 (MPU 10a) can therefore determine whether the match is a tournament or knockout match and perform the description generation process accordingly. In other words, if the match is a tournament or knockout match, the outcome of the match is more important than the content of the match or the players who performed well, so condition 1 (see Fig. 4) in the decision table 13 in Fig. 3 and the determination at step S11 in the flowchart in Fig. 19 (determining whether one team lost the match) are also important in the automatic description generation process.

[0168] Specifically, when the match record generation device 10 (MPU 10a) determines (detects) that the target match is a tournament or knockout match based on the received match data, it determines whether one of the competing teams won based on the match data (processing according to condition 1 shown in FIG. 4 or step S11 described above). If it determines that one team (e.g., one team) won the match, the MPU 10a automatically generates a match review based on the match data, the beginning of which states that the one team won the match (one team beat the other team). On the other hand, if it determines that one team lost the match, the MPU 10a automatically generates a match review based on the match data, the beginning of which states that the other team won the match (the other team beat the one team). This generation method makes it possible to provide a match review that is easy to understand regarding "who won," which is important in tournament or knockout matches.

[0169] It is also possible to include a description of the player who achieved the record in the game review. For example, by adding a rule that associates the player (batter or pitcher) who achieved the record with points exceeding the threshold in point table 16 (see FIG. 6) according to condition 7 (when there is a player (batter or pitcher) whose points calculated from the rule exceed the threshold) shown in FIG. 4, a game review including a description of the player who achieved the record may be automatically generated.

[0170] There are records that can be achieved in just one game (for example, hitting for the cycle, a no-hitter, or a perfect game), and records that can be achieved in conjunction with past performance (for example, reaching base in 20 consecutive at-bats, hitting 1,000 hits, or achieving the 100th win). All of these records can be determined from the game data that is distributed.

[0171] Records achieved in only one game can be determined from the play of each batter or pitcher based on the game data that is distributed, so by using point table 16 that includes the point regulations related to the achievement of the above-mentioned record, condition 7 above (or step S21 in the flowchart) is met, and MPU 10a automatically generates a game review article that includes information related to the achievement of the record (information indicating the player who achieved the record and the record that was achieved) based on, for example, pattern J (content related to one team losing the game, but having a player (batter or pitcher) who performed well on one team during the game).

[0172] Similarly, for records that are achieved in conjunction with past performance, information indicating that such a record is about to be achieved is included in the distributed game data, so by combining the playing details of the team or player (batter or pitcher) regarding the imminent achievement of such a record with the information about the imminent achievement of the record described above, MPU 10a determines whether the record has been achieved, and if it determines that the record has been achieved, condition 7 is met, and if the team to which the player who met condition 7 belongs loses, MPU 10a will automatically generate a game review article based on pattern J, including information about the past record achievement (information indicating the player who achieved the record and the record that was achieved).

[0173] It is of course possible to include in the match review any record such as the one described above (a record achieved in only one game or a record achieved in conjunction with past results) regardless of the outcome of the game, separate from condition 7, and in particular, team records (such as achieving 1,000 wins as a team) may be included in the match review regardless of the outcome of the game. Furthermore, it is also possible to automatically generate a match review that includes a description of any record achieved, regardless of points or thresholds, without using the points table 16, with the achievement of that record being the only condition.

[0174] Furthermore, the descriptions automatically generated by the present invention can be included not only in match review articles, but also in match summaries, headlines, short articles, etc. When included in summaries, the MPU 10a automatically generates a match summary including the descriptions according to the present invention. Because summaries contain less overall description than the above-mentioned match review articles, the descriptions according to the present invention are also shorter in length than when included in match review articles, and templates that are more concise in expression are extracted from the template table 17 (see FIG. 2) and automatically generated.

[0175] Furthermore, when a description automatically generated by the present invention is included in a match headline, the headline itself will contain less information than the above-mentioned summary, and will use expressions unique to headlines, such as ending with a noun. Therefore, the description according to the present invention will also be shorter in length than when included in the above-mentioned summary, and a template with an impactful title will be extracted from template table 17 (see Figure 2) and automatically generated.

[0176] Furthermore, when a description automatically generated by the present invention is included in a short article about a match, the description according to the present invention is automatically generated according to the length of the description in the short article. If the length of the description in the short article falls within the range between a match review and a summary, the description according to the present invention is automatically generated using a suitable template with a description length that falls within that range (shorter than a match review, but longer than a summary). Also, if the length of the description in the short article falls within the range between a summary and a headline, the description according to the present invention is automatically generated using a suitable template with a description length that falls within that range (shorter than a summary, but longer than a headline). Note that the various modifications of Example 1 described above can also be used in appropriate combinations. [Example]

[0177] A match description generation system including a match description generation device according to a second embodiment (Example 2) of the present invention automatically generates descriptions at a point in time during a match as progress reports related to the point in time of the match. In the first embodiment (Example 1) described above, a match review article reporting the results and details of the match was automatically generated based on the match data transmitted from the match data distribution system after the match ended, but in Example 2, the match description generation device receives and acquires the match data including information about the point in time of the match transmitted by the match data distribution system, and automatically generates descriptions explaining the situation up to that point in time of the match (descriptions related to the point in time of the match) based on the acquired match data.

[0178] In the second embodiment, a game in which one side competes against another in a scoring format may be, for example, a baseball game, and the game description generation device automatically generates progress reports up to that point based on game data corresponding to a good stopping point in the baseball game (for example, a point in the game that can be expressed as a fraction, such as the end of the bottom of the third inning, which is when 1 / 3 of the game has elapsed, the end of the top of the fourth inning, which is when 1 / 2 of the game has elapsed, or the end of the bottom of the seventh inning, which is when 2 / 3 of the game has elapsed). Note that the generation point (generation time) of the progress report, which is a good stopping point in the game, is set and determined in advance by the game description generation device, but the generation time of the progress report may be changed as appropriate according to user requests, specifications, etc.

[0179] There are two main methods that can be envisioned for a match data distribution system to transmit match data. The first method is a method in which the match data distribution system sequentially transmits match data containing update (difference) information at intervals, and the second method is a method in which the match data distribution system transmits match data containing information from the start of the match to any point in the match at any time. Which of these two methods is used depends on the specifications of the match data distribution system, but the present invention can accommodate either method.

[0180] In an example of the first method, when the first inning of a baseball game ends, the game data distribution system transmits the game data for the first inning at the end of the first inning, and thereafter, similarly, transmits the game data for the second inning (second inning) at the end of the second inning, and so on, transmitting updated (differential) game data as needed. Since the game data transmitted using the first method does not contain duplicate content, the game description generation device, which is the receiver of the game data, stores and accumulates the game data in order upon receiving it, and generates progress reports at predetermined points in time (convenient points during the game, such as the end of the bottom of the third inning as mentioned above) based on the accumulated multiple game data.

[0181] There are various possible intervals for the game data transmitted using the first method. In addition to transmitting every inning as described above, it is also possible to transmit at intervals longer than one inning, such as every two innings or every three innings, or conversely, it is also possible to transmit at intervals shorter than one inning, such as every out, every batter, or every pitch.

[0182] The second method described above involves transmitting game data including content from the start of the game to the middle of the game at multiple points during the game. For example, when transmitting game data for each inning, game data for the first inning (first inning) of a baseball game is transmitted at the end of the first inning, game data transmitted at the end of the second inning includes content from the start of the game to the end of the second inning, game data transmitted at the end of the third inning includes content from the start of the game to the end of the third inning, and similarly, game data transmitted at the end of the fourth through ninth innings includes content from the start of the game to the end of the fourth through ninth innings (the same applies in the case of extra innings). With this second method, when game data is received, the game description generation device can automatically generate a progress report for the game at the time of reception based on the game data acquired through the reception.

[0183] In the second embodiment of the present invention, progress reports can be generated based on either of the match data transmitted by the two methods described above. Note that, in the first embodiment of the present invention, by accumulating and storing match data by the first method described above, match review articles can be generated based on the accumulated match data, and in the case of match data by the second method, match review articles are generated based on the match data transmitted as the match ends.

[0184] The hardware configuration and processing contents of Example 2 of the present invention are basically the same as those of Example 1 described above, and therefore, the same reference numerals as those of Example 1 will be used in the following description of Example 2 for the same components as those of Example 1. Examples of components in Example 2 that are different from those of Example 1 include the contents of Conditions 1 to 7 and Patterns A to K in the judgment table 13 (see FIG. 3), which are suitable for generating entries during a match. However, the grid configuration of the judgment table 13 itself (the grid (matrix) configuration shown in FIG. 3) is the same in Example 2.

[0185] 21 shows a condition content list 54 that lists the specific contents of conditions 1 to 7 of Example 2, which are arranged in the leftmost vertical column 13a of the judgment table 13. Condition 1 of Example 2 specifies the content "when one team is leading the other team at some point during the game." Note that, in Example 2 as well, the judgment table 13 determines one of the opposing teams in a baseball game as one team (for example, the team batting first is provisionally determined as one team), and determines the remaining opposing team as the other team, and then specifies the content of each condition with a focus on that one team.

[0186] Condition 2 specifies that "one team's score falls below the other team's score at some point in the game, and up until that point, one team's score was zero or one team had not been in the lead." Below, condition 3 specifies that "one team's score falls below the other team's score at some point in the game, but there was a period up until that point when one team's score was higher than the other team's," condition 4 specifies that "one team's score falls below the other team's score at some point in the game, but there was a period up until that point when one team's score was higher than the other team's, and the difference in score at that time was more than four points," and condition 5 specifies that "one team's score falls below the other team's score at some point in the game, but Condition 6 stipulates that "up until that point, there has been a period in which one team's score has exceeded the other team's score, and that period has lasted for more than half of the number of games played up to that point," Condition 7 stipulates that "up until that point, there has been a period in which one team's score has fallen below the other team's score, but there has been a period in which one team's score has exceeded the other team's score, and there have been multiple such periods," and Condition 8 stipulates that "up until that point in the game, there is a player (batter or pitcher) whose points calculated based on the points table exceed the threshold."

[0187] Condition 1 in Example 2, "when one team is leading the other team at some point in the game," means "when one team's score exceeds the other team's score," and this condition corresponds to a situation in which one team is winning over the other team, even in an article reporting the progress of the game up to that point in the game.

[0188] Condition 2 in Example 2, "When one team's score is lower than the other team's score at some point in the game, and one team has had zero points or not been in the lead up to that point," means "When one team is losing to the other team at some point in the game, and one team has had no points or not been in the lead up to that point," and corresponds to the case when one team is losing at some point in the game, and the losing team has not had anything particularly noteworthy (nothing good) up to that point in the game.

[0189] Condition 3 in Example 2, "When one team's score is lower than the other team's score at some point in the game, but there was a time up until that point when one team's score was higher than the other team's," means "When one team is losing to the other team at some point in the game, but there was a time up until that point when one team was leading the other team," and corresponds to the fact that for a team that is trailing at some point in the game, there was a time when one team's score was higher than the other team's score (a time when they were leading), which is a highlight of the game up until that point.

[0190] Condition 4 in Example 2 states, "When one team's score is lower than the other team's at some point in the game, but there has been a period up until that point when one team's score was higher than the other team's, and the difference in score during that period was greater than four points" means "When one team is losing to the other team at some point in the game, but there has been a period up until that point when one team's score was higher than the other team's, and the difference in score during that period was greater than four points." This corresponds to the fact that, for a team that is trailing at the time, one team was leading by a score difference greater than the maximum score that can be obtained in a single play in baseball (four points from a grand slam), and this is a highlight of the game up until that point.

[0191] Condition 5 in Example 2 states, "When one team's score is lower than the other team's at some point in the game, but there has been a period up to that point in time when one team's score was higher than the other team's, and that period lasted for more than half the number of games played up to that point," meaning "When one team is losing to the other team at some point in the game, but there has been a period up to that point in time in which one team was ahead in points, and that period of lead (number of innings) lasted for more than half the duration of the game (the period up to the aforementioned point in the game)," and corresponds to the fact that, for a team that is behind at the point in the game, the highlight of the game up to that point is that the team has been leading for more than the majority of the period up to that point.

[0192] Condition 6 in Example 2 states, "When one team's score is lower than the other team's at some point in the game, but there has been a period up to that point when one team's score was higher than the other team's, and there were multiple periods where this was continuous and uninterrupted" means "When one team is losing to the other team at some point in the game, but there has been a period up to that point when one team's score was higher than the other team, and there were multiple periods where this was continuous and uninterrupted." This corresponds to the fact that, for a team that is trailing at the time of the game, the highlights of the game up to that point include the team making multiple comebacks during the game (making multiple comebacks from a losing position), or the team making multiple winning runs (when they were winning but were tied or overtaken, but then regained the lead at least once).

[0193] Condition 7, "When there is a player (batter or pitcher) whose points calculated based on the points table exceed the threshold value up to the middle of the game," is a condition for highlighting a player (batter or pitcher) who has performed above a certain level up to the middle of the game and introducing the details of his / her performance. Note that the determination of performance above a certain level is the same as in Example 1 (first embodiment).

[0194] 22 shows a pattern content list 55 listing the specific contents of patterns A to K of Example 2 arranged in the bottom row 13b of the judgment table 13. Like Example 1, each of patterns A to K of Example 2 also defines a guideline for the content to be included in the progress report to be generated, but since Example 2 automatically generates progress reports during a match, the content of each of patterns A to K of Example 2 is content that corresponds to the automatic generation of progress reports during a match. Specifically, pattern A of Example 2 defines a guideline of "content related to one team being in the lead in the match," and this pattern A is a guideline for including content related to one team being ahead of the other team at some point in the match.

[0195] Pattern B of Example 2 specifies the guideline, "One team is losing at some point in the game, but there was a time when they were leading the other team up until that point." This pattern B is a guideline for including in the progress report a description of the fact that one team was leading the game up until that point, as a good point of the team that is trailing at some point in the game, and the content is from the perspective of the team that is trailing the game, and specifically lists noteworthy points of the team that is trailing at a stage prior to that point.

[0196] Pattern C of Example 2 specifies in more detail the situation in which one team was leading up to a point in the game than Pattern B of Example 2 above, and prescribes the guideline that "one team was losing up to a point in the game, but at one time was leading by a score difference that was not easily recaptured." This Pattern C of Example 2 also provides a guideline for including in the progress report a description of the fact that, although one team is currently losing the game (to the other team), the losing team's good points up to that point were that the one team was leading the other team by a score difference (more than four points) that the opposing team (the other team) could not reverse, even if the other team hit a grand slam home run (even if the maximum score that can be obtained in one play in baseball was scored), and from the perspective of focusing on the losing team in the game, the content explains the situation of the losing team at a time when they were leading before that, based on the length of their lead.

[0197] Pattern D of Example 2 also defines the lead situation in more detail than Pattern B of Example 2 above, and prescribes the guideline that "one team is losing at some point in the game, but has been leading the other team for most of the period up to that point." Pattern D of Example 2 also provides a guideline for including in the progress report a description of the fact that, although one team is currently losing the game (the other team), the losing team's strengths up to that point were that the team had been leading the game for more than half the number of innings from the start of the game to the current point in the game (for example, if the current point in the game is the end of the bottom of the seventh inning, this would be when the team had been leading for more than three innings), and the content specifically explains the length of the lead that the losing team had during the game from the perspective of focusing on the losing team during the game.

[0198] Pattern E of Example 2 also defines the lead situation in more detail than Pattern B of Example 2 above, and prescribes the guideline that "one team is losing at some point in the game, but the other team has made a comeback or won again during the game." This Pattern E of Example 2 also provides a guideline for including in the progress report content that, although one team is currently losing the game (to the other team), the losing team's good points up to that point include multiple comebacks from a trailing situation up to that point in the game, or that the losing team was leading but was tied or overtaken, but then regained the lead at least once by that point in the game, and the content specifically explains the situation in which the losing team is struggling like a seesaw game from the perspective of focusing on the losing team during the game.

[0199] Pattern F of Example 2 specifies a situation that combines Patterns C and D of Example 2 above, and prescribes the guideline that "one team is losing at some point in the game, but up until that point, one team has been leading by a score difference that cannot be easily overtaken, and that lead has continued for most of the period up until that point in the game."

[0200] Furthermore, Pattern G of Example 2 specifies a situation that combines Pattern C and Pattern E of Example 2 described above, and defines the guideline as "one team is losing at some point in the game, but up until that point, one team has been leading by a score difference that cannot be easily caught up, and one team has had multiple comebacks or wins up until that point."

[0201] Furthermore, Pattern H of Example 2 specifies a situation that combines Pattern D and Pattern E of Example 2 above, and prescribes the guideline that "one team is losing at some point in the game, but has been in the lead for most of the period from the start of the game until the point in the game, and has made multiple comebacks or won by that point."

[0202] Furthermore, Pattern I of Example 2 specifies a situation that combines three patterns, namely Pattern C, Pattern D, and Pattern E of Example 2 above, and provides the following guideline: "One team is losing at some point in the game, but up until that point, one team has been leading by a score difference that cannot be easily caught up, and that lead has lasted for most of the period from the start of the game to the point in the game, and one team has made a comeback or won several times up until that point."

[0203] Pattern J of Example 2 specifies the guideline of "content related to one team being behind at some point in the game, but having a player (batter or pitcher) who has performed well on that team up until that point." Pattern J of Example 2 is a guideline for including in the progress article, when one team is trailing at some point in the game, but one of the teams has a player (batter or pitcher) who has performed above a certain level, as a good point of the team up until that point, and the content is from the perspective of the losing team, focusing on the batter or pitcher who performed well as a notable feature of the losing team.

[0204] Finally, pattern K of Example 2 specifies the guideline that "one team is losing at some point in the game, and the content is related to praising the opposing team (the other team)." This pattern K of Example 2 is a guideline for including in the progress article that the opposing team (the other team) was excellent, because when one team is trailing at some point in the game, and there is nothing particularly noteworthy about that team up to that point, or when none of the players on that team have performed well, it is difficult to mention the team that is trailing, even from the perspective of the losing team.

[0205] The description generation program P2 of the second embodiment basically performs the same processing as in the first embodiment, but the timing of the processing, that is, the point in time at which a progress report is generated during the game (for example, the end of the bottom of the third inning, the end of the bottom of the fifth inning, the end of the bottom of the seventh inning, etc.), is set in advance by an operation by an administrator or the like of the game description generation device 10. The processing related to setting the point in time at which such a progress report is generated is also possible by the programming contents of the description generation program P2, and once the point in time has been set, it can be changed as appropriate by an operation by an administrator or the like.

[0206] Fig. 23 is a flowchart showing the processing steps of a game description generation method by the game description generation device 10 (MPU 10a) based on the description generation program P2 of Example 2 (showing an example of the processing content of the game description generation method according to the present invention, which is executed by the game description generation device 10 (MPU 10a) by the description generation program P2). The flowchart of Fig. 23 shows the processing content corresponding to description generation at a point in the middle of a game, and the processing flow is basically the same as that of the flowchart shown in Fig. 7, but differs from the flowchart of Example 1 shown in Fig. 7 in that processing at step S31 is added.

[0207] Specifically, at step S30, it is determined whether or not game data distributed from the external game data distribution system 2 has been received and acquired (corresponding to step S1 in the flowchart of FIG. 7). If game data has not been received (S30: NO), the game description generation device 10 enters a state of waiting to receive (acquire) game data. Furthermore, if game data has been received (S30: YES), the description generation program P2 of the second embodiment specifies that the MPU 10a determines whether or not the most recent content of the game data acquired by the reception corresponds to a preset time point for generating progress stories (for example, a convenient time point such as the end of the top of an inning or the end of the bottom of an inning, such as the end of the bottom of the third inning, the end of the top of the fifth inning, or the end of the bottom of the seventh inning) (S31).

[0208] If the received match data does not correspond to a predetermined stopping point during the match (the point at which a progress report is generated) as the most recent content (S31: NO), the process returns to the initial step S30. If the method for transmitting match data is the first method described above (a method in which updated match data is transmitted sequentially), even if the result at step S31 is NO, the match description generation device 10 (MPU 10a) will store and accumulate the acquired match data in the storage unit 10g, RAM 10c, etc. If the method for transmitting match data is the second method described above (a method in which match data including the most recent content from the start of the match is transmitted), when the result at step S31 is NO, the acquired match data does not need to be stored, but may be stored as a backup.

[0209] On the other hand, if the acquired match data corresponds to a predetermined good stopping point during the match (the point at which a progress report is generated) as the most recent content (S31: YES), the match description generation device 10 (MPU 10a) then performs the processes of steps S32 to S35 as shown in the flowchart of Fig. 23. When performing the processes of steps S32 to S35, if the match data is transmitted by the above-mentioned first method, the processes of the subsequent steps S32 to S35 are performed based on a plurality of match data that combines the most recently acquired match data and the accumulated match data, whereas if the match data is transmitted by the second method, the processes of the subsequent steps are performed based on the most recently acquired match data.

[0210] The processing in steps S32 to S35 is basically the same as the processing in steps S2 to S5 shown in the flowchart of Fig. 7, but the processing content is such that a description is generated according to a point in time during the game. That is, in step S32, based on the game data acquired by reception (game data during the game), the game description generation device 10 (MPU 10a) judges whether each of the opposing teams corresponds to any of the cases of conditions 1 to 7 defined in the judgment table 13. In step S32, the contents shown in the condition content list 54 in Fig. 21 described above are used as conditions 1 to 7 defined in the judgment table.

[0211] The specific content of the processing at step S32 will be explained using the example of a baseball game between team XXX (batting first) and team YYY (batting second), as in the first embodiment. The MPU 10a of the game description generation device 10 basically, as in the first embodiment, provisionally identifies team XXX, batting first and one of the opposing teams included in the game data, as one of the teams and applies this to the judgment table 13, determines whether team XXX falls under any of conditions 1 to 7 specified in the judgment table 13, and identifies team YYY, the opposing team, as the other team.

[0212] Then, at this stage of S32, the game record generation device 10 (MPU10) judges the relative size of the score of team XXX and the score of team YYY based on the game data at this point in the game, and if the score of team XXX exceeds the score of team YYY, it determines that team XXX falls under condition 1 of Example 2. Also, if the score of team XXX falls below the score of team YYY at this point in the game, it means that team XXX is losing to team YYY at that point, so the game record generation device 10 (MPU10) analyzes the game data, compares the score of both teams in each inning, and determines whether team XXX falls under any of conditions 2 to 7.

[0213] At this stage of S32, similarly to the first embodiment, if it is determined that condition 3 is met, the MPU 10a will further determine whether any of conditions 4 to 6 is met. At this stage of S32, similarly to the first embodiment, the MPU 10a generates a match review article including a description focusing on team YYY, so it also determines the conditions when team YYY is one of the teams, calculates the points of each player up to that point, and compares them with thresholds to determine whether condition 7 is met or whether a record has been achieved.

[0214] Then, at step S33, the match description generation device 10 (MPU 10a) identifies, from the judgment table 13, a pattern of content corresponding to the condition judged at step S32 (patterns A to K of the second embodiment shown in the pattern content list 55 of FIG. 22).

[0215] The game record generation device 10 (MPU 10a) then automatically generates progress reports including descriptions according to the pattern identified in step S33 using the template table 17 (S34), issues an identification code for each of the automatically generated progress reports, and stores data related to the generated progress reports in the record DB 18 in association with the issued identification code (S35). As in the first embodiment, multiple types of progress reports are generated, such as those when team XXX is identified as one of the teams and those when team YYY is identified as the other team. Furthermore, it is preferable that the identification code issued for the generated progress report also includes information indicating the point in time during the game at which the progress report occurred, or that information indicating the point in time during the game at which the progress report occurred be associated with the identification code and be stored in the record DB 18 together, making it easier to search for the generated progress report data.

[0216] The progress report generated in Example 2 can be provided by various services as in Example 1, but since progress reports are generated at the midpoint of the game in Example 2, Example 2 is suitable for services that provide progress updates during the game. Even when providing a description (progress report) explaining the situation during the game, it is possible to convey content that mentions the good points of the losing team up to the midpoint from the perspective of fans who are rooting for the losing team. Below, a specific example of a progress report automatically generated using the present invention according to Example 2 (an example article about a baseball game between Team XXX and Team YYY) will be shown.

[0217] 24(a) is a score 60 for each inning showing Example 1 of the scoring situation at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and FIG. 24(b) shows Example 70 of a progress report automatically generated for Example 1 of the scoring situation at the end of the bottom of the seventh inning of the game. As shown in the score 60 in FIG. 24(a), at the end of the bottom of the seventh inning, Team XXX is leading Team YYY 3-0, and one team, Team XXX, falls under Condition 1 of Example 2 specified in the decision table 13 (see FIG. 21, when one team is winning the game at some point in the game). The example game review article 70 in FIG. 8(b) is automatically generated to include a description of Team XXX based on Pattern A of Example 2 (see FIG. 22, content related to one team winning at some point in the game).

[0218] Specifically, in the example progress article 70 shown in FIG. 24(b), a section 70a that reads, "Team XXX is leading Team YYY at the end of the bottom of the seventh inning," corresponds mainly to a description corresponding to Pattern A of Example 2. Also, in the example progress article 70, a section 70b that reads, "Team XXX got off to a good start in the first inning with a two-run home run by O with two outs and a runner on first base," and a section 70c that reads, "In the third inning, O's double added another run, extending the lead to three runs," also correspond to descriptions related to Pattern A of Example 2. In this way, the example game review article 70 in FIG. 24(b) contains content that describes key points about Team XXX's lead throughout the game, making it a suitable mid-game progress article for Team XXX fans and the like.

[0219] Figure 25(a) is a score 61 for each inning showing example 2 of the scoring situation at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 25(b) shows example progress article 71 automatically generated for example 2 of the scoring situation at the end of the bottom of the seventh inning in that game. As shown in score 61 in Figure 25(a), at the end of the bottom of the seventh inning, Team XXX is losing to Team YYY 3-4, but there was a situation where Team XXX was leading by the end of the bottom of the seventh inning (Team XXX was leading the game until the top of the third inning).

[0220] Therefore, one of the teams, Team XXX, falls under Condition 3 of Example 2 specified in the judgment table 13 (see Figure 21, when one team is losing at some point in the game, but there is a time when one team was leading up to that point), and the example progress article 71 of Figure 25(b) has been automatically generated to include a description for Team XXX based on Pattern B of Example 2 (see Figure 22, when one team is trailing at some point in the game, but there is content indicating that one team was leading up to that point).

[0221] Specifically, in the example game review article 71 shown in Figure 25(b), the following section 71a reads, "Team XXX got off to a good start in the first inning with a two-run home run by O○ with two outs and a runner on first base, and then added another run with a double by O○ in the third inning, widening the lead to three runs.", and the following section 71b reads, "Team XXX had a chance to tie the game in the seventh inning with one out and a runner on second base, but O○ hit an infield grounder and O○ struck out, and they were unable to catch up with the last hit." These mainly correspond to the description corresponding to pattern B of Example 2. In this way, the example game progress article 71 in Figure 25(b) contains content that explains the good points of Team XXX during the game, such as the fact that Team XXX was trailing at some point in the game, but that one team had been leading up to that point in the game, and the state of the offense related to that lead, making it a suitable game progress article for fans of Team XXX, etc.

[0222] Figure 26(a) is a score 62 for each inning showing example 3 of the scoring situation at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 26(b) shows an example progress article 72 automatically generated for example 3 of the scoring situation at the end of the bottom of the seventh inning in that game. As shown in the score 62 in Figure 26(a), at the end of the bottom of the seventh inning, Team XXX is trailing Team YYY 5 to 8, but at one point (until the end of the top of the fourth inning) at the end of the bottom of the seventh inning, Team XXX was leading by a score difference of 5 to 0.

[0223] Therefore, one of the teams, Team XXX, falls under Condition 4 of Example 2 specified in the decision table 13 (see Figure 21, when one team is losing at some point in the game, but up until that point, the other team had a lead by a score difference of more than 4 points), and the example match review article 72 of Figure 26(b) has been automatically generated to include a description for Team XXX based on Pattern C of Example 2 (see Figure 22, when one team is trailing at some point in the game, but at one point was leading by a score difference that could not be easily caught up).

[0224] Specifically, in the example progress article 72 shown in Figure 26(b), the following passage 72a, "Team XXX got off to a good start in the first inning with a two-run home run by Oh○ with two outs and a runner on first base, and in the third inning, Oh○ hit a three-run home run in his second consecutive at-bat, putting them in the lead at 5-0 early on," and the passage 72b, "In the sixth inning, Fuji○, the second pitcher, allowed a solo home run by △ to take the lead. They gave up two more runs in the seventh inning, falling behind 5-8, but hopes are high that Team XXX will make a comeback in the final stages of the game," mainly correspond to the passage that corresponds to pattern C. In this way, the game review article example 72 in Figure 26(b) describes Team XXX, who is trailing midway through the game, but up until that point, one team had been leading by such a large margin that even if the opposing team hit a grand slam home run, it could not catch up, and how that large lead was maintained, as well as expectations for a counterattack by Team XXX in the final stages of the game. This content includes descriptions of what was good about Team XXX during the game and hopes for future developments, making it a suitable progress article for fans of Team XXX, etc.

[0225] Figure 27(a) is a score 63 for each inning showing example 4 of the scoring situation at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 27(b) shows example progress article 73 automatically generated for example 4 of the scoring situation at the end of the bottom of the seventh inning in that game. As shown in score 63 in Figure 27(a), at the end of the bottom of the seventh inning, Team XXX is trailing Team YYY 3-4, but Team XXX was leading Team YYY from the top of the first inning until the end of the top of the sixth inning, and the period of time (number of innings) during which they were leading exceeded halfway through the entire period up to the midpoint of the game (7 innings of the game's elapsed period).

[0226] Therefore, one of the teams, Team XXX, falls under Condition 5 of Example 2 specified in the judgment table 13 (see Figure 21, when one of the teams is losing at some point in the game, but the number of times that Team XXX has been in the lead during the game exceeds half of the total number of times from the start of the game to some point in the game), and the example progress article 73 of Figure 27(b) has been automatically generated to include a description of Team XXX based on Pattern D of Example 2 (see Figure 22, when one of the teams is trailing at some point in the game, but has been leading for most of the time up to that point).

[0227] Specifically, in the example progress article 73 shown in Figure 27(b), the following passage 73a reads, "In the first inning, with two outs and a runner on first base, Team XXX got off to a good start with a two-run home run by Oh○, and in the third inning, Oh○ hit a solo home run in his second consecutive at-bat, putting them in the lead at 3-0 early on.", and the passage 73b reads, "However, in the sixth inning, Iwa○ lost control of his pitching, allowing the bases to be loaded with three walks, and then giving up a double to Saka○ that cleared the bases, bringing the score to just one run, before he was taken out of the game. However, Fuji○, who was entrusted with the game after him, was unable to stop the momentum of Team YYY's batting line, and allowed a double to Moto○ and a timely hit to Oki to turn the game around, but from here on, Team XXX was expected to fight back.", which mainly corresponds to the passage 73b of Example 2. In this way, the progress article example 73 in Figure 27(b) is suitable for fans of Team XXX because it contains content that describes the good points of Team XXX during the game and the desired future developments, such as that Team XXX was behind during the game, but that Team XXX had been leading the game for most of the time up until that point, how Team XXX had taken that lead, and the expectation of Team XXX's counterattack in the final stages of the game.

[0228] Figure 28(a) is a score 64 for each inning showing example 5 of the scoring situation at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 28(b) shows example progress article 74 automatically generated for example 5 of the scoring situation at the end of the bottom of the seventh inning in that game. As shown in the score 64 in Figure 28(a), in this game, Team XXX is losing to Team YYY 6-7 at the end of the bottom of the seventh inning, but up until that point, Team XXX has made two comebacks (in the top of the second inning and the top of the sixth inning).

[0229] Therefore, one of the teams, Team XXX, falls under Condition 6 of Example 2 specified in the judgment table 13 (see Figure 21, when one team is losing at some point in the game, but has had multiple periods of leading up to that point in the game), and the example progress article 74 of Figure 28(b) has been automatically generated to include a description of Team XXX based on Pattern E of Example 2 (see Figure 22, when one team is trailing at some point in the game, but has had multiple comebacks or wins up to that point).

[0230] Specifically, in the example progress article 74 shown in Figure 28(b), the following sections correspond to pattern E of Example 2: section 74a, which reads, "In the first inning, Team XXX's starting pitcher Iwa○ was hit with consecutive hits by Team YYY's batting line, allowing them to take a two-run lead, but immediately afterwards in the second inning, O○ hit a three-run home run and successfully turned the game around." section 34b, which reads, "Team XXX, desperate to catch up, once again managed to turn the game around with consecutive hits from Ito○ and O○ in the sixth inning, making the score 6 to 5." section 34b; and section 74c, which reads, "In the seventh inning, Fuji○, who took the mound early, walked Saka○ with two outs and then allowed O○ to hit a two-run home run to turn the game around, once again allowing Team YYY to take the lead, but we can hope for a third turnaround for Team XXX from the eighth inning onwards." In this way, the progress case 74 in Figure 28(b) is about Team XXX, which was trailing midway through the game, but which had made multiple comebacks up to that point in the game, and the situation at the time of those comebacks, as well as descriptions of the desired future developments, making it a suitable progress article for fans of Team XXX, etc.

[0231] Figure 29(a) is a score 65 for each inning showing an example of the scoring situation 6 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 29(b) shows an example progress article 75 automatically generated for the example of the scoring situation 6 at the end of the bottom of the seventh inning of that game. As shown in score 65 in Figure 29(a), in this game, Team XXX is losing to Team YYY 5 to 6 at the end of the bottom of the seventh inning, but up until that point, Team XXX had been in the lead, and that lead period (the number of innings from the top of the first to the top of the seventh inning) was more than half of the entire period up to the middle of the game (all innings from the start of the game to the seventh inning), and there was a situation where Team XXX's maximum lead was 5 points (a 5-0 lead).

[0232] Therefore, one team, Team XXX, falls under Condition 4 of Example 2 (see Figure 21, when one team is losing at some point in the game, but up until that point, the other team had a lead by a score difference of more than four points) and Condition 5 of Example 2 (when one team is losing at some point in the game, but the number of times that one team XXX has been in the lead during the game exceeds half of the total number of times up to that point in the game), which are specified in the judgment table 13. The example progress article 75 in Figure 29(b) has been automatically generated to include a description for Team XXX based on Pattern F of Example 2 (see Figure 22, when one team is trailing during the game, but has led by a score difference that cannot be easily caught up, and that this lead has continued for most of the period up to that point).

[0233] Specifically, in example progress article 75 shown in Figure 29(b), the following passage 75a, "In the first inning, with two outs and a runner on first base, Team XXX got off to a good start with a two-run home run by Oh○, and in the third inning, Oh○ hit a three-run home run in his second consecutive at-bat, putting them in a large lead of 5-0 and in the early stages of the game," and passage 75b, "However, in the seventh inning, Iwa○ lost control of his pitching, walking three batters to load the bases, and gave up a double to Saka○ that cleared the bases, bringing the score to within two runs, before being taken out of the game. However, Fuji○, who was entrusted with the game afterwards, was unable to stop the momentum of Team YYY's batting line, and gave up a double to Moto○ and a home run to Oki, turning the game around, but we can look forward to a comeback from Team XXX in the final stages," mainly correspond to the passage 75a in Example 2. In this way, the progress article example 75 in Figure 29(b) is about Team XXX, who is trailing midway through the game, but up until that point in time they had a huge lead with a score difference that the opposing team could not catch up with even if they hit a grand slam home run, and they had been leading for most of the time up to that point, the status of their offense regarding the lead, etc., and the content includes descriptions that express the desired developments for Team XXX in the future, making it a suitable progress article for fans of Team XXX, etc.

[0234] Figure 30(a) is a score 66 for each inning showing an example of the scoring situation 7 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 30(b) shows an example progress article 76 automatically generated for the example of the scoring situation 7 at the end of the bottom of the seventh inning in that game. As shown in the score 66 in Figure 30(a), at the end of the bottom of the seventh inning, Team XXX is losing to Team YYY 9 to 10, but up until that point, Team XXX has made two comebacks (in the top of the second and top of the sixth innings), and the largest score difference during Team XXX's initial lead was 5 points (7 to 2).

[0235] Therefore, one of the teams, Team XXX, falls under Condition 4 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but up until that point, the other team had a lead by a score difference of more than four points) and Condition 6 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but up until that point, there were multiple times during the game when the other team had been in the lead), which are specified in the judgment table 13. The example progress article 76 in FIG. 30(b) has been automatically generated to include a description for Team XXX based on Pattern G of Example 2 (see FIG. 22, when one team is trailing at some point in the game, but up until that point, Team XXX had a lead by a score difference that was not easy to catch up, and had made multiple comebacks or won).

[0236] Specifically, in the example progress article 76 shown in Figure 30(b), there is a section 76a that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take a two-run lead, but immediately after in the second inning, O○ hit a three-run home run and successfully turned the game around.", a section 76b (content corresponding to condition 4) that reads, "In the third inning, Ogawa hit a grand slam, and in the early stages of the game, they took a large lead of 7 to 2 and advanced in the game.", and a section 76c that reads, "However, in the sixth inning, Team XXX The statements that correspond to pattern G are mainly in 76c (content corresponding to condition 6), which reads, "With one out and a runner on second base, O○ hits his second home run of the day, turning the game around again." and in 76d, which reads, "However, in the seventh inning, Team XXX's reliever Fuji○ walks Saka○ with two outs, and then allows ○ to hit a two-run home run to take the lead, giving Team YYY the lead again. Team XXX is expected to make a further comeback in the final stages of the eighth and ninth innings."

[0237] In this way, the game review article example 76 in Figure 30(b) describes how Team XXX was trailing the opposing team (Team YYY) at one point in the game, but that up until that point, Team XXX had made multiple comebacks, that Team XXX had taken a large lead with a score difference that the opposing team could not catch up with even if they hit a grand slam home run, how the offense performed when they made a comeback or had a large lead, and expectations for Team XXX to make a comeback in the future, making it a suitable progress article for fans of Team XXX, etc.

[0238] Figure 31(a) is a score 67 for each inning showing an example of the scoring situation 8 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 31(b) shows an example progress article 77 that was automatically generated for the example of the scoring situation 8 at the end of the bottom of the seventh inning in that game. As shown in score 67 in Figure 31(a), in this game, Team XXX was trailing Team YYY 6 to 8 at the end of the bottom of the seventh inning, but up until that point Team XXX had made two comebacks (in the top of the second and the top of the fifth inning), and there were also situations in which Team XXX had been leading for more innings than the majority of innings up to that point (leading from the top of the second to the top of the fourth inning, and from the top of the fifth to the top of the seventh inning).

[0239] Therefore, one team, Team XXX, falls under Condition 5 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but the number of times Team XXX has been ahead in the game exceeds half of all innings up to that point in the game) and Condition 6 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but there have been multiple innings in which Team XXX has been ahead up to that point in the game), which are specified in the judgment table 13. The example progress article 77 in FIG. 31(b) was automatically generated to include a description for Team XXX based on Pattern H of Example 2 (see FIG. 22, when one team is behind at some point in the game, but Team XXX had been leading for most of the time up to that point (from the top of the second inning to the top of the fourth inning and from the top of the fifth inning to the top of the seventh inning), and also includes a description related to the fact that there have been multiple comebacks or wins during the game (comebacks in the top of the second inning and the top of the fifth inning)).

[0240] Specifically, in the example of the game review article 77 shown in Figure 31(b), there is a section 77a (related to conditions 5 and 6) that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take the lead by two points, but in the second inning immediately after that, O○ hit a three-run home run and they succeeded in turning the game around.", a section 77b (related to conditions 5 and 6) that reads, "However, in the fourth inning, Iwa○ lost three points from a barrage of hits, allowing them to take the lead.", and a section 77c (related to conditions 5 and 6) that reads, "However, in the fifth inning immediately after that, Team XXX The statements in 77c (related to conditions 5 and 6) that "With one out and a runner on second base, Oh○ hit his second home run of the day and turned the game around again," and the statement in 77d (related to conditions 5 and 6) that "However, in the seventh inning, Team XXX's reliever Fuji○ walked Saka○ with two outs, and then allowed ○moto to hit a two-run home run to take the lead, again losing the lead. We look forward to Team XXX making another comeback as the game draws to a close," mainly correspond to the statements in pattern H of Example 2.

[0241] In this way, the match review article example 77 in Figure 31(b) shows that Team XXX was behind the opposing team (Team YYY) at one point in the match, but that Team XXX had made multiple comebacks up to that point, that Team XXX had been leading for most of the period up to that point, the state of their attacks when they made comebacks or took the lead, and expectations for future comebacks by Team XXX, making it a suitable match review article for fans of Team XXX, etc.

[0242] Figure 32(a) is a score 68 for each inning showing an example score situation 9 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and another team, Team YYY, and Figure 32(b) shows an example progress article 78 that was automatically generated for the example score situation 9 at the end of the bottom of the seventh inning of that game. As shown in the score 68 in Figure 32(a), at the end of the bottom of the seventh inning, Team XXX was losing to Team YYY 9 to 10, but up until that point, Team XXX had made two comebacks (top of the second inning and top of the sixth inning), and the largest score difference during Team XXX's initial lead was 5 points (Team XXX led 7 to 2 from the top of the fourth inning to the top of the fifth inning), and there were times when Team XXX was leading for a total of more innings than the majority of the period from the start of the game to the end of the bottom of the seventh inning (leading from the top of the second inning to the top of the fifth inning and from the top of the sixth inning to the top of the seventh inning).

[0243] Therefore, the one team, Team XXX, falls under Condition 4 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but up until that point, the other team had a lead by a score difference of more than four points), Condition 5 of Example 2 (see FIG. 21, when one team is losing at some point in the game, but the number of times Team XXX was leading during the game exceeds half of the total number of innings up to that point in the game), and Condition 6 (see FIG. 21, when one team is losing at some point in the game, but up until that point, there were multiple innings in which Team XXX was leading due to a break in the game), all of which are specified in the judgment table 13. The example progress article 78 in FIG. 32(b) was automatically generated to include a description for Team XXX based on Pattern I of Example 2 (see FIG. 22, when one team was trailing at some point in the game, but up until that point, Team XXX was leading by a score difference that was not easy to catch up, and this lead had lasted for most of the period from the start of the game up to that point, and the team had made multiple comebacks or won).

[0244] Specifically, in the example progress article 78 shown in Figure 32(b), there is a section 78a (related to conditions 5 and 6) that reads, "In the first inning, Team XXX's starting pitcher, Iwa○, was hit by consecutive hits from Team YYY's batting line, allowing them to take a two-run lead. However, immediately after that, in the second inning, O○ hit a three-run home run and the team succeeded in turning the game around."; there is a section 78b (related to condition 4) that reads, "In the fourth inning, Ito○ hit a grand slam, widening the lead to 7 to 2, a one-sided score difference."; and there is a section 78c (related to condition 4) that reads, "However, in the fifth inning, Team XXX's second pitcher, O○, was hit by a barrage of hits, allowing Team YYY to take the lead by six runs." The statements that correspond mainly to Pattern I are 78c (a statement related to conditions 5 and 6), which states, "However, in the sixth inning, with one out and a runner on second base, Dai○ hit his second home run of the day and turned the game around again," 78d (a statement related to conditions 5 and 6), which states, "However, in the seventh inning, Team XXX's reliever Fuji○ walked Saka○ with two outs, and then allowed ○ to hit a two-run home run to take the lead, once again losing the lead. We are looking forward to a comeback by Team XXX in the final stages," 78e (a statement related to conditions 4, 5 and 6).

[0245] In this way, the progress article example 78 in Figure 32(b) shows that Team XXX was trailing the opposing team (Team YYY) at some point in the game, but that up until that point, Team XXX had made multiple comebacks, had taken a large lead with a score difference that the opposing team could not catch up with even if they hit a grand slam home run, had been in the lead for most of the period from the start of the game until the bottom of the seventh inning, the state of the offense when they made a comeback or took the lead, and expectations for future comebacks by Team XXX, making this a suitable game review article for fans of Team XXX, etc.

[0246] Figure 33(a) is a score 69 for each inning showing an example of the scoring situation 10 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 33(b) shows an example progress article 79 automatically generated for the example of the scoring situation 10 at the end of the bottom of the seventh inning of that game. As shown in the score 69 in Figure 33(a), at the end of the bottom of the seventh inning, Team XXX was losing to Team YYY 3 to 8, and up until that point, although Team XXX had scored points, they had never been in the lead, and all of Team XXX's points were due to two home runs by player O○.

[0247] Therefore, one of the teams, Team XXX, falls under Condition 2 of Example 2 (see Figure 22; when one team is trailing and has zero points or no lead during the game) and Condition 7 of Example 2 (same as Condition 7 of Example 1; when there is a player (batter or pitcher) whose points calculated from a certain formula exceed a threshold) specified in the judgment table 13, and the example progress article 79 in Figure 33(b) has been automatically generated to include a description for Team XXX based on Pattern J of Example 2 (content related to one team being trailing at a point in the game, but having a player who had performed well on one of the teams up to that point).

[0248] Specifically, in example progress article 79 shown in Figure 33(b), the following passage 79a (a passage relating to condition 7) reads, "No. 4 Dai○ is showing his spirit despite the difficult game development. He hit a solo home run in his third at-bat in the sixth inning, sparking a counterattack, and in the seventh inning, he hit two consecutive home runs in two at-bats with a walked runner on first base, giving him three RBIs that day."; and passage 79b (a passage relating to condition 2) reads, "Although the team is trailing 3 to 8 due to a large number of runs conceded in the early stages of the game, we can expect Team XXX's batting line-up to follow No. 4 Dai○'s lead in the final stages." are mainly descriptions that correspond to pattern J of Example 2. In this way, the progress article example 79 in Figure 33(b) is about Team XXX, which was trailing the opposing team (Team YYY) at the midpoint of the match, and there was nothing particularly noteworthy about it. However, there was a player on Team XXX who performed well, so the article includes a description of how that player performed and a description of the hope that Team XXX will fight back in the future. This makes it a suitable match review article for fans who want to know how the players on Team XXX were doing, even when they were losing.

[0249] Figure 34(a) is a score 80 for each inning showing an example of the scoring situation 11 at the end of the bottom of the seventh inning in a baseball game between one team, Team XXX, and the other team, Team YYY, and Figure 34(b) shows an example progress article 81 automatically generated for the example of the scoring situation 11 at the end of the bottom of the seventh inning in that game. As shown in the score 80 in Figure 34(a), at the end of the bottom of the seventh inning, Team XXX was shut out by Team YYY, 0-10, and the game content was such that it was difficult to point out any good points of Team XXX or any outstanding players.

[0250] Therefore, one of the teams, Team XXX, only meets Condition 2 of Example 2 (see Figure 22; when one team is trailing and has zero points or no lead during the match) specified in the judgment table 13, and the example progress article 81 of Figure 34(b) has been automatically generated to include a description for Team XXX based on Pattern K of Example 2 (see Figure 22; when one team is trailing at a point in the match and has content related to praising the opposing team (the other team)).

[0251] Specifically, in the example progress article 81 shown in Figure 34(b), the section 81a (a section related to condition 2) that reads, "By the seventh inning, Team XXX's batting line-up had been racking up strikeouts against Team YYY's starting pitcher, Nobuo O, who continued to pitch spectacularly, and had been limited to just three hits without even reaching second base," and the section 81b (a section related to condition 2) that reads, "I hope Team XXX will somehow pull together in the final stages to get revenge," are primarily descriptions that correspond to pattern K. In this way, the example game review article 81 in Figure 34(b) contains a section praising the opposing team because Team XXX had been defeated badly midway through the game, and there had been no notable events up to that point, and no players on Team XXX had performed well, so the progress article contains a section that reluctantly includes a section praising the opposing team, as well as a section that expects Team XXX to pull together in the future.

[0252] As described above, in the second embodiment (Example 2) of the present invention, even if the losing team is losing during the game, if the losing team had been leading up to that point, if the score difference due to that lead was so large that it could not be caught up even with the maximum score in one play, if the lead had lasted for most of the period up to that point, or if there had been multiple comebacks, etc., a progress article including descriptions explaining these details is automatically generated, so that even fans of the losing team can be provided with a description that they can enjoy.

[0253] Note that the various modifications described in the first embodiment (Example 1) can also be applied to the second embodiment (Example 2) of the present invention. For example, Example 2 can be applied to various sports other than baseball in which one player competes against another player in a scoring format, and any of Conditions 2 to 6 in Example 2 can be omitted depending on the specifications of the service to be provided.

[0254] 19 and 20 can also be applied to Example 2, but in that case, the processing at each stage changes to the processing for generating progress reports at a point in time during the match, and similar to the flowchart shown in Figure 23, processing equivalent to step S31 in the flowchart shown in Figure 23 (processing for determining whether the acquired match data corresponds to the time at which a progress report should be generated (a good point in the match)) is inserted between steps S10 and S11 shown in Figure 19. Below, the processing flow in one of the modifications of Example 2 corresponding to the flowcharts shown in Figures 19 and 20 will be explained with reference to the flowcharts shown in Figures 35 and 36.

[0255] First, as shown in the flowchart of Fig. 35, in this modified example of Example 2, during a game, the game description generation device 10 (MPU 10a) determines whether or not it has received and acquired game data transmitted from the external game data distribution system 2 (S40). If it has not acquired the data (S40: NO), it waits to receive the data. If it has acquired the data (S40: YES), it determines whether the most recent information included in the acquired game data corresponds to a convenient point in time (such as the end of the bottom of the third inning or the end of the bottom of the seventh inning) that will be the point in time for generating a progress report (whether it includes information at the time of article generation) (S41).

[0256] If the latest information contained in the acquired match data does not correspond to the time when the progress report was generated (S41: NO), the process returns to the initial step of S40. Note that the way in which the match description generation device 10 responds when the transmission method of the match data from the external match data distribution system 2 is the first method or the second method is the same as that described in the flowchart of FIG.

[0257] On the other hand, if the most recent information included in the acquired game data corresponds to the time when the progress report was generated (S41: YES), the game record generation device 10 (MPU 10a) provisionally identifies the team batting first as one team, compares the scores of one team and the other team at a point in the game based on the acquired game data, and determines whether the former score exceeds the latter score (whether one team is leading the other team) (S42). If it is determined that the score of one team exceeds the score of the other team at a point in the game (S42: YES), the game record generation device 10 determines that the one team satisfies Condition 1 of Example 2, and the MPU 10a automatically generates a progress report using a template based on the game data, including a description indicating that one team is winning (leading) at that time (S50). Then, the process proceeds to S48 of the flowchart of FIG. 36, where data related to the generated progress report is stored in the description DB 18.

[0258] If the game record generation device 10 (MPU 10a) determines that one team's score is lower than the other team's score (one team is behind) (S42: NO), it determines based on the acquired game data whether or not there was a period up to that point in time when one team's score exceeded the other team's score (S43).If the game record generation device 10 (MPU 10a) determines that there was a period up to the middle of the game when one team's score exceeded the other team's score (S43: YES), it determines whether or not that period was more than halfway through the game (S44).

[0259] If the period during which one team's score exceeded the other team's score exceeds half of the game's elapsed time up to the midpoint of the game (S44: YES), that is, when one team meets condition 5 of Example 2, the game record generation device 10 (MPU 10a) automatically generates a progress report using a template based on the game data, including a description based on the fact that one team was losing up to the midpoint of the game but had been leading the other team for most of the game up to the midpoint of the game (S54). After the progress report is generated at step S54, the data related to the generated progress report is stored at step S48 of the flowchart in FIG. 36, as described above.

[0260] Furthermore, if the period during which one team's score exceeded the other team's score does not exceed half of the elapsed time of the game (S44: NO), as shown in the flowchart of FIG. 36, the game record generation device 10 (MPU 10a) determines based on the game data whether there were multiple consecutive periods during which one team's score exceeded the other team's score up to the midpoint of the game (S45). If there were multiple consecutive periods during which one team's score exceeded the other team's score (S45: YES), the game record generation device 10 (MPU 10a) automatically generates a progress report using a template based on the game data, including a description of one team's re-overtake or re-winning record up to the midpoint of the game (S55). After generating the progress report at S55, the process proceeds to S48, where storage processing is performed in the same manner as described above.

[0261] Furthermore, if there is no continuous period in which one team's score exceeds that of the other team (S45: NO), MPU 10a determines based on the game data whether one team has ever led the other team by a score difference of more than four points (the maximum score that can be obtained in a single play) (a score difference of five points or more) up to the middle of the game (S46). If one team has led the other team by a score difference of more than four points (S46: YES), MPU 10a automatically generates a progress report using a template based on the game data, including a description of the fact that the other team was once leading the other team by a score difference that was difficult to catch up with (S56). Note that after generating the progress report at this step S56, the process proceeds to step S48, where storage processing is performed in the same manner as described above.

[0262] Furthermore, if one team is not leading by a score difference of more than four points (S46: NO), the history of each process up to that point indicates that one team is trailing the other team at some point in the game, but that one team has been leading the other team for a period up to that point, so the MPU 10a automatically generates a progress report using a template based on the game data, including a description appropriate to such a case (S47). Note that after generating the progress report at this S47 stage, the process proceeds to S48, where storage processing is performed in the same manner as above.

[0263] 35, if one team is not leading by the time the game is over (S43: NO), the MPU 10a calculates points for each player on one team based on the game data, determines whether there is a player (batter or pitcher) whose points exceed a threshold value (for example, 50 points) (S51), and if there is a player whose calculated points exceed the threshold value (S51: YES), the MPU 10a automatically generates a progress report including a description of the performance of the player who exceeded the threshold (a description of the player who performed well on the team that is losing the game) (S52). After generating the progress report at this S52 stage, the process proceeds to S48, where storage processing is performed in the same manner as above.

[0264] Furthermore, if there is no player who has exceeded the threshold for the calculated points (S51: NO), the MPU 10a determines that up until the middle of the match, one team has not done well and is being led by the other team, and that there is no player on either team who has performed well, and the MPU 10a automatically generates a progress report including a description that praises the other leading team (S53). Note that after generating the progress report at this S53 stage, the process proceeds to S48, and storage processing is performed in the same manner as above.

[0265] As described above, one variation of the second embodiment is the processing flow shown in the flowcharts of Figures 35 and 36. The judgment processes at steps S43, 44, 45, 46, and 51 can be omitted as appropriate depending on the specifications of the service to be provided, as in the case of using the judgment table 13 described above. In the above explanation, the first team was identified as one of the teams and processing was performed. However, if the first team is winning at some point in the game, the second team is identified as one of the teams and processing is performed according to the flowcharts of Figures 35 and 36. If the first team is losing at some point in the game, the second team is winning. In this case, processing is performed for the second team according to step S50, and a progress report is automatically generated to indicate that the second team is winning at some point in the game.

[0266] Furthermore, a specific modification of the second embodiment (Example 2) is to vary the scope and generation time of the progress reports. For example, the scope of the progress reports may be from the start of the game to a point in the middle of the game, or, depending on the content and specifications of the service provided, the invention of Example 2 may be applied to a range from a certain point in the middle of the game to a subsequent point in the game. That is, Example 2 can be applied not only to creating progress reports for baseball games from the start of the game to the end of the bottom of the fifth inning or from the start of the game to the end of the bottom of the seventh inning, but also to generating a first progress report from the start of the game to the end of the bottom of the third inning, a second progress report from the top of the fourth inning to the end of the bottom of the sixth inning, and a third progress report from the top of the seventh inning to the end of the bottom of the eighth inning, etc.

[0267] In such a modified example, the second progress report generated for the period (range) from the top of the fourth inning to the end of the bottom of the sixth inning and the third progress report generated for the period (range) from the top of the seventh inning to the end of the bottom of the eighth inning do not include the period from the start of the game, but even in such cases where the period from the start of the game is not included, the invention of Example 2 can generate progress reports that include descriptions according to the target period.

[0268] In this modification, the description generation program P2 of the second embodiment is configured to allow the time point for generating progress reports to be set multiple times, and when the set time point arrives, it is determined whether the acquired game data contains the content of the set time point, for example, at step S31 in the flowchart of Fig. 23 or step S41 in the flowchart of Fig. 35. If it is determined that the game data contains the content of the set time point, the processing from step S32 in the flowchart of Fig. 23 onwards or the processing from step S42 in the flowchart of Fig. 35 onwards is executed for each period (range) to be generated, and progress reports for the range to be generated are automatically generated and saved (recorded).

[0269] Condition 5 in this modified example is determined by whether one team's lead period (number of innings) exceeds half of the period to be generated (for example, the range from the top of the fourth inning to the bottom of the sixth inning). This modified example can also be applied to sports games where the game is determined by time, and it is possible to generate progress reports for each range (period), for example, from the start of the game until 20 minutes have elapsed, or from 20 minutes to 45 minutes have elapsed.

[0270] Furthermore, as in the above-described modified example, when generating progress reports at multiple game innings, it is possible for an administrator or the like of the game record generation device 10 to set the generation time points. However, instead of being set by an administrator or the like, multiple generation time points may be pre-programmed as defaults in the description generation program P2. In this case, as an example of a short interval for the specified generation time points, in the case of a baseball game, a rule for generating progress reports at the end of each inning (bottom of the inning) may be programmed. Also, for games other than baseball, when the game duration is determined by time, progress reports may be generated in accordance with the rules of the description generation program P2 at intervals such as every 5 minutes, every 10 minutes, or every 15 minutes (note that in this case, game data would also be transmitted at the above-described time intervals). Furthermore, in Example 2, the various modified examples of Example 2 described above may be used in combination as appropriate. [Industrial Applicability]

[0271] The present invention automatically generates match reviews or progress reports that include descriptions that focus on the side that lost the match or the side that is losing during the match, and can therefore be suitably used for providing services that convey information that fans of the losing or losing side want to know. [Explanation of symbols]

[0272] 1. Match Record Generation System 2. Match data distribution system 10. Match Record Generator 13 Decision Table 14, 54 Conditions List 15, 55 Pattern Contents List 16 Point Table 17 Template Tables 18. Description DB (Database) 30~39, 41 Battle Review Article Examples 70~79, 81 Progress report examples P1 OS Program P2 description generation program

Claims

1. A game description generation device acquires game data from an external device that indicates details related to the results of a scoring-type game in which multiple players from each team participate, and generates multiple types of descriptions related to the game results based on the acquired game data, a means for generating a description of whether one of the competitors has won or lost the match based on match data acquired from an external source; a detection means for detecting a player whose playing content satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; a means for generating a description of the detected player when the detecting means detects the player; A game description generation device comprising:

2. A game description generation device acquires game data from an external device that indicates details of a game that is being played during a scoring game in which multiple players from each team participate, and generates multiple types of descriptions of the game based on the acquired game data, a means for generating a description of content relating to one of the opponents being in the lead or being behind based on match data acquired from an external source; a detection means for detecting a player whose playing content satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; a means for generating a description of the detected player when the detecting means detects the player; A game description generation device comprising:

3. A points table that defines the points awarded according to the content of play; a calculation means for calculating points of players who participated in a match from the point table based on match data acquired from an external source; Equipped with The match record generation device described in claim 1 or claim 2, wherein the detection means detects players whose play content meets the condition that the play content exceeds a threshold by determining whether the points calculated by the calculation means exceed a threshold.

4. A points table that defines the points awarded according to the content of play; a calculation means for calculating points of players who participated in a match from the point table based on match data acquired from an external source; Equipped with the detection means detects a player whose play meets a condition that specifies that the points calculated by the calculation means exceed a threshold by determining whether the points exceed a threshold, 2. The game record generating device according to claim 1, wherein the threshold value for a losing team is set lower than the threshold value for a winning team.

5. A points table that defines the points awarded according to the content of play; a calculation means for calculating points of players who participated in a match from the point table based on match data acquired from an external source; Equipped with the detection means detects a player whose play meets a condition that specifies that the points calculated by the calculation means exceed a threshold by determining whether the points exceed a threshold, 3. The game record generating device according to claim 2, wherein the threshold value for the team that is behind at the midpoint of the game is set lower than the threshold value for the team that is leading.

6. A game description generation method in which a game description generation device externally acquires game data indicating details related to the results of a scoring-based game in which multiple players from each team participate, and generates multiple types of descriptions related to the game results based on the acquired game data, The game description generator comprises: generating a description of whether one of the competitors won or lost the match based on match data obtained from an external source; A step of detecting a player who satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; if a player is detected, generating a description of the detected player; A method for generating a match description, comprising:

7. A game description generation method in which a game description generation device externally acquires game data indicating details relating to the course of a game in a scoring-type game in which multiple players from each team participate, and generates multiple types of descriptions relating to the course of the game based on the acquired game data, The game description generator comprises: generating a description of content relating to one of the opponents being in the lead or being behind based on match data acquired from an external source; A step of detecting a player who satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; if a player is detected, generating a description of the detected player; A method for generating a match description, comprising:

8. A computer program for causing a computer to externally acquire game data indicating details related to the results of a scoring-based game in which multiple players from each team participate, and to generate multiple types of descriptions related to the game results based on the acquired game data, comprising: The computer, generating a description of whether one of the competitors won or lost the match based on match data obtained from an external source; A step of detecting a player who satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; if a player is detected, generating a description of the detected player; A computer program characterized by causing the computer to execute the following:

9. A computer program for causing a computer to externally acquire game data indicating details relating to the course of a game in a scoring format in which multiple players from each team participate, and to execute a process of generating multiple types of descriptions relating to the course of the game based on the acquired game data, the computer program comprising: The computer, generating a description of content relating to one of the opponents being in the lead or being behind based on match data acquired from an external source; A step of detecting a player who satisfies a condition that specifies that the playing content exceeds a threshold or a player who has achieved a record, based on match data acquired from an external source; if a player is detected, generating a description of the detected player; A computer program characterized by causing the computer to execute the following:

Images