Amusement park systems

Abstract

To provide a system for amusement parks that realizes highly accurate operation predictions. [Solution] In an amusement facility system, amusement facilities are divided into groups according to predetermined conditions, and predicted data is output by inputting operation information into a trained model that includes operation information divided by group. This makes it possible to predict operation information with high accuracy even after a predetermined period of time has passed since the installation of amusement machines.

Description

[Technical Field]

[0001] The present invention relates to a system for an amusement park. [Background technology]

[0002] Conventionally, there have been provided management systems for gaming parlors that predict the operation of newly introduced machines after a certain period of time has passed since their introduction (see, for example, Patent Document 1). In such management systems, a calculation formula suited to each gaming parlor is obtained to predict the future operation (operation) status of the new machine. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2017-176586 Summary of the Invention [Problem to be solved by the invention]

[0004] Now, even if the operating conditions of gaming machines are similar immediately after introduction, there are many gaming machines whose operating conditions deviate from those of the gaming machines after a certain period of time has passed since introduction. In this regard, the above-mentioned management system uses multiple regression analysis to predict the number of outs that will decrease with the number of days installed based on the initial characteristic values ​​immediately after installation and the operating characteristic values ​​after a certain period of time has passed since installation.However, it cannot be said that this takes into account factors that may cause deviations in the operating status after a certain period of time has passed since installation, and there is a demand for even more accurate operation predictions. On the other hand, the decline in operation of gaming machines such as those described above tends to be gradual or sudden depending on the gaming facility in which they are installed, and it is possible that this tendency is influencing the discrepancy factors mentioned above.

[0005] The present invention has been made in view of the above circumstances, and its object is to provide an amusement arcade system that realizes highly accurate operation prediction. [Means for solving the problem]

[0006] The game center system of the present invention comprises: A gaming information collecting means (e.g., a management server 10) that collects gaming information (e.g., various daily gaming information of each store A to C in FIG. 2) including operation information (e.g., out) specified based on either an operation input (e.g., an input on a keyboard 6b operated by the management device 6) or a gaming signal output from a gaming machine installed in the gaming facility from a plurality of gaming facilities (e.g., the management device 6 of each gaming facility including stores A to C); an input means (e.g., the control unit 30 of the management server 10) for inputting the operation information into a trained operation information prediction model (e.g., the trained model 32 of FIG. 2); an output means (for example, the control unit 30 of the management server 10) that outputs predicted operation information from the learned operation information prediction model; A classification means (for example, the control unit 30 of the management server 10) that groups the gaming facilities according to predetermined conditions (for example, conditions for ranking gaming facilities A to C in a predetermined order) and classifies the gaming information by group; Equipped with The learned operation information prediction model is a model constructed by including operation information (for example, operation information divided into ranks on a group basis, including data Re corresponding to the operation rates of top-ranked amusement arcades color-coded in the histogram of Figure 3 and data Bl corresponding to the operation rates of middle-ranked amusement arcades) among the game information divided by the division means as a learning target.

[0007] According to the above-mentioned configuration, gaming facilities are divided into groups according to predetermined conditions, and prediction data is output by inputting operation information into a trained model that includes operation information divided by group, so that it becomes possible to predict operation information with high accuracy after a predetermined period has elapsed since the introduction of gaming machines. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a schematic diagram showing the overall configuration of an amusement park system according to one embodiment; [Figure 2]Overview of the learning model and its input / output data [Figure 3] Histogram showing the distribution of operation rates immediately after installation [Figure 4] Diagram showing game information recorded in the management server database (part 1) [Figure 5] Diagram showing game information recorded in the management server database (part 2) [Figure 6] Diagram showing game information recorded in the management server database (part 3) [Figure 7] FIG. 10 is a diagram showing an example of predicted operation information output by a management device; DETAILED DESCRIPTION OF THE INVENTION

[0009] An embodiment of the present invention will be described below with reference to the drawings. "Store A," "Store B," and "Store C" shown in Figure 1 are schematic examples of three gaming parlors A, B, and C, and since a common gaming parlor system is constructed, the following description will focus on the configuration of store A, i.e., store A.

[0010] In the gaming facility system shown in Fig. 1, a plurality of gaming machines 1, such as pachinko machines and slot machines, are installed in gaming facility A, and a card unit 2 is attached to each gaming machine 1. Two gaming machines 1 and two card units 2 are connected to relay devices 4, and relay devices 4 are connected to a management device 6 via a LAN 5. In addition, an information display device 3 is attached above each gaming machine 1, and receives various gaming signals (game information) output from the gaming machine 1 via the card unit 2 and relay device 4, and creates and displays various display information. The relay device 4 collects information from the gaming machine 1, the card unit 2, and the information display device 3.

[0011] The management device 6 is installed in, for example, an office or the like within the amusement facility A. Although detailed illustration is omitted, the control unit of the management device 6 is mainly composed of a microcomputer having a CPU and a memory unit 6m (shown only in FIG. 2) such as a ROM and RAM. A monitor 6a and a printer (not shown) are connected to the control unit as output means, and a keyboard 6b and a mouse (not shown) operated by the amusement facility manager are also connected to the control unit.

[0012] The management device 6 receives game signals output from the gaming machines 1 side via the card unit 2 and relay device 4 in sequence, thereby managing game information for each gaming machine 1 and personal data for each registered member, as well as managing information on the gaming machines 1, the card unit 2, and the information display device 3 (stored in the memory unit 6m). Although omitted in Fig. 1, it is assumed that the management device 6 manages, for example, several hundred gaming machines 1 of various models, and that multiple gaming machine islands with different rental prices (rates by type) for gaming media are formed.

[0013] The control unit of the management device 6 (simply referred to as "management device 6") compiles, by model, the gaming information collected from multiple models of gaming machines 1. In addition, the management device 6 can display and view various gaming information (for example, predicted operation information described below) compiled and obtained by the management server 10 on the monitor 6a, as will be described in detail later.

[0014] The gaming machine 1 of this embodiment is configured as, for example, a so-called enclosed type pachinko gaming machine or a medal-less type slot machine in which gaming media are managed only by information. Specifically, the pachinko machine shown on the left side of Fig. 1 is an enclosed type pachinko gaming machine that uses circulating gaming media (for example, circulating pachinko balls) enclosed in advance inside the machine. The gaming value (gaming media) used in the gaming machine 1 is virtual pachinko balls, but it is a gaming value stored in a memory unit inside the machine, and unlike conventional payout type gaming machines, the balls shot onto the board 11 circulate and are not paid out outside the gaming machine even if a prize is won, and the gaming value (game points) in the memory unit increases or decreases depending on the game.

[0015] When the pachinko machine receives the game points from the card unit 2, it becomes possible to fire pachinko balls equal to the game points, and when the game points reach "0", it becomes impossible to fire pachinko balls. The pachinko machine transmits the game points to the card unit 2 in response to a counting operation (a touch operation on the touch panel described below) and subtracts the game points. The gaming machine 1 displays various game data (number of jackpots, number of special drawing draws, etc.) on the information display unit 13 in response to an information display operation (an operation on the same touch panel).

[0016] 1, the pachinko machine has a face 11 from which pachinko balls are shot, a launch handle 12 that constitutes the launching device, and a touch panel type information display unit 13 having a touch panel on its surface. The face 11 is provided with a liquid crystal display unit 14 and a pattern display unit (special pattern display unit, normal pattern display unit) on the display unit 14, a special pattern 1 start slot 15, a special pattern 2 start slot 16, and a big prize slot 17. In pachinko machines, a jackpot lottery is held on special prize draw 1 by winning the so-called center hole (special prize draw 1 start slot 15) (or on special prize draw 2 by winning the electric chute). One jackpot lottery is also called a start. When a jackpot is won in the jackpot lottery, the jackpot game begins (transition to jackpot state), and the large prize draw slot 17 is opened for a number of rounds (R) corresponding to the type of special prize draw. In this case, for example, the maximum number of prizes per round is 10, and when a prize is won in the large prize draw slot 17, 15 game points are added.

[0017] The slot machine shown on the right side of Figure 1 is a medalless slot machine that allows play without using actual medals. The slot machine also stores and manages game values ​​(game points) stored in its internal memory, and a game can be started on the condition that the game points have a specified number (for example, 3 points) required to play at least one game. When a player acquires points according to the type of winning combination, the points are added to the game points in the data without paying out medals, and the slot machine mainly manages the increase and decrease of the game points.

[0018] When the slot machine receives the gaming points from the card unit 2, it becomes possible to bet points equal to the gaming points. That is, as shown in FIG. 1, the slot machine has a display window 18, a start lever 19, stop buttons 20 (left stop button, center stop button, and right stop button), a touch panel type information display unit 21, a BET button 22, etc. A player can see the symbols depicted on the reels provided inside through the display window 18. By operating the BET button 22, the player bets a predetermined number of credit medals (betting the aforementioned three points). When the start lever 19 is operated in this state, an internal lottery is executed and the variation of the symbols is started (the game is started), and by operating the stop button 20, the variation of the reels is stopped by so-called pull-in control.

[0019] As is well known, slot machines have bonus roles (BB and RB roles) in addition to small roles and replay roles. During the internal lottery, if a player operates the stop button 20 while an internally winning role has been obtained, a win occurs when the symbols corresponding to the internally winning role are stopped and displayed on a pre-set pay line. When a win occurs, game points are awarded according to the winning role. For example, there are 3-point winning roles, 5-point winning roles, 10-point winning roles, and 15-point winning roles. If the winning role is a 15-point winning role, 15 game points are awarded. If the winning role is a bonus role (BB or RB), the machine transitions to a bonus state (BB state or RB state). The BB state and RB state are game states (jackpot states) that are more advantageous to the player than the normal state. Since the number of game points awarded for small roles and the probability of winning are improved, game points can be earned all at once. In addition, like the information display unit 13 of a pachinko machine, the information display unit 21 of a slot machine is capable of displaying not only counting buttons for the counting operation, but also game points and various operating data (number of games played, number of bonuses generated, etc.).

[0020] From the gaming machine 1 such as the pachinko machine or slot machine, various gaming signals including the following gaming signals are transmitted (output) to the card unit 2. ·Out signal: A signal indicating the number of points used in the game. Safe signal: A signal indicating the number of points obtained as a result of the game. · Jackpot signal: A signal that is output when a jackpot occurs. As mentioned above, the gaming value acquired by the player through play, which is the number of points managed primarily on the gaming machine 1 side, is also referred to as gaming points, and the number of points managed primarily on the card unit 2 side is also referred to as owned points.

[0021] The above gaming signals may be a data item (for example, "out" = a data item indicating the consumed (used, put) value (out) since the previous data transmission in units of one point), or a data signal summarizing multiple pieces of gaming information in a telegram. In this case, data may be aggregated in units of a predetermined period (for example, 200 ms), and the aggregated data and a data signal in a telegram that can identify the status at that time may be transmitted (output). In addition to the data signal, gaming signals may of course include pulse signals, status signals, etc. Furthermore, various gaming information is transmitted from the gaming machine 1 to the card unit 2 side (relay device 4 side) as a signal output in accordance with the execution of a game using gaming points, and various signals may be output depending on the model, not just the above gaming signals.

[0022] As shown in FIG. 1, there are two types of card units 2: one for pachinko machines that is placed on the left side of the pachinko machine, and one for slot machines that is placed on the right side of the slot machine. For ease of explanation, the same reference numerals will be used to designate components that are common to both types, and they will be described together. A card unit 2 for a pachinko machine or slot machine (hereinafter simply referred to as "card unit 2") has a bill insertion slot 23 into which bills (1,000 yen) are inserted, a touch panel type LCD display unit 24 that accepts operation inputs from the player and displays various game data, a loan button 25 for performing loan operations, a return button 26 for performing return operations, a replay button 27 for performing replay operations, and a card insertion slot 28 into which an IC card (not shown) is inserted.

[0023] The card unit 2 operates as follows. (1) The device receives the various gaming signals described above from the gaming machine 1, tallying and storing them, and transmitting them to the management device 6 and the information display device 3 via the relay device 4. The device also transmits the player's deposit balance and gaming points to the management device 6. The device receives and stores various setting information (gaming machine number, model name, score unit cost, etc.) from the management device 6. The model name is a name that can identify various models, and is expressed as "AAA, BBB, ..." for pachinko machines (see Figures 4, 6, etc.), and is managed in association with the gaming information identified in the management device 6.

[0024] (2) When a banknote is inserted into the banknote insertion slot 23, it is added to the deposit balance and stored. When an IC card is inserted into the card insertion slot 28, it reads out and displays the deposit balance and the number of points associated with the card ID. For example, in the pachinko machine card unit 2, when the loan button 25 is operated, it converts points corresponding to a certain amount (for example, 1,000 yen) into game points in increments of 250 points and sends them to the gaming machine 1. When the replay button 27 is operated, it converts the number of points into game points in increments of 250 points and sends them to the gaming machine 1. It stores and displays the number of points, game points, and total number of points (the sum of the number of points and game points). It receives point addition information and point subtraction information from the gaming machine 1, updates the stored number of points according to the received point addition information and point subtraction information, and sends point number information to the relay device 4 according to the update of the number of points.

[0025] (3) When the return button 26 is operated when there is at least one of the deposited balance and the number of points held, the deposited balance and the number of points held are recorded on the IC card and issued. When the number of points held is recorded, the type information (for example, the rate as the type) corresponding to the gaming machine 1 that acquired them is also recorded in association with it. The gaming points are not recorded on the IC card, but remain on the gaming machine 1. The number of points held is sometimes referred to as held balls. (4) Up to 10 IC cards are stocked.

[0026] The card unit 2 described above can identify various information such as currency acceptance processing, value granting processing, balance, balls held (number of points held), number of loaned balls (number of points at the time of loaning), number of converted balls, deposit amount, number of counted balls, number of loaned balls, sales amount that is the consideration for loaned balls, and recording medium acceptance and issuance processing through serial communication with the relay device 4. These may be identified by pulse signals (for example, one pulse for every 1,000 yen deposited, one pulse for every 100 yen sold, etc.). Note that communication with the gaming machine 1 may be direct communication between the gaming machine 1 and the card unit 2 without going through the relay device 4, or communication may be via the relay device 4.

[0027] The management device 6 manages the operating status of all machines installed in the above-mentioned gaming facility A and collects gaming information of multiple models of gaming machines 1. The gaming information is stored in the memory unit 6m of the management device 6 as various daily gaming information (see FIG. 2) associated with date information of business days. 1 and 2, the management device 6 can be connected to the Internet 100 (public communication line) and can communicate with the management server 10 installed in the center via the Internet 100, etc. The same configuration as that of the gaming center A is also realized in the gaming centers B and C, and the management devices 6 of the gaming centers B and C can also communicate with the management server 10 via the Internet 100, etc., just like the management device 6 of the gaming center A. Here, it is assumed that the management server 10 is operated by an information provider outside the game centers A to C, for example.

[0028] 2, in the management server 10, the control unit 30 is mainly composed of a microcomputer having a CPU and a memory unit 31m. The control unit 30 (also simply referred to as "management server 10") is equipped with communication means (transmitting / receiving units not shown) for communicating with the management devices 6 of each of the gaming centers A to C, and functions as a game information collecting means, input means, output means, and classification means.

[0029] That is, first, the management server 10 receives game information specified based on game signals output from the gaming machines 1 (as various daily game information) from the multiple gaming parlors A to C via the management device 6. In this way, the management server 10 functions as a game information collecting means that collects various daily game information from the multiple gaming parlors A to C. The various daily gaming information includes, for example, outs (operation information) and safes for each machine type and machine number for each store A to C, difference in balls, number of jackpots, operation rate per regular time, playing time for the machine, sales or gross profit, etc.

[0030] As shown in FIG. 2, the management server 10 records the received various types of daily game information in a storage unit 31m (hereinafter also referred to as "database 31m") as nationwide daily game information. The nationwide daily gaming information refers to, for example, gaming information for each machine type and type calculated from various daily gaming information, and is calculated as a national average. The nationwide daily gaming information is accumulated and recorded as daily gaming information for a sufficiently long period of time.

[0031] As will be described in detail later, the nationwide daily gaming information is stored in association with a gaming facility ID (identification information unique to gaming facilities A to C). This makes it possible to obtain the results of each gaming facility A to C from the nationwide daily gaming information, and also makes it possible to group each gaming facility A to C according to predetermined conditions described later and classify the gaming information by group. The database 31 also records gaming information (classified gaming information) obtained by classifying gaming facilities A to C into groups.

[0032] The operation information prediction model of this embodiment is a trained model 32 that inputs operation information as explanatory variables by importing and inputting nationwide daily gaming information to be predicted, for example, as a CSV file, into the control unit 30 of the management server 10, and outputs prediction results for a specific future as objective variables using a predetermined learning model. In this case, the management server 10 functions as an input means for inputting the operation information into the trained model 32 and as an output means for outputting predicted operation information from the trained model 32.

[0033] In this way, the trained model 32 is a model constructed by including operation information as a learning target among the nationwide daily gaming information shown in Figure 2, that is, a trained operation information prediction model generated by machine learning based on learning data including operation information, and is trained according to the error between the predicted operation information, which is the prediction result, and the actual operation information, which is the actual operation information.

[0034] The operation information used for machine learning is the operation information for a relatively short period after the machine is introduced into the gaming facility, for example, for seven days after the machine is introduced. The target prediction date is a predetermined period of time, for example, four weeks or eight weeks into the future, after the target gaming machine 1 is introduced into the gaming facility.

[0035] Here, the predetermined learning model is a known learning model (learning algorithm), and is appropriately selected from, for example, a linear regression model, a neural network, a decision tree, a random forest, and the like. The error between the prediction results output by the selected learning model and the actual operation information for the target day corresponding to the nationwide daily gaming information accumulated and recorded for a sufficiently long period is evaluated using known indices, such as mean squared error (MSE) or mean absolute error (MAE).The parameters of the learning model are adjusted appropriately to minimize the error obtained by this evaluation.

[0036] Incidentally, as described in the "Background Art" section, gaming parlors have traditionally performed predictions of the operation of a newly introduced model (gaming machine 1) after a certain period of time has passed since its introduction. However, even if the operating information immediately after introduction to an amusement facility, such as the national averages for outs and operating rates, is the same for gaming machines 1, there are many gaming machines 1 where a discrepancy occurs in the operating information after a certain period of time has passed since their introduction. Therefore, it is difficult to predict accurate values ​​in operation forecasts immediately after installation, and conventional operation forecasts also do not take into account such deviation factors in operation conditions, so there is a demand for even more accurate operation forecasts.

[0037] Therefore, in this embodiment, attention is focused on the characteristics that appear in the distribution of operation information when the amusement facilities A to C in which the gaming machine 1 is installed are grouped according to predetermined conditions, and the management server 10 adds these characteristics to explanatory variables to construct the trained model 32. In other words, the management server 10 uses a classification means to group gaming facilities A to C, classifies the nationwide daily gaming information by group, and has a trained model 32 constructed by incorporating the operation information (group-based operation information) from the classified nationwide daily gaming information as learning data.

[0038] Here, the predetermined conditions for grouping are, for example, conditions for ranking amusement facilities A to C in a predetermined order, and the ranking is based on operation information such as the number of people in operation and operation rate included in the nationwide daily gaming information, and gross profit information indicating the gross profit of amusement facilities A to C. The management server 10 ranks and groups the gaming facilities A to C based on such operation information, gross profit information, and predetermined conditions (as predetermined information), and classifies the operation information, gross profit information, etc. by group. Note that the gaming facilities that are the targets of grouping (subject to management) by the management server 10 are assumed to be a large number of gaming facilities including gaming facilities A to C in Fig. 1, and therefore will hereinafter also be referred to simply as "gaming facilities."

[0039] For example, if gaming parlors are divided into 10 groups based on their operating rates, the gaming parlors are sorted in ascending order according to their operating rates, and then these are divided and ranked according to predetermined conditions related to operating rates (standard values ​​for dividing into 10 levels), and the operating information for each group is classified. In this case, the groups with the highest operating rates are expressed as rank 1, rank 2, ..., rank 10.

[0040] In this regard, in gaming parlors with good operating rates, i.e., those with high ranks (for example, ranks 1 and 2) when classified into 10 levels, the operating rates of all gaming machines 1 tend to be good regardless of the popularity or performance of the gaming machines 1. On the other hand, in gaming parlors with average operating rates, that is, medium ranks (for example, ranks 3 to 6), differences in operating rates arise according to the popularity and performance of the gaming machines 1. For this reason, by dividing gaming information such as operating rates into high-rank or medium-rank groups and using them as explanatory variables, in other words, as variables corresponding to input values ​​to the trained model 32, highly accurate prediction values ​​can be obtained.

[0041] 3(a) and 3(b) are histograms showing the distribution of operation information immediately after the introduction of gaming machines 1, where even if the operation information immediately after introduction is similar, there is a discrepancy in the operation information after a certain period of time has passed since the introduction. In this histogram, the horizontal axis represents operation information such as "operation rate (%)" in 5% intervals, and the vertical axis represents the "number of data points" (e.g., the number of stores or the number of machines) in each interval.

[0042] Of these, Figure 3(a) aggregates the operation information of gaming machines 1 whose operation rates decline gradually after a predetermined period has elapsed since their introduction into gaming facilities, while Figure 3(b) aggregates the operation information of gaming machines 1 whose operation rates decline rapidly, and shows these in separate histograms (a) and (b). Such differences in the decline in operation rates will cause a discrepancy in operation rates after a predetermined period has elapsed.

[0043] Furthermore, the symbols "Re," "Bl," and "Pu" in the histogram of Figure 3 correspond to the color coding of the ranks, with red Re representing the scores (frequency) of the top-ranked groups, blue Bl representing the frequency of the middle-ranked groups, and purple Pu representing the frequency of the overlapping parts of these in each interval (operation rate). When analyzing the standard deviation in this histogram, no difference is seen in the peak position (interval) of any of the gaming machines 1 in the top ranks, but in the middle ranks, the gaming machine 1 in (a), which has a gradual decline in operation, has a peak that is further to the right than the gaming machine 1 in (b), which has a sudden decline in operation. The operation information having such characteristics, i.e., the group unit operation information including the operation rate of the gaming machine 1 in the histogram of Fig. 3, is used for learning (input to the trained model 32). Note that the operation information is not limited to the operation rate of the gaming machine 1, and the number of operating gaming machines 1 may also be used for learning.

[0044] FIG. 4 shows the game information recorded in the database 31m of the management server 10. As shown in FIG. The management server 10 manages (records) gaming information such as the "release date" for the "machine name" AAA shown in the figure, the "operation rate" of the nationwide daily gaming information, and the "operation rates" of the top and bottom ranks in groups related to the ranking of gaming facilities.

[0045] Specifically, the "model name" AAA indicates the name of the gaming machine 1, and the "release date" indicates the date when the model AAA was first introduced into the gaming facility. The Nth day operating rate ("1st day operating rate" ... "7th day operating rate") indicates the operating rate on the Nth day after the introduction of machine type AAA into the gaming facility. The Nth-day top-ranked operating rate ("first-day top-ranked operating rate" ... "seventh-day top-ranked operating rate") indicates the operating rate on the Nth day after the AAA model is introduced in the top-ranked gaming facility. The Nth-day median rank operating rate ("1st-day median rank operating rate" ... "7th-day median rank operating rate") indicates the operating rate on the Nth day after the AAA model is introduced in the median-ranked gaming facility.

[0046] In Figure 4 and onwards, the numerical value in the operating rate column is set to "0.999" for convenience, but as explained in Figure 3, for example, in a mid-ranking arcade, the operating rate will vary depending on the popularity and performance of the AAA machine, and this difference will be reflected in the actual numerical value. 4 shows only a portion of the gaming information recorded in the database 31m, and it goes without saying that the database 31m may include gaming information other than that shown in the figure. Furthermore, although the operating rate is shown as an example of the operating information, the operating number, such as outs, may also be used.

[0047] FIG. 5 shows "predicted operation information" which is the result of predicting the operation rate by the trained model 32, and which is recorded in the database 31m in addition to or separately from the gaming information of FIG. The target prediction date may be any date, and may be any one of the target days from "22nd day" to "56th day" as shown in Fig. 5, or may be a period such as 22nd to 28th day, including 4 weeks or 8 weeks later, or 50th to 56th day. Also, all days included in the above-mentioned period may be the target date, or days after the above-mentioned period may be the target date. In any case, the prediction results from the trained model 32 in the same figure will result in highly accurate predictions of operating rates by dividing the gaming information into high-rank and medium-rank as described above and using them as explanatory variables.

[0048] The database 31m of the management server 10 further stores game play information for a period sufficient for use in training the trained model 32. Specifically, as shown in Fig. 6, the prediction results are output based on the operation rate from the 1st to 7th days immediately after the introduction of model AAA, with the 28th and 56th days, which are one and two months later, being the prediction target dates, and the prediction results are compared with the actual operation information recorded in database 31m for learning. Note that while the database 31m is configured to accumulate daily game information for each model, it is also possible to accumulate game information for each gaming machine 1 by day. In this case, the information is accumulated in association with the date.

[0049] In this way, the management device 6 of each of the gaming centers A to C can output the predicted operation information predicted by the management server 10 to the monitor 6a. Fig. 7 shows an example of game information including predicted operation information displayed on the display screen of the monitor 6a. As shown in the figure, the forecast operation information is displayed together with the corresponding "model name" AAA and its "release date."

[0050] Specifically, the "introduction week operating rate" indicates the average operating rate for one week from the day that model AAA is introduced to the amusement parlor. The "predicted operating rate N weeks later" indicates the average predicted operating rate for one week N weeks later. In this case, the number of weeks later to be targeted can be arbitrarily set by the amusement parlor manager using keyboard 6b or the like, and the average predicted operating rate for the set week can be displayed.

[0051] As described above, according to the gaming center system of this embodiment, gaming centers are divided into groups based on factors that affect operation information after a predetermined period has elapsed since the introduction of gaming machine 1, and prediction data is output by inputting operation information into trained model 32 that has been trained and includes operation information divided into groups. This makes it possible to predict operation information with high accuracy even after a predetermined period has elapsed since the introduction of gaming machine 1.

[0052] Since the game arcades are ranked according to predetermined information and learning is performed using operation information classified by rank according to the ranking, it becomes possible to predict operation information with high accuracy, taking into account operation information after a predetermined period of time has elapsed. Learning is performed using outs (operating numbers) and operating rates, which affect the performance of amusement parlors, so the information obtained from predictions can be used in sales.

[0053] The present invention is not limited to the above-described embodiment, but may be modified or expanded as follows, or each modified example may be combined with the above-described embodiment, or each modified example may be combined. Although the example shows a case where gaming facilities are grouped based on gaming information, the gaming information may include the number of operating players and the number of corresponding gaming machines 1. In addition, gaming facilities may be grouped based on whether new machines are likely to be introduced, or on the operating ranking within a specific commercial area or nationwide. The number of people in operation corresponds to the average number of people (customers) in operation per hour in a particular amusement facility. The operation rate (%) can be calculated by dividing the number of people in operation by the number of machines x 100, but the calculation method for the operation rate is not limited to this.

[0054] The operation information to be learned may be the average value, median value, or standard deviation of the operation information in addition to or instead of the out or operation rate. Furthermore, by comparing game information (unit period information) for a specific gaming machine 1 in a unit period defined by a management server company or the like with reference game information, contribution information indicating the degree of contribution to the gaming facility may be output. That is, for example, the management server 10 serves as contribution information identification means and compares the unit period information of the gaming machine 1 with reference information to identify an index (contribution information) indicating the degree of contribution to the gaming facility. In this case, the index identified in the management server 10 may be calculated and output by comparing the respective prediction results, or a trained model 32 (trained operation information prediction model) including the index as a learning target may be constructed and output directly from the trained model. In either case, by configuring the system to output an index indicating the degree of contribution to the amusement facility, the amusement facility manager can objectively grasp the index on the target day and appropriately consider increasing or decreasing the number of corresponding gaming machines 1.

[0055] The target period of the operation information used for learning may be any of the exemplified periods, and may be increased or decreased as appropriate, such as 3 days or 10 days. The target prediction date may be set in any way, or all days in a predetermined period may be set as the target dates. Furthermore, multiple periods may be set as the target dates, and multiple periods may be set as the target dates so that they can be output for comparison.

[0056] An index indicating the degree of error that occurred during learning may also be output in the output from the management device 6. For example, an index such as reliability A to C may be output depending on the degree of error. The trained model 32 may be trained by the management server 10 each time, or an externally trained model may be introduced. Furthermore, game information collected after the model is introduced may be used for training, or a newly trained model may be introduced.

[0057] A cloud-based server may be used as the management server 10, and the aggregation and other processes may be performed on the cloud. The processes performed by the management server 10 may be performed by the management device 6, and vice versa. The information of the management server 10 can be viewed on the monitor 6a of the management device 6, but may also be viewed on, for example, an ordinary personal computer (a monitor of an ordinary terminal device, not shown). In this case, it is advisable to identify the viewer by entering an account ID (ID of the terminal device) or the like. Furthermore, the output of the predicted operation information and the like is not limited to display output on the monitor 6a or the like, but can be printed out on a printer or the like, and can be performed using various output means.

[0058] Although the operation information is determined based on the game information transmitted from the management device 6, the actual aggregated values ​​may be determined based on operational input using an input means such as the keyboard 6b on the management device 6. Furthermore, game information of gaming facilities that cannot transmit game information may be collected by inputting operations at other gaming facilities. Exemplary numerical ranges may be inclusive or exclusive. [Explanation of symbols]

[0059] In the drawing, 1 is a gaming machine, 10 is a management server (gaming information collection means, input means, output means, classification means, contribution information identification means), and 30 is a control unit (gaming information collection means, input means, output means, classification means, contribution information identification means).

Claims

1. a gaming information collecting means for collecting gaming information from a plurality of gaming facilities, the gaming information including operation information specified based on either an operation input or a gaming signal output from a gaming machine installed in the gaming facility; an input means for inputting the operation information into a trained operation information prediction model; an output means for outputting predicted operation information for each model from the learned operation information prediction model; a classification means for classifying the gaming facilities into groups so as to include a middle-ranking group by ranking the gaming facilities according to operation information immediately after the introduction of the gaming facilities, and classifying the gaming information by the group; Equipped with The learned operation information prediction model is a model constructed for an amusement facility system that includes as learning targets the operation information classified by the classification means, which is prior to the prediction period of the predicted operation information output by the output means, and which is for the period immediately after the gaming machine is introduced into the amusement facility, and information indicating that the gaming machine belongs to the middle group.

2. 2. The gaming facility system according to claim 1, wherein the operation information includes at least one of the number of operating gaming machines and the operation rate.

3. a contribution information specifying means for specifying contribution information indicating a degree of contribution to the gaming facility by comparing unit period information of the gaming information with reference information; 3. The game arcade system according to claim 1, wherein the output means outputs the contribution information including the contribution information identified by the contribution information identifying means.

Images