Athlete competition performance real-time evaluation method and system

By registering athletes and integrating data on a sports assessment platform, and combining iterative updates of predicted and real-time win rates, the problem of insufficient accuracy in evaluating athletes' competitive level has been solved. This has led to the convergence of rating scores and score deviation values, achieving a precise, objective, and dynamic evaluation of competitive performance.

CN121302129APending Publication Date: 2026-01-09SHENZHEN MINGSHI INTELLIGENT TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511493625.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-10-20
Publication Date
2026-01-09

AI Technical Summary

Technical Problem

Existing technologies lack sufficient accuracy in assessing athletes' competitive level, cannot update dynamically in real time, and rely excessively on subjective judgment.

Method used

By registering athletes on a sports assessment platform, retrieving historical matchup data and uploading it to the platform server, performing data classification and integration, importing it into the calculation engine, establishing initial values ​​for athletes, including rating scores and score deviation values, identifying matchup groups and predicting win rates, collecting data in real time for iterative updates, and making corrections by combining time decay factors and style restraint coefficients.

Benefits of technology

It has achieved gradual convergence between athlete rating scores and scoring deviation values, more realistically and accurately reflecting their competitive level, solving the problems of insufficient assessment accuracy and subjective dependence, and achieving precise, objective, and dynamic assessment of competitive performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121302129A_ABST
    Figure CN121302129A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of data mining, and provides an athlete competition performance real-time evaluation method and system. The method comprises the following steps: registering on an exercise evaluation platform, calling and uploading first pair data to a server, classifying and integrating the first pair data, and importing the first pair data into a calculation engine; the calculation engine establishes an athlete initial value in a user database according to the first pairing data; identifying a registration pair group, and predicting and determining a winning rate based on the initial value; second pairing data is collected in real time and imported into the calculation engine through the server, and the initial value is updated in combination with prediction and the real-time winning rate. The technical problems that in the prior art, athlete competitive level evaluation precision is insufficient, real-time dynamic updating cannot be achieved, and subjective judgment is excessively depended on are solved, and the purposes that the accuracy of athlete competitive level evaluation is improved by collecting pairing data in real time and conducting iterative updating in combination with the predicted winning rate and the actual result are achieved. And the deviation value of the grade score and the score of the athlete is gradually converged to the real competitive level, so that the technical effect of accurate, objective and dynamic competitive performance evaluation is realized.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of data mining technology, specifically to a method and system for real-time evaluation of athletes' competitive performance. Background Technology

[0002] In modern competitive sports, assessing an athlete's performance is a complex and challenging task. This is especially true in sports like tennis, where an athlete's skill level and performance are often influenced by various factors, such as match data, physical condition, playing style, and opponent. Therefore, traditional assessment methods, such as those relying on manual scoring or simplified rating systems, often fail to accurately reflect an athlete's true skill level or dynamically and in real-time reflect their competitive state. Existing tennis rating systems, such as the NTRP (National Tennis Rating Program), can provide athletes with an initial level classification through grading and adjustments based on match results. However, the NTRP system has certain limitations, primarily in its long update cycle, insufficient accuracy, and over-reliance on subjective judgment. Especially in dynamic match environments, how to update athletes' ratings in real-time and accurately measure their competitive level remains a problem that urgently needs to be solved. Summary of the Invention

[0003] This application provides a method and system for real-time evaluation of athletes' competitive performance, aiming to solve the technical problems in the prior art of insufficient accuracy in evaluating athletes' competitive level, inability to update dynamically in real time, and excessive reliance on subjective judgment.

[0004] The first aspect disclosed in this application provides a method for real-time evaluation of athlete performance. The method includes: registering for competition on a sports evaluation platform; retrieving first matchup data and uploading it to a platform server; performing data classification and integration and importing it into a calculation engine; establishing a first initial value for each athlete in the user database based on the first matchup data in the calculation engine's database, wherein the first initial value includes a first rating score and a first score deviation value, the score deviation value representing the uncertainty of the athlete's rating score; identifying matchup groups based on the competition registration; performing matchup prediction based on the first initial value to determine the predicted win rate; acquiring second matchup data through real-time competition data collection; importing the data into the calculation engine via the platform server; and updating the first initial value based on the predicted win rate and the real-time win rate, wherein, with iterative updates, the rating score and score deviation value converge to the athlete's competitive level. Another aspect of this application discloses a real-time athlete performance evaluation system, comprising: a data retrieval module: registering athletes on a sports evaluation platform, retrieving first matchup data and uploading it to the platform server, and performing data classification, integration, and importing into a calculation engine; an initial value establishment module: establishing a first initial value for each athlete in the user database based on the first matchup data in the database of the calculation engine, wherein the first initial value includes a first rating score and a first score deviation value, and the score deviation value characterizes the uncertainty of the athlete's rating score; a matchup prediction module: identifying matchup groups based on the sports registration, performing matchup prediction based on the first initial value, and determining the predicted win rate; and an iterative update module: acquiring second matchup data through real-time sports data collection, importing it into the calculation engine via the platform server, and updating the first initial value based on the predicted win rate and the real-time win rate, wherein, with iterative updates, the rating score and score deviation value converge to the athlete's competitive level.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages: The aforementioned real-time athlete performance evaluation method first calls upon and uploads historical match-up data, which is then categorized, integrated, and fed into a calculation engine. Within the engine, a basic evaluation is established for each athlete based on this initial data, including a rating score and a score deviation value. The deviation value reflects the uncertainty of the evaluation result. Subsequently, predictions are made for match-up groups based on these initial values, resulting in a win probability estimate. During the competition, new match-up data is collected in real time and fed into the calculation engine. The initial athlete evaluation is dynamically corrected by combining the prediction results with the actual results. With continuous iteration, the athlete's rating score and score deviation value gradually converge, more realistically and accurately reflecting their competitive level.

[0006] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating a method for real-time evaluation of athlete performance in one embodiment.

[0009] Figure 2 This is a system architecture diagram for real-time evaluation of athlete performance in one embodiment. Reference numerals: Data retrieval module 11, initial value establishment module 12, match prediction module 13, iterative update module 14. Detailed Implementation

[0010] This application provides a method and system for real-time evaluation of athletes' competitive performance, which solves the technical problems in the prior art such as insufficient accuracy in evaluating athletes' competitive level, inability to update dynamically in real time, and excessive reliance on subjective judgment.

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0012] It should be noted that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product, or device.

[0013] Example 1, as Figure 1 As shown, this application provides a method for real-time evaluation of athlete performance, the method comprising: Register for the competition on the sports assessment platform, retrieve the first match data and upload it to the platform server, and then perform data classification, integration and import into the calculation engine.

[0014] In this embodiment, athletes first complete competition registration on the sports assessment platform. The registration process includes not only inputting basic identity information but also binding their competition account and historical match records to ensure data uniqueness and traceability. After registration, the athlete's first matchup data is retrieved through an interface with the sports monitoring system. This first matchup data typically comes from past match records and covers multiple dimensions such as match time, opposing teams, score results, and shot trajectories. The retrieved data is then uniformly formatted by the platform server, including timestamp standardization and removal of outlier data. Next, the processed first matchup data is categorized and organized according to data type; for example, all match times are grouped into one set, and all shot trajectories are grouped into another. Finally, the integrated data is imported into the calculation engine. The calculation engine writes various types of information into the corresponding database forms according to preset parsing rules and data table structures to support the establishment of subsequent initial values ​​and matchup prediction calculations.

[0015] Furthermore, this application provides, before establishing the first initial value for each athlete in the user database based on the first matchup data, the following steps are included: Connect to the competitive monitoring system and retrieve the first matchup data, which is historical matchup data and includes at least hitting data, movement data, and tactical data; based on the first matchup data, perform style clustering and generate a sports profile to determine a technical style vector; store the technical style vector in the user database, wherein there is a one-to-one correspondence between the athlete and the user database.

[0016] Preferably, a connection is first established with the competitive monitoring system to retrieve the athlete's historical performance data as the first match data. This first match data includes not only basic match result information but also multi-dimensional data such as shot data, movement data, and tactical data. Shot data typically includes ball speed, spin, hitting point position, and the type of shot chosen, such as whether it is an offensive or defensive shot. Movement data includes the athlete's movement speed during the match, court coverage, number of sudden stops and starts, and the smoothness of recovery footwork. Tactical data involves the frequency of net play, the success rate of serving rallies, and the conversion rate of break points. Subsequently, based on the data recorded in the first matchup, shot characteristics, movement characteristics, and tactical characteristics are extracted. In shot characteristic extraction, the percentage of offensive shots (forehand drive frequency / total number of shots) and average shot depth (distance from the baseline to the landing point) are calculated. In movement characteristic extraction, the percentage of maximum speed maintenance time, number of changes of direction per minute, and recovery footwork efficiency (number of steps required to restart after a sudden stop) are extracted. In tactical characteristic extraction, the handling of key points (such as the probability of serving and volleying at break point) and the third shot scoring rate are statistically analyzed. Then, principal component analysis (PCA) is used to project the original feature space to a low-dimensional subspace, retaining principal components whose cumulative variance contribution rate is greater than or equal to a preset value (such as 95%) to eliminate feature redundancy. Then, a feature weight matrix is ​​constructed by combining domain expert knowledge (e.g., a service point percentage weight of 0.3 and a shot depth weight of 0.25). A density threshold and a minimum sample size are set. Density clustering is then performed on the retained principal components to automatically identify three basic styles: offensive (core point density ≥ 15), defensive (core point density ≤ 5), and balanced (between the two), used to represent the athlete's athletic profile. Finally, the athlete ID, athletic profile, and last update timestamp are concatenated according to a vector template to form a technical style vector, which is then labeled with style tags such as offensive or defensive. A one-to-one correspondence is established between the generated technical style vector and the athlete's user database, and the vector is stored in the user database for subsequent use in match prediction, rating adjustments, and dynamic updates, achieving a comprehensive characterization of the athlete's competitive level and style characteristics.

[0017] Furthermore, this application provides, after determining the technical style vector, the following: Construct a style restraint matrix, wherein the style restraint matrix is ​​used for grade score correction; perform matching based on the style restraint matrix according to the technical style vector to determine the style restraint coefficient.

[0018] Optionally, after determining the style vector, a style restraint matrix is ​​also established to consider the interaction between different style vectors during rating calculation and correction. During its establishment, common style types are first classified and summarized, such as offensive, defensive, and balanced. Win-loss relationships between different styles are statistically analyzed based on historical match data. By analyzing the win rate distribution of these matchups, this style restraint matrix is ​​constructed. The rows and columns of the style restraint matrix correspond to different style types, and the values ​​in the matrix represent the degree to which one style restrains another, typically the ratio of the number of wins of one style against another to the total number of matches played against that style. For the generated style vector, the restraint degree of that style vector against other styles is matched from the style restraint matrix, and this restraint degree is defined as the style restraint coefficient. This style restraint coefficient is used to correct the original rating prediction results, making them not only dependent on the base rating difference but also reflecting the mutual restraint relationship between styles, thereby improving the accuracy and rationality of the evaluation results.

[0019] Furthermore, this application provides the introduction of a time decay factor, wherein the time decay factor is used to measure the decay of the scoring deviation value of an athlete during a competitive window; wherein the scoring deviation value is constrained within a preset range, which is bounded by the high uncertainty of new players and the low uncertainty of active players.

[0020] Optionally, to ensure that the rating deviation value more accurately reflects the uncertainty of an athlete's competitive level, a time decay factor is introduced into the calculation. This rating deviation value (RD) is initially set within a preset range based on the athlete's competition experience. This range defines the boundary conditions of uncertainty; that is, for newly registered athletes or those with a small sample size, their rating deviation value is in a higher range, indicating greater uncertainty in their rating score (default value is 70). Conversely, for active athletes who compete frequently and have accumulated a large amount of data, the rating deviation value converges to a lower range, indicating higher reliability of their rating score. This time decay factor can be dynamically adjusted based on the time zone of the competition gap within this period. The specific calculation formula for the time decay factor is as follows: ;in, This is the attenuated score deviation value. This is the current rating deviation value. The time decay factor represents the number of days since an athlete's last competition, i.e., the time zone of their competitive gap. c is a constant used to adjust the sensitivity of the decay factor, typically set to a fixed value (e.g., 15.411035007422441). By introducing the time decay factor, we can avoid over-reliance on limited data for blindly evaluating new players and prevent the ratings of players who haven't competed for a long time from being considered stable. This ensures that the calculation of ratings and deviation values ​​always balances historical data accumulation with current competitive state, making the evaluation results more scientific and reasonable.

[0021] In the database of the computing engine, based on the first matchup data, a first initial value is established for each athlete in the user database, wherein the first initial value includes a first rating score and a first rating deviation value, and the rating deviation value characterizes the uncertainty of the athlete's rating score.

[0022] In one embodiment, after the computing engine receives and stores the first matchup data in its internal database, it obtains a unique identifier for each athlete based on this data. Then, based on this unique identifier, it establishes a first initial value for each athlete. This first initial value includes a first rating score and a first rating deviation value. The first rating score is obtained by mapping discrete match results to a continuous rating score range. For example, a 3.0 rating is 2000–2100 points, and a 3.1 rating is 2100–2200 points. In tennis, each 0.5 rating is generally considered a technical step, so a higher-than-usual score difference is set for each 0.5 rating, for example, a 3.4 rating is 2400–2600 points. The first rating deviation value is used to quantify the reliability and uncertainty of the rating score. For newly registered athletes or those with fewer matches, the initial rating deviation value is set to a higher value (e.g., 70) to reflect the uncertainty of their evaluation. As the number of matches increases, the rating deviation value gradually converges to a lower range (e.g., 35), representing the stability of their rating score. Finally, the calculation engine combines the first-level score with the corresponding scoring deviation value, and writes the combination result into the user database according to each athlete's unique identifier. It establishes an initial value record that corresponds one-to-one with the athlete's identity, ensuring that there is a reliable benchmark value available for subsequent prediction and update stages, thereby realizing dynamic tracking and accurate evaluation of the athlete's competitive level.

[0023] Furthermore, this application provides initial values ​​for each athlete, including: A first initial value is determined, wherein the first initial value includes a first rating score and a second rating deviation value; based on the first match data, a rating score mapping based on player rating is performed to determine the first rating score, wherein each athlete corresponds to a first rating score; based on the first match data, the second rating deviation value is calculated using the number of matches and time series, and the second rating deviation value is corrected according to the time decay factor and the style restraint coefficient; the first rating score and the second rating deviation value are added to the first initial value.

[0024] Optionally, when establishing the initial value for each athlete, a unique identifier for each athlete is first obtained based on the first match data. Then, based on these unique identifiers, the athlete's historical match data is backtracked to extract the number of matches played, match results, and match time series. Next, based on the athlete's level, the athlete's level score range is obtained. Following the score calculation rules, the first-level score for each athlete is calculated in the order of matches, based on the extracted data. The score calculation rules can involve increasing or decreasing a predetermined score (e.g., 30) for each match. Then, based on the extracted number of matches played and match time series, the time interval between each match and the previous match is obtained. Then, based on the interval of the second match and the athlete's initial score deviation value, a time decay coefficient is used to correct the score. The corrected result is then multiplied by a style restraint coefficient to obtain the score deviation value for the second match. This process is repeated for other matches until the score deviation value for the last match is obtained, and this score deviation value is used as the second score deviation value. Finally, the first-level score and the second score deviation value are stored to form the initial value for each athlete, providing a solid foundation for subsequent predictions and updates.

[0025] Identify matchup groups based on competitive registration, perform matchup prediction based on the first initial value, and determine the predicted win rate.

[0026] In one embodiment, after athletes complete their competition registration, the corresponding matchup group is automatically identified based on the registration information; that is, the participating athletes are grouped into matchup units according to the competition schedule. After identification, the calculation engine calls the initial value established for each athlete in the user database and uses the win rate prediction formula to predict the matchup. This win rate prediction formula is as follows: ;in, The predicted winning probability for both sides; q is the match type correction coefficient. The more games there are, the larger q becomes. In a match that is decided by six games, q is usually 0.00690775528. This represents the rating difference, specifically the difference between the participating athlete's rating and the opponent's rating. Finally, the predicted win rate calculated by the engine is written into the match prediction database, providing a foundation for pre-match result prediction, match analysis, and subsequent real-time updates. This ensures that match predictions not only rely on static scores but also dynamically reflect uncertainty and individual differences.

[0027] The system collects second-match data in real-time and imports it into the calculation engine via the platform server. Based on the predicted win rate and the real-time win rate, the initial value is updated. As the system iterates, the rating score and score deviation value converge to the athlete's competitive level.

[0028] In one embodiment, after the match begins, a second pairing data is generated in real time by a front-end sensor array (scoring system, video / radar, wearable IMU, etc.). The platform server acts as a relay to synchronize the real-time collected data and import it into the calculation engine. Subsequently, the calculation engine calculates the real-time win rate based on the predicted win rate obtained before the match and the cumulative statistics of the current round. It then corrects the ratings of both players according to a preset incremental update rule and updates the rating deviation value. At the same time, it uses the time decay factor and style restraint coefficient for correction to complete the update of the first initial value. After each update, the new rating and rating deviation value are sent back to the user database as priors for the next update. When the changes in rating and rating deviation value in consecutive batches are all below the threshold, or when the correction amount obtained based on real-time data is consistent with the correction amount obtained based on the expected win rate, it is determined that the rating and rating deviation value have converged to the athlete's true competitive level, providing a reliable basis for subsequent pairing predictions and strategy formulation.

[0029] Furthermore, this application provides a method for performing match prediction based on the first initial value, determining the predicted win rate, and then including: For the matchup group, the calculation engine performs a win rate prediction under a single match to determine a first expected win rate, wherein the first expected win rate is the prediction result under any single match in any matchup group; based on the first expected win rate, Bayesian inference calculation is performed to determine a first correction score; based on the first correction score, the first initial value is updated to determine a first updated rating score.

[0030] Preferably, for the current matchup, the initial values ​​of both players in the user database are retrieved. The first expected win rate for each player in a single match is calculated using the aforementioned win rate prediction formula. This first expected win rate represents the theoretical probability of any player winning in this matchup, without considering real-time match factors. Subsequently, based on this first expected win rate, the first rating in the initial value is revised using Bayesian methods derived from observed match results. Specifically, the first expected win rate is input into the derivation formula to calculate the first revised rating. This derivation formula is as follows: ,in, This is the first corrected score; This represents the ratio of the score difference to the total score. For example, when the score is 6:3, set_diff is (6-3) / 6=0.5. , ; For the match result, a win is assigned a value of 1, and a loss a value of 0. After calculating the first corrected score, the scoring deviation value is updated according to the scoring deviation value update formula to obtain the second corrected deviation. The specific formula for updating the scoring deviation value is as follows: Then, the first initial value is replaced and updated based on the calculated first correction score and second correction deviation to obtain the first updated rating score and the second updated deviation value. This update process not only integrates the athlete's historical data, but also introduces a correction amount in the sense of probability statistics, making the value closer to the athlete's actual competitive level and realizing a dynamic update mechanism for the gradual convergence of rating scores.

[0031] Furthermore, this application provides a method for updating the first initial value based on the predicted win rate and the real-time win rate after determining the first updated rating, including: The system connects to a competition monitoring system and retrieves second matchup data. The competition monitoring system is connected to a front-end sensor array, and the second matchup data is real-time competition data. The second matchup data is uploaded to a platform server, where it is categorized, integrated, and imported into the calculation engine to determine a first actual win rate. Based on the first actual win rate, the first updated rating is corrected. Optionally, the system first connects to a reserved interface. This competition monitoring system is linked in real-time with a front-end sensor array deployed at the venue, including high-speed cameras, radar speedometers, wearable motion sensors, pressure sensors, etc., to collect real-time data of athletes during the competition, thereby obtaining second matchup data. This second matchup data includes multi-dimensional information such as real-time scores, landing point distribution, and running trajectories. Subsequently, the retrieved second matchup data is uploaded to the platform server and categorized, such as technical indicators, tactical performance, and result statistics. The categorized data is then packaged into data packets and imported into the calculation engine. After receiving this data, the calculation engine combines it with the real-time score to calculate the first actual win rate, and replaces the first expected win rate with this first actual win rate to recalculate the first updated rating, so as to achieve dynamic updates of the athlete's competitive level.

[0032] Furthermore, this application provides a method to determine that when the corrected score determined based on the expected win rate is consistent with the corrected score determined based on real-time competitive data, the rating and score deviation value converge to the athlete's competitive level.

[0033] Optionally, when the corrected score determined based on the expected win rate matches the corrected score determined based on real-time competitive data—that is, when the deviation between the two is within a predetermined threshold—it indicates that the prediction result has reached a stable match with the actual competition performance. At this point, the fluctuation range of the athlete's rating and score deviation value will significantly decrease. To avoid the influence of random errors, the condition that the difference in corrected score is lower than the threshold in multiple consecutive updates can also be used as a criterion. Once this condition is met, it is determined that the athlete's rating and score deviation value has converged, and can truly and objectively reflect their current competitive level. This status is written into the user database as a convergence marker. This ensures both the scientific nature of the evaluation and the credibility and stability of the prediction results in practice.

[0034] Furthermore, this application provides a correction to the first updated grade score, including: Establish a mapping relationship between NTRP levels and continuous level scores, wherein the mapping relationship is non-uniformly segmented; based on the mapping relationship, convert the updated first initial value from level scores to NTRP levels to determine the athlete's level.

[0035] Optionally, to ensure the assessment results are consistent with the existing tennis rating system, a mapping relationship between NTRP levels and continuous rating scores is first established. Unlike the traditional NTRP's uniform division in 0.5-level increments, this mapping relationship uses a non-uniform segmentation method. That is, continuous rating score intervals are mapped according to subtle differences in athletes' skill levels. For example, the intervals are wider in lower skill levels (e.g., 2.0–3.0) to quickly cover the differences between beginners and amateurs; while the intervals are narrower in higher skill levels (e.g., 4.0–5.0) to more accurately distinguish players of similar skill levels. This non-uniform segmentation effectively reflects the sensitivity to differences in athletic performance across different skill levels. After establishing the mapping relationship, the updated initial value is mapped to a specific NTRP level interval. For example, when an athlete's rating is 2450, the mapping relationship indicates that the score falls within the 2400–2600 range, ultimately converting it to an NTRP level of 3.4. Simultaneously, the rating deviation value can be used as supplementary information to indicate the reliability of the rating. This approach not only achieves a seamless connection between continuous rating scores and traditional NTRP ratings, but also makes the assessment results internationally applicable and comparable, allowing coaches, athletes, and event organizers to directly apply them in training, grouping, and competition arrangements.

[0036] For example, a newly registered player A with a rating of 3.8, a rating of 3100, and an RD of 60 plays against an older player B with a rating of 2.5, a rating of 1600, and an RD of 40. The final score is 3:6, calculated as follows: For player A, calculate the estimated win rate before the match. ; but, ; Calculate the change in its grade score: ; ; ; ; The same logic applies to player B. ; ; Calculate the change in its grade score: ; ; ; ; After the match, Player A's rating becomes 2950.86, corresponding to a grade of 3.7, and their RD (Real Rating) becomes 59.58. Player B's rating becomes 1667.94, corresponding to a grade of 2.6, and their RD becomes 39.87.

[0037] As can be seen, Player A lost the match against Player B, who was 1.3 levels lower, despite a pre-match prediction of over 90% win rate. This suggests that Player A's skill level may not have reached 3.8. Furthermore, as a relatively new player with a limited sample size, Player A's rating is more likely to be inaccurate. These two factors combined lead to greater fluctuations in the rating. With more matches played, Player A will eventually converge to their true skill level.

[0038] In summary, the embodiments of this application have at least the following technical effects: This embodiment first registers athletes on a sports assessment platform, retrieves first matchup data, uploads it to the platform server, and performs data classification and integration before importing it into the calculation engine. Then, in the calculation engine's database, based on the first matchup data, a first initial value is established for each athlete in the user database. This first initial value includes a first rating score and a first score deviation value, with the score deviation value representing the uncertainty of the athlete's rating score. Next, matchup groups based on the registered athletes are identified, and matchup prediction is performed based on the first initial value to determine the predicted win rate. Finally, real-time sports data is collected to obtain second matchup data, which is imported into the calculation engine via the platform server. The first initial value is updated based on the predicted win rate and the real-time win rate. With iterative updates, the rating score and score deviation value converge to the athlete's actual athletic level. These technical effects collectively solve the technical problems of insufficient accuracy in athlete athletic level assessment, inability to update dynamically in real time, and excessive reliance on subjective judgment in existing technologies. This achieves the technical effect of accurate, objective, and dynamic athletic performance assessment by iteratively updating the athlete's rating score and score deviation value through real-time collection of matchup data combined with predicted win rates and actual results, gradually converging the athlete's rating score and score deviation value to their true athletic level.

[0039] Example 2, based on the same inventive concept as the real-time evaluation method for athlete performance in the foregoing examples, such as... Figure 2As shown, this application provides a real-time athlete performance evaluation system, which includes: a data retrieval module 11: registering athletes on a sports evaluation platform, retrieving first matchup data and uploading it to the platform server, and performing data classification, integration, and importing into the calculation engine; an initial value establishment module 12: establishing a first initial value for each athlete in the user database based on the first matchup data in the database of the calculation engine, wherein the first initial value includes a first rating score and a first score deviation value, and the score deviation value characterizes the uncertainty of the athlete's rating score; a matchup prediction module 13: identifying matchup groups based on the sports registration, performing matchup prediction based on the first initial value, and determining the predicted win rate; and an iterative update module 14: acquiring second matchup data through real-time sports data collection, importing it into the calculation engine via the platform server, and updating the first initial value based on the predicted win rate and the real-time win rate, wherein, with iterative updates, the rating score and score deviation value converge to the athlete's competitive level.

[0040] Furthermore, the data retrieval module 11 is also used to perform the following method: Connect to the competitive monitoring system and retrieve the first matchup data, which is historical matchup data and includes at least hitting data, movement data, and tactical data; based on the first matchup data, perform style clustering and generate a sports profile to determine a technical style vector; store the technical style vector in the user database, wherein there is a one-to-one correspondence between the athlete and the user database.

[0041] Furthermore, the data retrieval module 11 is also used to perform the following method: Construct a style restraint matrix, wherein the style restraint matrix is ​​used for grade score correction; perform matching based on the style restraint matrix according to the technical style vector to determine the style restraint coefficient.

[0042] Furthermore, the data retrieval module 11 is also used to perform the following method: A time decay factor is introduced, which is used to measure the decay of the score deviation value during the athlete's competitive window time zone; wherein the score deviation value is constrained within a preset range, which is bounded by the high uncertainty of new players and the low uncertainty of active players.

[0043] Furthermore, the initial value establishment module 12 is also used to perform the following method: A first initial value is determined, wherein the first initial value includes a first rating score and a second rating deviation value; based on the first match data, a rating score mapping based on player rating is performed to determine the first rating score, wherein each athlete corresponds to a first rating score; based on the first match data, the second rating deviation value is calculated using the number of matches and time series, and the second rating deviation value is corrected according to the time decay factor and the style restraint coefficient; the first rating score and the second rating deviation value are added to the first initial value.

[0044] Furthermore, the iterative update module 14 is also used to perform the following method: For the matchup group, the calculation engine performs a win rate prediction under a single match to determine a first expected win rate, wherein the first expected win rate is the prediction result under any single match in any matchup group; based on the first expected win rate, Bayesian inference calculation is performed to determine a first correction score; based on the first correction score, the first initial value is updated to determine a first updated rating score.

[0045] Furthermore, the iterative update module 14 is also used to perform the following method: Connect to the competition monitoring system and retrieve the second matchup data, wherein the competition monitoring system is connected to the front-end sensor array, and the second matchup data is real-time competition data; upload the second matchup data to the platform server, perform data classification and integration and import it into the calculation engine to determine the first actual win rate; and correct the first updated rating based on the first actual win rate.

[0046] Furthermore, the iterative update module 14 is also used to perform the following method: When the corrected score determined based on the expected win rate is consistent with the corrected score determined based on real-time competitive data, the rating score and the score deviation value are considered to converge to the athlete's competitive level.

[0047] Furthermore, the iterative update module 14 is also used to perform the following method: Establish a mapping relationship between NTRP levels and continuous level scores, wherein the mapping relationship is non-uniformly segmented; based on the mapping relationship, convert the updated first initial value from level scores to NTRP levels to determine the athlete's level.

[0048] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0049] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0050] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for real-time evaluation of athlete performance, characterized in that, The method includes: Register for the competition on the sports assessment platform, retrieve the first match data and upload it to the platform server, and then perform data classification, integration and import into the calculation engine. In the database of the computing engine, based on the first matchup data, a first initial value is established for each athlete in the user database, wherein the first initial value includes a first rating score and a first rating deviation value, and the rating deviation value characterizes the uncertainty of the athlete's rating score. Identify matchup groups based on competitive registration, perform matchup prediction based on the first initial value, and determine the predicted win rate; The system collects second-match data in real-time and imports it into the calculation engine via the platform server. Based on the predicted win rate and the real-time win rate, the initial value is updated. As the system iterates, the rating score and score deviation value converge to the athlete's competitive level.

2. The real-time evaluation method for athlete performance as described in claim 1, characterized in that, Before establishing the initial values ​​for each athlete in the user database based on the first pairing data, the process includes: Connect to the competitive monitoring system and retrieve the first matchup data, which is historical record data and includes at least hitting data, movement data, and tactical data. Based on the first pairing data, style clustering and motion profile generation are performed to determine the technical style vector; The technical style vectors are stored in the user database, where there is a one-to-one correspondence between athletes and users in the database.

3. The real-time evaluation method for athlete performance as described in claim 2, characterized in that, After determining the technical style vector, the following is included: Construct a style restraint matrix, wherein the style restraint matrix is ​​used for rating score correction; Based on the stated technical style vector, perform matching based on the style restraint matrix to determine the style restraint coefficient.

4. The real-time evaluation method for athlete performance as described in claim 3, characterized in that, A time decay factor is introduced, wherein the time decay factor is used to measure the decay of the score deviation value of the athlete during the time zone of the competition gap; The scoring deviation value is constrained within a preset range, with the boundary defined by the high uncertainty of new players and the low uncertainty of active players.

5. The real-time evaluation method for athlete performance as described in claim 4, characterized in that, Establish initial values ​​for each athlete, including: Determine a first initial value, wherein the first initial value includes the deviation value between the first grade score and the second score; Based on the first matchup data, a rating mapping based on player level is performed to determine the first rating, where each athlete corresponds to a first rating. Based on the first match data, a second score deviation value is calculated using the match number and time series, and the second score deviation value is corrected based on the time decay factor and the style restraint coefficient. Add the deviation value between the first grade score and the second score to the first initial value.

6. The real-time evaluation method for athlete performance as described in claim 1, characterized in that, Based on the first initial value, perform match prediction, determine the predicted win rate, and then include: For the matchup group, the calculation engine performs a win rate prediction for a single matchup to determine a first expected win rate, wherein the first expected win rate is the prediction result for any single matchup in any matchup group; Based on the first expected win rate, Bayesian inference calculation is performed to determine the first correction score; Based on the first corrected score, the first initial value is updated to determine the first updated grade score.

7. The real-time evaluation method for athlete performance as described in claim 6, characterized in that, After determining the first updated rating, the initial value is updated based on the predicted win rate and the real-time win rate, including: Connect to the competition monitoring system and retrieve the second matchup data, wherein the competition monitoring system is connected to the front-end sensor array, and the second matchup data is real-time competition data; The second matchup data is uploaded to the platform server, the data is classified and integrated, and then imported into the calculation engine to determine the first actual win rate. The first updated rating is adjusted based on the first actual win rate.

8. The real-time evaluation method for athlete performance as described in claim 7, characterized in that, When the corrected score determined based on the expected win rate is consistent with the corrected score determined based on real-time competitive data, the rating score and the score deviation value are considered to converge to the athlete's competitive level.

9. The real-time evaluation method for athlete performance as described in claim 7, characterized in that, After correcting the first updated grade score, the following is included: Establish a mapping relationship between NTRP levels and continuous level scores, wherein the mapping relationship is non-uniformly segmented; Based on the mapping relationship, the updated first initial value is converted from a grade to an NTRP grade to determine the athlete's grade.

10. A real-time evaluation system for athlete performance, characterized in that, The system is used to execute the real-time evaluation method for athlete performance according to any one of claims 1-9, the system comprising: Data retrieval module: Register for competition on the sports assessment platform, retrieve the first match data and upload it to the platform server, and perform data classification, integration and import into the calculation engine; Initial value establishment module: In the database of the calculation engine, based on the first match data, the first initial value of each athlete is established in the user database, wherein the first initial value includes the first rating score and the first rating deviation value, and the rating deviation value represents the uncertainty of the athlete's rating score; Match Prediction Module: Identifies match groups based on competitive registration, performs match prediction based on the first initial value, and determines the predicted win rate; Iterative update module: Real-time competitive data collection for the second matchup, importing the data into the calculation engine via the platform server, and updating the initial value based on the predicted win rate and the real-time win rate. As the update progresses, the rating score and score deviation value converge to the athlete's competitive level.