IP Library Granted Patent US 12,611,586
Granted Patent B2
US 12,611,586 · App. 17/807,582 · Granted Apr 28, 2026

System and method for generating daily-updated rating of individual player performance in sports

Inventor: Nathan Walker (Durham, NC)
Assignee: STATS LLC
A63B71/0669A63B24/0062G06F18/24155
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,611,586
App. No.
17/807,582
Granted
Apr 28, 2026
Kind
B2
Abstract

A computing system identifies a target player. The computing system generates rookie priors for the target player based on characteristics of the target player. The computing system generates time series data points for the player based on at least one of the rookie priors and historical statistics of the target player. The computing system projects a game position of the target player based on the historical statistics of the target player. The computing system projects next game projections for the target player based on at least one of the rookie priors, the time series data points, the game position, and the historical statistics of the target player. The computing system generates a contribution of the target player to a team's production based on the next game projections.

Claims (75)

1 . A computer-implemented method for refining a player prediction model based on tracking data generated by processing video frames of a broadcast feed, the method comprising:

processing, by a tracking system, the video frames of the broadcast feed to digitally transform video frame data from the video frames into continuous two-dimensional coordinate data;

generating, by the tracking system, tracking data associated with one or more players, wherein the tracking data includes the continuous two-dimensional coordinate data and player motion data and object motion data during one or more events, and wherein the tracking system is in electronic communication with a computing system;

generating, by a pre-processing agent of the computing system, one or more game files formatted to include the video frames associated with included corresponding tracking data;

identifying, by the computing system, a target player of the one or more players, wherein the target player is associated with a plurality of attributes defining a plurality of characteristics of the target player, the plurality of characteristics extracted using tracking data from the one or more game files stored in a data store in electronic communication with the computing system;

generating, by the computing system, rookie priors for the target player using only the plurality of characteristics of the target player;

generating, by the computing system, time series data points for the target player based on the generated rookie priors of the target player;

classifying, by the computing system, a game position of the target player based on the generated rookie priors;

providing, by the computing system, the player prediction model for the target player based on the time series data points and the classifying;

generating, by the player prediction model, one or more next game projections for the target player;

generating, by the computing system, an adjustment weighting, wherein the adjustment weighting is based on a comparison of the one or more next game projections for the target player with an average statistic for the target player; and

providing, by the computing system, the adjustment weighting to the player prediction model as training data.

2 . The method of claim 1 , further comprising:

comparing, by the computing system, the one or more next game projections for the target player to actual statistics of the target player;

determining, by the computing system, that the one or more next game projections for the target player differ from the actual statistics by at least a threshold amount in one category of statistics; and

based on the determining, adjusting, by the computing system, the one or more next game projections.

3 . The method of claim 1 , wherein generating, by the computing system, the rookie priors for the target player using only the plurality of characteristics of the target player comprises:

generating an adjusted game one metric for each statistical category based at least in part on attributes of the target player, the attributes comprising one or more of a height, a weight, an age, and a draft pick number.

4 . The method of claim 1 , wherein generating, by the computing system, the time series data points for the target player based on the generated rookie priors of the target player comprises:

padding at least one of the rookie priors of the target player with league average data.

5 . The method of claim 4 , further comprising:

generating a baseline value for each statistic using a bayes filter based on the padded rookie priors.

6 . The method of claim 1 , wherein classifying, by the computing system, the game position of the target player based on the generated rookie priors of the target player comprises:

generating a position score for the target player, wherein the position score ranges from a true point guard to a true center.

7 . The method of claim 1 , wherein one or more of the rookie priors, the time series data points, the game position, and a box score of the target player is provided to the player prediction model as input.

8 . A non-transitory computer readable medium comprising one or more sequences of instructions, which, when executed by one or more processors, causes a computing system to perform operations for refining a player prediction model based on tracking data generated by processing video frames of a broadcast feed, the operations comprising:

processing, by a tracking system, the video frames of the broadcast feed to digitally transform video frame data from the video frames into continuous two-dimensional coordinate data;

generating, by the tracking system, tracking data associated with one or more players, wherein the tracking data includes the continuous two-dimensional coordinate data and player motion data and object motion data during one or more events, and wherein the tracking system is in electronic communication with a computing system;

generating, by a pre-processing agent of the computing system, one or more game files formatted to include the video frames associated with included corresponding tracking data:

identifying, by the computing system, a target player of the one or more players, wherein the target player is associated with a plurality of attributes defining a plurality of characteristics of the target player, the plurality of characteristics extracted using tracking data from the one or more game files stored in a data store in electronic communication with the computing system;

generating, by the computing system, rookie priors for the target player using only the plurality of characteristics of the target player;

generating, by the computing system, time series data points for the target player based on the generated rookie priors of the target player;

classifying, by the computing system, a game position of the target player based on the generated rookie priors;

providing, by the computing system, the player prediction model for the target player based on the time series data points and the classifying;

generating, by the player prediction model, one or more next game projections for the target player;

generating, by the computing system, an adjustment weighting, wherein the adjustment weighting is based on a comparison of the one or more next game projections for the target player with an average statistic for the target player; and

providing, by the computing system, the adjustment weighting to the player prediction model as training data.

9 . The non-transitory computer readable medium of claim 8 , the operations further comprising:

comparing, by the computing system, the one or more next game projections for the target player to actual statistics of the target player;

determining, by the computing system, that the one or more next game projections for the target player differ from the actual statistics by at least a threshold amount in one category of statistics; and

based on the determining, adjusting, by the computing system, the one or more next game projections.

10 . The non-transitory computer readable medium of claim 8 , wherein generating, by the computing system, the rookie priors for the target player using only the plurality of characteristics of the target player comprises:

generating an adjusted game one metric for each statistical category based at least in part on attributes of the target player, the attributes comprising one or more of a height, a weight, an age, and a draft pick number.

11 . The non-transitory computer readable medium of claim 8 , wherein generating, by the computing system, the time series data points for the target player based on the generated rookie priors of the target player comprises:

padding at least one of the rookie priors of the target player with league average data.

12 . The non-transitory computer readable medium of claim 11 , the operations further comprising:

generating a baseline value for each statistic using a bayes filter based on the padded rookie priors.

13 . The non-transitory computer readable medium of claim 8 , wherein classifying, by the computing system, the game position of the target player based on the generated rookie priors of the target player comprises:

generating a position score for the target player, wherein the position score ranges from a true point guard to a true center.

14 . The non-transitory computer readable medium of claim 8 , wherein one or more of the rookie priors, the time series data points, the game position, and a box score of the target player is provided to the player prediction model as input.

15 . A system, comprising:

a processor; and

a memory having programming instructions and a player prediction model stored thereon, which instructions, when executed by the processor, causes the system to perform operations for refining a player prediction model based on tracking data generated by processing video frames of a broadcast feed, the operations comprising:

processing, by a tracking system, the video frames of the broadcast feed to digitally transform video frame data from the video frames into continuous two-dimensional coordinate data;

generating, by the tracking system, tracking data associated with one or more players, wherein the tracking data includes the continuous two-dimensional coordinate data and player motion data and object motion data during one or more events, and wherein the tracking system is in electronic communication with a computing system;

generating, by a pre-processing agent of the system, one or more game files formatted to include the video frames associated with included corresponding tracking data;

identifying, by the processor, a target player of the one or more players, wherein the target player is associated with a plurality of attributes defining a plurality of characteristics of the target player, the plurality of characteristics extracted using tracking data from the one or more game files stored in a data store in electronic communication with the computing system;

generating, by the processor, rookie priors for the target player using only the plurality of characteristics of the target player;

generating, by the processor, time series data points for the target player based on the generated rookie priors of the target player;

classifying, by the processor, a game position of the target player based on the generated rookie priors;

providing, by the processor, the player prediction model for the target player based on the time series data points and the classifying;

generating, by the player prediction model, one or more next game projections for the target player;

generating, by the processor, an adjustment weighting, wherein the adjustment weighting is based on a comparison of the one or more next game projections for the target player with an average statistic for the target player; and

providing, by the processor, the adjustment weighting to the player prediction model as training data.

16 . The system of claim 15 , wherein the operations further comprise:

comparing the one or more next game projections for the target player to actual statistics of the target player;

determining that the one or more next game projections for the target player differ from the actual statistics by at least a threshold amount in one category of statistics; and

based on the determining, adjusting the one or more next game projections.

17 . The system of claim 15 , wherein generating the rookie priors for the target player using only the plurality of characteristics of the target player comprises:

generating an adjusted game one metric for each statistical category based at least in part on attributes of the target player, the attributes comprising one or more of a height, a weight, an age, and a draft pick number.

18 . The system of claim 15 , wherein generating the time series data points for the target player based on generated rookie priors of the target player comprises:

padding at least one of the rookie priors of the target player with league average data.

19 . The system of claim 18 , further comprising:

generating a baseline value for each statistic using a bayes filter based on the padded rookie priors.

20 . The system of claim 15 , wherein one or more of the rookie priors, the time series data points, the game position, and a box score of the target player is provided to the player prediction model as input.

Assignments (2)
SECURITY INTEREST Recorded Apr 14, 2026
From: STATS LLC
To: MORGAN STANLEY SENIOR FUNDING, INC., AS COLLATERAL AGENT
Reel/Frame 075390/0491 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 17, 2022
From: WALKER, NATHAN
To: STATS LLC
Reel/Frame 060240/0299 →
Continuity (2)
Provisional Application 63262665 · Oct 18, 2021
Related Publication 20230116986A1 · Apr 20, 2023
References Cited (22)
US 8622797B2 · Noonan · 2014 [cited by examiner]
US 20070279494A1 · Aman · 2007 [cited by examiner]
US 20110237317A1 · Noonan · 2011 [cited by examiner]
US 20120283858A1 · Lapadula · 2012 [cited by examiner]
US 20130282640A1 · LaPadula, III · 2013 [cited by examiner]
US 20140155178A1 · Bloodworth · 2014 [cited by applicant]
US 20150258443A1 · McAuley et al. · 2015 [cited by applicant]
US 20150297949A1 · Aman · 2015 [cited by examiner]
US 20160296843A1 · Stickel · 2016 [cited by examiner]
US 20190224556A1 · Ruiz et al. · 2019 [cited by applicant]
US 20220080263A1 · Brush · 2022 [cited by examiner]
US 20230116986A1 · Walker · 2023 [cited by examiner]
Medvedovsky, et al., “Daily Adjusted and Regressed Kalman Optimized projections,” DARKO Exploration, https://apanalytics.shinyapps.io/DARKO/, printed on Jun. 17, 2022, 8 pages. [cited by applicant]
Goldstein, “Nylon Calculus: Introducing Player Impact Plus-Minus,” Fansided.com, https://fansided.com/2018/01/11/nylon-calculus-introducing-player-impact-plus-minus/, Jan. 11, 2018, 10 pages. [cited by applicant]
Bball Index, “Lebron Database,” The Alchemists, https://www.bball-index.com/lebron-database/, printed Jun. 17, 2022, 4 pages. [cited by applicant]
Dunks&Threes, “2022 NBA Player Stats,” Snarr Data Science LLC, https://dunksandthrees.com/epm, 2019, 2 pages. [cited by applicant]
FiveThirtyEight, “The Best NBA Players, Accoridng to RAPTOR,” ABC News Internet Ventures, https://projects.fivethirtyeight.com/nba-player-ratings/, Jun. 17, 2022, 10 pages. [cited by applicant]
Basketball Reference, “2020-21 NBA Player Stats: Advanced,” The Sports Reference, https://www.basketball-reference.com/leagues/NBA_2021_advanced.html, printed Jun. 17, 2022, 28 pages. [cited by applicant]
ESPN, “NBA Real Plus-Minus—2021-22,” ESPN.com, http://www.espn.com/nba/statistics/rpm, Apr. 10, 2022, 7 pages. [cited by applicant]
PCT International Application No. PCT/US22/77963, International Search Report and Written Opinion of the International Searching Authority, dated Jan. 3, 2023, 8 pages. [cited by applicant]
Miller, A. et al., “Factorized Point Process Intensities: A Spatial Analysis of Professional Basketball”, arxiv.org, Cornell University Library, 201 Olin Library Cornell University Ithaca, NY 14853, Jan. 5, 2014, total … [cited by applicant]
Yue, Y. et al.: “Learning Fine-Grained Spatial Models for Dynamic Sports Play Prediction”, 2014 IEEE International Conference on Data Mining, Dec. 1, 2014, pp. 670-679, XP055639137. [cited by applicant]