Influence of Player Tracking Statistics on Basketball Team Success
Explore the impact of player tracking statistics on winning basketball teams, analyzing various metrics such as DIST, SPD, TCHS, PASS, AST, and more. The study delves into how these statistics affect team performance and strategies for success.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
ANALYZING THE INFLUENCE OF PLAYER TRACKING STATISTICS ON WINNING BASKETBALL TEAMS IGOR STAN IN* AND ALAN JOVI * *UNIVERSITY OF ZAGREB FACULTY OF ELECTRICAL ENGINEERING AND COMPUTING / DEPARTMENT OF ELECTRONICS, MICROELECTRONICS, COMPUTER AND INTELLIGENT SYSTEMS MIPRO 2018, OPATIJA
TOPICS Introduction What we analyze? How we do it? Results Novel statistical category Conclusion
INTRODUCTION Basketball is a dynamic sport with a lot of different kinds of events. New computer vision system for measuring position of the ball and all the players at the court 25 times/second. Many statistical categories were improved. Many new statistical categories.
WHAT WE ANALYZE? Player tracking statistics: DIST, SPD, TCHS, PASS, AST, SAST, DFGM, DFGA, DFG%, ORBC, DRBC, RBC, FG%, CFGM, CFGA, CFG%, UFGM, UFGA, UFG% and FTAST Hustle statistics: screen assists, deflections, loose balls recovered, charges drawn, contested 2PT shots, contested 3PT shots and contested shots Importance of passes and touches? Contested vs. uncontested shots? Importance of new defensive statistics?
HOW WE DO IT? Collect the data for 2016-2017 season of National Basketball Association (NBA) Make three separations: marking winner of each game as winning team, marking teams with 50+ wins in a season as a winning team and marking teams with 50+ wins in a season as a winning team but considering only their winning games. Calculate significance of difference between means of these groups with Mann- Whitney U test
RESULTS Table 1. Most significant results from first separation of data set. Mean Standard deviation p-value Winning team Losing team Winning team Losing team Statistical category FG% UFG% AST UFGM CFG% DRBC SAST CFGM SCREEN ASSISTS FTAST 0.48 0.46 24.27 18.52 0.50 59.51 5.90 22.44 10.49 2.24 0.43 0.40 20.97 16.04 0.46 53.99 4.89 20.98 9.42 1.98 0.049 0.074 5.26 4.06 0.073 9.04 2.79 4.43 4.09 1.55 0.046 0.073 4.65 3.88 0.07 9.61 2.33 4.31 3.81 1.43 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 1.79E-114 3.24E-67 2.72E-54 1.31E-49 2.10E-49 4.94E-48 3.83E-19 2.02E-15 1.99E-11 3.52E-05 Table 2. Most significant results from second separation of data set. Mean Standard deviation p-value Winning team Losing team Winning team Losing team Statistical category SCREEN ASSISTS FG% UFG% UFGM AST SAST CFG% DIST TCHS PASS 10.89 0.47 0.45 18.36 23.61 5.89 0.49 16.59 417.89 296.76 9.17 0.45 0.42 17.04 21.94 5.09 0.47 16.92 430.12 307.55 4.08 0.05 0.08 4.33 5.88 2.97 0.07 0.72 35.02 33.35 3.60 0.05 0.07 3.93 4.67 2.42 0.07 0.73 35.08 31.49 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 6.25E-17 1.99E-14 5.11E-13 1.46E-09 2.28E-09 3.89E-07 2.99E-05 1.27E-22 4.29E-10 7.51E-09
RESULTS Table 3. Most significant results from third separation of data set. Mean Standard deviation p-value Winning team Losing team Winning team Losing team Statistical category FG% UFG% AST UFGM CFG% SCREEN ASSISTS DRBC SAST DIST DFG% PASS 0.49 0.48 25.13 19.47 0.50 11.22 58.79 6.42 16.63 0.52 298.97 0.44 0.41 21.12 16.43 0.46 8.93 53.81 4.89 16.94 0.57 308.82 0.05 0.08 5.79 4.13 0.07 4.18 8.43 3.08 0.74 0.12 33.12 0.05 0.07 4.49 3.84 0.07 3.51 9.42 2.37 0.74 0.12 32.17 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 3.7E-4 6.66E-56 8.43E-38 2.52E-30 1.44E-30 9.08E-23 8.98E-19 3.40E-19 2.92E-16 2.16E-14 1.31E-11 2.70E-05
NOVEL STATISTICAL CATEGORY Effective passing ratio EPR: ??? =??? + ???? + ????? ???? Percentage of passes that ends up with assists We trained logistic regression model only on a single feature, EPR, and compare it with random model for all three separations. Model gave us accuracy of 63.3%, 62.8% and 65.7%, respectively.
CONCLUSION The most important differences are: 1. field goal percentage, especially uncontested (UFG%), but also contested (CFG%) 2. uncontested field goal made (UFGM) 3. better teams have lower distance covered 4. better teams have less passes 5. better teams have more defensive rebound chances 6. better teams have more assists and secondary assists The current results suggest that teams start considering the measure of quality of their passes (EPR)
THANK YOU! Questions?
