Top Banner
Play Predict Seiji Armstrong
10
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Play predict

Play Predict

Seiji Armstrong

Page 2: Play predict

15 players on a team

5 players on the court

The coach rotates players

between the bench and the court

Data-driven approach:

real-time performance prediction

Performance predictions

Page 3: Play predict

Game play has patterns

Positive: shots, assists, steals… Negative: turnovers, fouls…

Plays come in streaks

Page 4: Play predict

33

4

5

4 3 5

Plays come in streaks

Positive: shots, assists, steals… Negative: turnovers, fouls…

Game play has patterns

Page 5: Play predict

33

4

5

4 3 5

Plays come in streaks

Positive: shots, assists, steals… Negative: turnovers, fouls…

Game play has patterns

3 pointer

2 pointer

Page 6: Play predict

Play Predict :: algorithm

Basketballvalue.com

API Bulk download Scrape

2500 game events per player

Probability density

functionsConditional probabilities

Performance predictions

Unique streak

profilesClassifying

streaks

Complete game logs

Page 7: Play predict

Play Predict :: validation

Training set

Test

set

0.90.2

0.2

0.9

Validation set:

Played over 75% of

games (89 players).

Testing predictions

to continue streak

from 2 to 3.

Pearson R = 0.7

Page 8: Play predict

Seiji Armstrong

Page 9: Play predict
Page 10: Play predict