Skip to content

LeSingh1/edge-wnba

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

edge-wnba

WNBA player prop projection model. One of the edge-models family.

Give it a player's recent game log and a betting line. It returns a calibrated probability that the player goes over, plus a quarter Kelly stake. Trained on 11.5M synthesized lines pulled from real WNBA game logs.

Results

Held out test split (2.33M rows that the model never trained on):

accuracy brier log loss baseline lift
87.7% 0.087 0.291 0.307 +5.1%

Lift is how much lower the calibrated log loss is than a no-skill baseline (0.016 absolute, 5.1% relative). Smaller league, single-game stats swing a lot. The calibration fix matters most here, it stops the raw model from overclaiming on a hot streak.

Run it

git clone https://github.com/LeSingh1/edge-wnba
cd edge-wnba
node src/predict.js --stat "Points" --line 18.5 --log 22,15,19,24,17

Output is the projection, the raw over probability, the calibrated probability, and a suggested stake.

How it works

  1. Weight the recent games up and fit a mean and standard deviation.
  2. Turn the line into a raw over probability with a normal model.
  3. Correct that probability with the trained isotonic calibrator in models/calibration.json. The correction is clamped to +/-0.20 so a thin bucket cannot fake confidence.
  4. Size the bet with quarter Kelly.

The calibrator is keyed by stat type, so "Points" gets a different correction than other stats.

What is not here

The scraping pipeline and the raw training rows live in the private app this came out of. This repo ships the trained model and the code to run it, which is the useful part.

License

MIT. Not financial advice. The house edge is real, bet responsibly.

About

WNBA player prop projection model (isotonic calibration + EV)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors