AFL 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 15.7M synthesized lines pulled from real AFL game logs.
Held out test split (3.17M rows that the model never trained on):
| accuracy | brier | log loss | baseline | lift |
|---|---|---|---|---|
| 88.3% | 0.085 | 0.284 | 0.298 | +4.7% |
Lift is how much lower the calibrated log loss is than a no-skill baseline (0.014 absolute, 4.7% relative). Aussie rules. Thin edge. The disposal lines are priced fairly well, so there is not much to beat.
git clone https://github.com/LeSingh1/edge-afl
cd edge-afl
node src/predict.js --stat "Disposals" --line 22.5 --log 25,19,28,21,24
Output is the projection, the raw over probability, the calibrated probability, and a suggested stake.
- Weight the recent games up and fit a mean and standard deviation.
- Turn the line into a raw over probability with a normal model.
- 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. - Size the bet with quarter Kelly.
The calibrator is keyed by stat type, so "Disposals" gets a different correction than other stats.
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.
MIT. Not financial advice. The house edge is real, bet responsibly.