Bookmakers’ numbers, reconstructed faithfully.

Original strategy

For each home, draw, and away outcome, the 2017 implementation averaged the decimal odds offered across bookmakers, used the inverse of that average as a consensus probability, subtracted a fixed 0.05 adjustment, and tested the best available price.

A candidate exists only when the result is positive and more than three odds are available. If several outcomes qualify, the outcome with the greatest expected payoff is selected. This implementation preserves that rule under BTB-ORIGINAL-2017-CS-1.0.

Enhanced strategy

The enhanced score is deliberately separate. It removes each complete bookmaker market’s overround, uses a median-based consensus, calculates best-price edge and expected value, and adds robust median-absolute-deviation anomaly and confidence scores.

Interpretation

An odds anomaly is a price disagreement—not evidence that a match is fixed. The platform never labels fixtures as fixed or guaranteed. Source records are marked LIVE, HISTORICAL, or REPLAY.

Provenance

This GPL-3.0 rebuild acknowledges Lisandro Kaunitz, Shenjun Zhong, Javier Kreiner, the paper Beating the bookies with their own numbers — and how the online sports betting market is rigged, and the original Lisandro79/BeatTheBookie repository.