Hand-made page by Bet Better, Melbourne. Everything here is CC BY 4.0.
Bet Better publishes free, open data on sports betting markets so that students, researchers and journalists can check the numbers themselves.
Everything is licensed CC BY 4.0. Model estimates, not betting advice. 18+.
Every betting price has a cut built in. The margin index measures that cut each day across 12 sports from the prices quoted by around 30 bookmakers. The full, daily-updated file and its codebook live at betbetter.world/studies/margin-index.
| Market type | Average bookmaker margin (%) | Daily observations |
|---|---|---|
| h2h (3way) | 7.50 | 843 |
| player_props (2way) | 7.45 | 426 |
| other (2way) | 6.65 | 269 |
| totals (2way) | 6.15 | 1426 |
| spreads (2way) | 5.35 | 1360 |
| h2h (2way) | 5.11 | 1071 |
Three-way match results and player props carry the biggest cut, and plain two-way match results are the cheapest market. The measured-versus-estimated split and the method are in the bookmaker margins study.
Data: CC BY 4.0, free to reuse with a credit to Bet Better.
The favourite-longshot bias is one of the most replicated findings in the economics of wagering, yet most of the evidence still comes from historical parimutuel horse racing. This working paper tests the calibration of contemporary online fixed-odds markets using both sides of 18,238 settled two-outcome markets (36,476 outcome observations) across five sports between November 2024 and July 2026, with the bookmaker margin removed from every quoted probability.
Two findings:
1. De-vigged market probabilities are well calibrated: in every probability decile, realised outcome frequencies fall within roughly two percentage points of the vig-free implied probabilities.
2. The residual deviation is a systematic, antisymmetric tilt: outcomes priced below 50% overperform by one to two points while favourites underperform by the same amount, and the tilt roughly doubles in player-proposition markets, the same markets that carry the highest measured margins.
Read the full paper and download the data: betbetter.world/studies/market-calibration. Working-paper DOI on Zenodo; a free PDF copy is also on Leanpub.
Author: Edward Glush, Bet Better Pty Ltd, Australia. Data CC BY 4.0.