| Team | Played | Scored | Conceded |
|---|---|---|---|
Step 1 — Full Match Stats: Enter team names, games played (6–15 recommended), total goals scored & conceded.
Step 2 — 1st Half Stats (Optional): Tap the [＋ 1st Half Stats] button to enter half-time goals. 2nd half values show automatically (read-only). The 1st Half tab becomes visible after calculating.
Step 3 — League Averages: Defaults: Home 1.50, Away 1.20. Toggle Advanced to override for specific leagues.
Step 4 — Calculate: Hit ⚡ Calculate. Results appear across all tabs.
Step 5 — EV Finder: Enter bookmaker 1X2 odds in the Margin Calculator — they auto-fill the EV table below. Enter remaining market odds manually.
Enter team data and press Calculate.
Enter team data on the Inputs tab and press Calculate.
Enter team data on the Inputs tab and press Calculate.
Calculate predictions first, then enter bookmaker odds to find value bets.
OddsBreaker uses the Dixon-Coles model, which improves on standard Poisson by correcting the underestimation of low-scoring results (0–0, 1–0, 0–1, 1–1) via a τ correction factor set at 0.12. Probabilities are computed via a Bivariate Poisson distribution that accounts for the statistical correlation (ρ = 0.10) between both teams' scores.
Each team's Attack Rating = Goals Scored ÷ Games ÷ League Average. Their opponent's Defence Rating = Goals Conceded ÷ Games ÷ League Average. Then: λH = AttH × DefA × LgAvgH. This cross-multiplication is the core of Dixon-Coles.
The home/away league averages (e.g. 1.50 vs 1.20) already encode home advantage at the league level. The asymmetry between these two values is the home advantage factor. Adding a separate multiplier would double-count it.
Standard models treat both teams' scores as independent. In reality, events like red cards or tactical shifts affect both. The bivariate correction adds a ρ correlation term that slightly increases draw probability and reduces the weight of extreme scorelines. Combined with Dixon-Coles' low-score correction, this improves log-loss accuracy by roughly 2–4%.
Expected Value (EV%) = (Model Probability × Bookmaker Odds − 1) × 100. A positive EV means the bookmaker is offering above fair value — theoretically profitable over the long run. Enter 1X2 odds in the Margin Calculator and they auto-fill the table to avoid double entry.
1st half league averages are automatically set to 43% of full-match averages; 2nd half to 57%. 2nd half goals are auto-derived as Full Time minus 1st Half. The 1st Half results tab becomes visible only after you enter half-time stats.
The original Dixon-Coles paper estimated τ ≈ 0.13 on English football. Cross-league studies find 0.10–0.12 is more stable across different scoring environments. OddsBreaker uses 0.12 — it slightly underweights the low-score correction, which is safer than overweighting it on leagues with different profiles.
OddsBreaker is a statistical research tool. Predictions are probabilistic — not guarantees. Past model accuracy does not guarantee future results. Please gamble responsibly. — by Victor Korir