About 11Stat & Methodology

How the model works

Positioning: every 11Stat output is a model confidence score, data-driven scenario, simulation result or historical backtest. No outcome is guaranteed, and words like "certain", "guaranteed" or "will win" are never used. 11Stat accepts no bets, sells no coupons and gives no betting tips.

Instead of relying on a single formula, 11Stat uses a multi-engine approach that combines several independent statistical models. Each engine views a match from a different angle and the results are merged into a probability plus a confidence score:

• Poisson & Dixon-Coles: compute expected goals (xG) and score distributions from teams' scoring and conceding rates. Dixon-Coles corrects the dependency in low-scoring outcomes.
• Elo / form model: a rating system that updates team strength and recent form over time, accounting for home advantage and momentum.
• Monte Carlo simulation: each match is simulated thousands of times to produce probability distributions for outcome, over/under and both-teams-to-score markets.
• Calibration & backtest: model probabilities are calibrated against past results; accuracy is measured with the Brier score and calibration error.

Confidence score & consensus

For each match the engines' probabilities are compared. When several independent models agree on the same outcome, data quality is high and historical calibration is good, the confidence score (0-10) rises. A high score means more model agreement — not a guarantee. Matches where models disagree or data is weak are flagged low-confidence.

Data sources & freshness

Fixtures, team form, squad/league info and market odds are pulled near-real-time from licensed sports-data providers and cached server-side. Odds come from a verified market average, not a single bookmaker, and no bookmaker brand is shown to users. Data refreshes regularly on match day; repeat same-day queries are served from cache to protect the API quota.

Backtest, paper ROI & transparency

A backtest is a leakage-free, walk-forward simulation of how the model would have performed on past matches. Paper ROI is a quality metric computed on paper without real money — it is not real profit. In the interest of honesty, 11Stat states clearly that its model does not systematically beat the closing line and that real-odds ROI can be negative. Synthetic (made-up) odds are never used to advertise ROI.

Every published model signal is timestamped at publish time with a SHA256 fingerprint (an immutable snapshot). The result is added later as a separate record. Signals captured before kickoff are labelled "Verified Publish"; results archived afterward are labelled "Archive Record", so past performance is shown honestly and is independently verifiable.

Limitations (honest disclosure)

• Football is high-variance; no model can guarantee a single match result.
• The model produces probabilities, not certainties; low-probability outcomes still happen.
• Paper/backtest metrics are past performance and do not guarantee future results.
• Matches with weak data (youth/reserve/lower leagues) are filtered or flagged low-confidence.
• 11Stat gives no financial advice, betting advice or stake recommendation.

See the FAQ for more, or use the live analysis panel on the home page.