Not a match prediction, a probability analysis: read a match in numbers
Everyone who searches "match prediction" wants the same thing: how will this game end? But a one-word prediction — "home win" — doesn't answer that question, it only hides it. The honest answer to a football match isn't a name, it's a probability distribution: percentages for home / draw / away, an xG edge, form, and how many models agree. 11Stat is a football data-analytics platform, not a betting service; it shows probabilities, not advice to bet.
What you call a "prediction" is really a probability distribution
A classic match prediction is a single sentence: "home win." It sounds decisive, yet it leaves out everything you need to understand a game. A football match isn't a deterministic event — it's a probability distribution over many possible results.
- A table, not a word: "home win" hides whether the home side is a 78% favourite or a 52% coin-flip — two completely different reads.
- The draw is a real outcome: in most leagues a draw sits around 24–30%. A bare prediction ignores it entirely.
- Uncertainty is measurable: the right question isn't "who wins?" but "what is each outcome's probability, and how reliable is this read?"
That's why 11Stat never issues a prediction sentence. It turns a fixture into a probability table you can run match analysis on: home / draw / away percentages, goal expectations and uncertainty flags.
Reading a match in numbers: 1X2, xG, form, model agreement
Reading a match in numbers means looking at several complementary layers instead of one pick. The core signals on an 11Stat match card are:
- 1X2 probability split — home / draw / away percentages that sum to 100%. The full distribution, not a single label.
- xG (expected goals) edge — the chance quality each side creates, independent of the final score.
- Form index (last 5 / last 10) — the direction and stability of momentum, stripped of single-match noise.
- Model agreement (N/M) — how many independent engines point the same way. High agreement is a firmer signal; low agreement is an uncertainty flag.
- Data-quality & uncertainty flags — thin data, missing squads or high-variance games are marked openly.
The difference is simple: a prediction hands you the answer; a data readout hands you the evidence to reach it yourself.
A sample card: reading a prediction like an analyst
Here is how an analyst reads a prediction, using an anonymised sample card (not a real team). Say Team A hosts Team B:
- 1X2: Home 50% · Draw 27% · Away 23%. The home side is favoured, but nearly one match in two ends otherwise — hardly a certainty.
- xG edge: +0.4 to the home side. A modest advantage, not a dominant one.
- Form (last 5): Home 3W-1D-1L, away 2W-2D-1L. Momentum leans slightly home.
- Model agreement: 4/5 engines point the same way — reasonable consensus, not unanimity.
- Uncertainty: "medium" — a key player's status is unresolved.
An analyst's takeaway isn't "back the home win." It's: a moderate, well-supported home lean, a live draw, and no outsized signal. What's valuable isn't the pick itself, but knowing exactly how strong or weak it is.
How is this probability computed? Poisson, Dixon-Coles, Monte Carlo
So where do these percentages come from? The short answer: not one "tipster," but several statistical engines that check each other:
- Goal-rate estimate (λ): each team's expected scoring rate is derived from attack and defence strength, with home advantage added as a separate term.
- Poisson / Dixon-Coles: those rates become the probability of every scoreline (0-0, 1-0, 2-1 …). The Dixon-Coles correction adjusts low-score dependence with a rho term, because plain Poisson mis-prices 0-0 and 1-1.
- Monte Carlo simulation: the match is simulated thousands of times (typically 10,000) to derive 1X2, over/under and both-teams-to-score probabilities.
- Elo / form engine: a secondary engine tracking team strength and momentum produces an independent second opinion.
- Model consensus: the engines are combined and reported as N/M model agreement.
The last step is calibration: per league and market, we check "when we say 70%, does it really land near 70%?" You can see how that's verified on our model calibration and methodology pages.
Probabilities are not promises — and why 11Stat shows it that way
Being honest here is the most important part of the product. A probability is not a promise:
- Our accuracy and calibration figures are simulated, paper-backtest results — past performance that never guarantees or implies any future outcome.
- High probability is not certainty. Even a 75% read goes the other way in roughly one match in four — that's the nature of probability, not a model error.
- Football markets are efficient; most public information is already priced in. The goal isn't to "beat the market" but to ground your decision in transparent numbers.
11Stat is a football data-analytics platform, not a betting service; its outputs are not betting advice and carry no guarantee of results. It measures uncertainty instead of hiding it. See the full probability breakdown of today's matches yourself: open the live match analysis or start a free 3-day trial.
Frequently Asked Questions
Is this betting advice?
No. 11Stat is a football data-analytics platform, not a betting service. It never tells you what to bet; it shows 1X2 probability, xG, form and model agreement. The outputs are probabilities, not advice.
How is a match prediction calculated?
Not with a single guess, but with statistical engines. Each team's expected goal rate is estimated, converted to scoreline probabilities via Poisson/Dixon-Coles, and the match is simulated 10,000 times with Monte Carlo. An Elo/form engine adds a second opinion, and the results are combined as N/M model agreement.
Is it free?
You can start with a free 3-day trial and see the full probability readout of today's matches. After the trial, Core plans continue with regional pricing.
Are the accuracy and ROI figures real winnings?
No. All accuracy and calibration figures are simulated, paper-backtest metrics on historical data. They are past performance and never guarantee or imply any future outcome.
Does a high probability mean the result is certain?
No. High probability is not certainty — even a 75% read can go the other way in about one match in four. Football markets are efficient; probabilities are decision-support inputs, not promises.