Methodology
Soccer Stats Hub is built to make football prediction research more transparent. The site combines league-level trends, team form, match odds, xG, PPG, goal markets and modelled probabilities so users can see why a fixture may be interesting.
Data Inputs
Core match and league statistics come from industry leading stat websites used throughout the app, including season, fixture and team fields. Selected competition and player ranking views use additional industry leading stat website metric feeds where available.
Prediction Signals
- Recent form and home/away performance windows
- xG for, xG against and rolling xG difference
- PPG, win/draw/loss rates and league position context
- BTTS, Over 2.5, Under 2.5 and clean sheet percentages
- Market odds and implied probability comparisons
- Correct score probabilities generated from goal expectation models
Goal Expectation (Lambda)
At the heart of our fixture modelling is a pair of expected goals values, called lambdas: one for the home side and one for the away side. Each lambda estimates how many goals that team is likely to score in the specific match, before we turn those expectations into a full range of scoreline probabilities.
We start with the league scoring environment. Season average goals are split into home and away baselines so a typical home attack is not treated the same as a typical away attack. Each team's attacking strength is then weighed against the opponent's defensive weakness to produce a raw goal expectation.
Those raw values are tuned rather than taken at face value. Small sample sizes are dampened towards the league average when a team has played fewer than ten matches, so early-season outliers do not dominate. Missing players and lineup absences can nudge attack or defence lambdas up or down. xG efficiency is used for a light regression adjustment when a side has scored more or fewer goals than chance quality suggests. Rolling xG form and recent results can shift the final multiplier within safe bounds. A new manager flag can temporarily lift a side's expectation while the squad settles.
On continental and international fixtures we blend the statistical expectation with market odds more heavily, because squad rotation, travel and knockout context can move results away from domestic league form. Every lambda is clamped to a realistic floor and ceiling before it enters the probability engine.
Poisson Distribution and Market Probabilities
Once home and away lambdas are set, we model each team's goals as a Poisson distribution. In plain terms, Poisson maths answers the question: if a team is expected to score 1.6 goals, how likely are 0, 1, 2, 3 or more? We build a score matrix by combining the home and away Poisson outcomes for every scoreline up to five goals each.
A Dixon-Coles style adjustment is applied to low-scoring outcomes such as 0-0, 1-0, 0-1 and 1-1. Football draws and tight games happen slightly more often than a naive independent Poisson model suggests, so this step keeps the matrix better aligned with real match results.
The matrix is normalised so all scoreline probabilities sum to 100%, then lightly calibrated so extreme long-shot scores do not dominate the output. From that single matrix we derive:
- Home win, draw and away win probabilities
- Both teams to score (BTTS) yes and no
- Over and Under 2.5 goals
- The most likely correct score
These model probabilities are shown alongside bookmaker implied prices on fixture pages so you can compare what the stats suggest with what the market is pricing in. The predicted scoreline on the page is the highest-probability outcome from the same matrix, not a separate guess.
How To Use The Outputs
Treat every table and prediction as a research aid, not a guarantee. Start with league and team tendencies, then check the fixture page for form, head-to-head data, venue splits, model probability and the current market price.
For a longer, fan-friendly walkthrough of the same process, read How we predict a game.
Update Cadence
Fixture and league data refreshes as source data updates throughout the day. Some competition pages and tournament previews update less frequently when the underlying season or editorial data changes.
Responsible Use
Soccer Stats Hub is an analytical tool. Predictions are statistical estimates and can be wrong. Users should follow local laws and gamble responsibly if they use football stats for betting research.