GUIDE · 5 min read

How to Read Our Match Predictions

Our predictor outputs a wall of numbers: 1X2 probabilities, fair odds, totals, BTTS, a heatmap of scorelines. Here's what each one means — and the three mistakes almost everyone makes with them.

1. Probabilities are frequencies, not predictions

"Home win 48%" does not mean the home team will win. It means: in 100 matches with exactly these characteristics, the home team wins about 48 of them. The single most likely scoreline (say, 1–1 at 13%) still won't happen 87% of the time. If a model's output surprises you, that's normal — probability is a long-run language.

2. Fair odds are for comparison, not for betting

We show fair odds (1 ÷ probability) so you can compare against bookmaker prices. If our fair price on the draw is 3.42 and the bookmaker offers 3.60, the model sees value. But the model doesn't know about the suspended striker, the waterlogged pitch, or the manager who rests everyone before a cup final. A gap between model and market is a question to investigate, not a bet to place. See our calibration notes on why model-vs-market divergence usually means the model is missing something.

3. The three classic mistakes

  1. Chasing the favorite. A 70% home-win probability feels safe, but at odds of 1.35 it needs to win 74% of the time just to break even. Probability without price is meaningless.
  2. Ignoring the margin. Bookmaker odds include a ~5% margin. Always strip it (our no-vig tool does this) before comparing to model probabilities.
  3. Overweighting one match. A model that's right 55% of the time still loses 45% of individual bets. Judge models over hundreds of predictions, never over a weekend.
The one-sentence version: use predictions to understand how likely things are, use prices to decide whether anything is worth doing — and never confuse the two.

GoalMath tools are for informational and educational purposes only. Nothing here is betting advice.

Related reading: How the Dixon-Coles Model Predicts Soccer Scores · Expected Goals (xG) Explained: What the Number Really Means