Football match probability calculator

Give each team an expected goals figure for the match and this works out how often each result and scoreline should happen, using the same Poisson maths that sits under most public football models.

Where the numbers come from

Goals in football are rare, roughly independent events, which is the textbook situation for a Poisson distribution. If a team is expected to score 1.6 goals, the chance it scores exactly 0 is e^-1.6, about 20%, exactly 1 is about 32%, exactly 2 about 26%, and so on. Do that for both teams, multiply the two together for every scoreline, and add up the cells where home is ahead, level or behind. That is all this page does.

Where to get the two inputs

The cleanest source is each team’s expected goals in the match: the xG you will see on a Ninety match page once it is over, or a forecast of it before it starts. For a quick estimate before a game, average each team’s xG created and conceded over its last ten matches, then nudge the home side up a little (home teams typically score about a quarter more than away teams in the same matchup). A pair of 1.6 and 1.1 is a typical home favourite in a mid-table fixture.

Try some pairs

• 1.5 v 1.5: evenly matched. Each side wins about 38% and the draw is about 24%, with 1-1 the single most likely score.
• 2.4 v 0.7: a heavy favourite. The favourite wins about 75% of the time and 2-0 is the most likely scoreline.
• 0.9 v 0.8: a cagey game. The draw is about 34%, and 0-0 is the single most likely scoreline, ahead of 1-0.

Where it falls short

Two honest limits. First, goals are not perfectly independent: a team that goes ahead changes how the match is played, and draws come up a little more often than a pure Poisson says. Models such as Dixon-Coles correct for this. Second, an xG number for a match is itself an estimate. Garbage in, garbage out: the calculator is only as good as the two figures you give it. For a model that handles both, see our 90Machine track record.