What Is a Poisson Model, and Why Football?

A Poisson distribution describes how often a rare, roughly independent event happens in a fixed period. Goals in football fit that description well enough to be useful: they are infrequent, they can occur at any point, and to a first approximation one goal does not cause the next.

Give the model an expected number of goals for each side and it produces the probability of every scoreline. Aggregate those and you have probabilities for the match result, both teams to score, every over and under line, and more, all internally consistent with each other.

The assumptions do break down, and it is worth knowing where. Goals are not fully independent: a team that scores often changes how it plays, and a team that concedes early frequently opens up. Red cards break the model badly. Very high-scoring matches are underpredicted. It is a strong starting point that needs correction, not an oracle.

Key points

Poisson describes rare, roughly independent events over a period
Produces every scoreline probability from two expected-goal figures
All market probabilities derive consistently from one distribution
Independence assumption breaks on red cards and game state

Frequently asked

Why use Poisson for football?

Because goals are infrequent and spread across the match, which fits the distribution's assumptions closely enough to be genuinely useful.

Where does the model fail?

Where goals stop being independent: red cards, a side protecting a lead, and matches with unusual game states.

Is Poisson enough on its own?

No. It is a strong baseline that needs adjustment for context, which is why every serious model layers corrections on top.

Put it to use

What Is a Poisson Model, and Why Football?

Key points

Frequently asked

Put it to use

Read next