Goals, Streaks and the Hot Hand: What Football Teaches Us About Randomness

A striker scores in five straight matches and commentators crown him unstoppable. A team loses four in a row and the manager is "finished." Football is a perfect laboratory for an uncomfortable truth: randomness produces patterns far more often than our intuition expects.

Goals follow a Poisson pattern

Goals are rare events scattered across 90 minutes, which is the situation the Poisson distribution describes. If a team averages 1.4 goals per match, the Poisson distribution calculator gives about a 25% chance of a blank, 35% of exactly one goal, and roughly 17% of three or more.

Scoring three is no miracle. It happens about once every six matches.

The model is not perfect, since stronger opponents and red cards change expectations mid-game, but it captures the essential shape. Analysts use it as a baseline, and a baseline is exactly what we need to judge whether a result is surprising or ordinary.

Home advantage adds a twist. Teams usually score more at home, so analysts fit separate averages for home and away sides and feed each into the model. Assuming goals are independent, multiplying the two distributions estimates how often a match ends 1-1 or 2-0, which is the backbone of many forecasting models.

Streaks happen by accident

Suppose a team has a 60% chance of winning every match. Five wins in a row has probability 0.6⁵, about 8%, for any given run of five. Across a 38-match season there are many such runs, and the chance of at least one five-win streak is roughly three in four. No momentum is required.

The binomial distribution calculator shows that same team wins at least seven of ten matches about 38% of the time, so "great form" is often just ordinary variance. The geometric distribution calculator answers a related question: if each match gives a 40% chance of a win, how long until the first one? On average, 2.5 matches.

Is the league table mostly luck?

Even a genuinely strong team has a wide range of plausible seasons. If a side wins each match with probability 0.6, the expected total over 38 matches is about 23 wins, but chance alone gives a typical spread of roughly six wins either way, from about 17 to 29. Draws complicate the picture, yet the lesson holds: league positions contain real skill and a lot of noise.

This also explains why overperformers tend to slide back. A team that wins far more than its underlying quality predicts has usually enjoyed some good luck, and luck does not repeat on schedule. Statisticians call this regression to the mean, and it quietly makes many managerial "turnarounds" look more impressive than they are.

The hot hand debate

In 1985, psychologists Gilovich, Vallone and Tversky concluded that the basketball "hot hand" was an illusion. In 2018, Miller and Sanjurjo showed that the original method contained a subtle bias, and that small streak effects may exist after all. The honest lesson is not that streaks are fake. It is that explaining them needs proper tests, and that skill and variance usually work together.

How to think about the next streak

Before crediting a manager, a tactic or a lucky pair of boots, compare the streak with what chance alone would produce. If a plain binomial model makes it unremarkable, the story deserves less certainty than the commentary gives it.

None of this spoils the game. Knowing the numbers simply adds a second layer of appreciation: the thrill of the match, and the quiet recognition of how much of it is chance.