How to predict football scores: a method that beats guessing

Most people predict a scoreline by picturing the game and writing down whatever number the picture suggests. That is fine for one match. Over a season it produces the same three results over and over, because the picture in your head is mostly made of the last thing you saw.

Here is a method you can repeat every week, in about ninety seconds a match.

Step one: start from what usually happens

Across Europe's big leagues, matches finish with about 2.7 goals in them, and the home side wins a bit more often than not. Before you know anything about the two teams, the honest starting point is a low-scoring home win or a draw.

That base rate is not the answer. It is the number you have to argue your way away from, which is a very different way of thinking than starting with a blank box.

Step two: adjust for the two teams, in that order

1. How good are they, really. The table, not the reputation. A famous club sitting fourteenth is a fourteenth-placed team.
2. How do they score and concede. Goals for and against tell you more than points. Two teams on 15 points play very differently if one has scored 20 and the other has scored 8.
3. Recent form, but lightly. Three wins in a row is a signal. It is not a bigger signal than a whole season of evidence.
4. Anything unusual. A suspended striker, a European trip on the Thursday, a manager who was sacked on Tuesday.

Each adjustment moves the score by roughly one goal, not three. If your reasoning takes you from 1:1 to 4:0, you have talked yourself into a story rather than a prediction.

Step three: sanity check the number

Ask two questions about the scoreline you just wrote:

• Has it happened before, between these sorts of teams? 2:1 and 1:0 happen constantly. 4:3 almost never does.
• Would I be surprised? If the answer is no for both your score and the opposite result, you have found a genuinely tight game, and those are worth attention because they usually pay more.

The scorelines that actually happen

Long-run frequencies across the top European leagues, rounded. This is the table worth having in your head:

That table is also why prediction games have to reward rarity. If everyone converges on 1:1 and 2:1, the exact-score points would all land in the same place, so a bonus for calling something nobody else called is what keeps bravery worth anything.

The mistakes that cost the most

• Predicting the team, not the match. Your team's away game at the league leaders is not the week to be loyal.
• Chasing the last result. A 5:0 on Wednesday does not make a 5:0 on Saturday likelier. It usually makes it less likely.
• Ignoring the goalless draw. People almost never call 0:0, and it lands more often than 3:1.
• Being brave everywhere. Bravery is a budget. Spend it on the matches where you have an actual reason, not across a whole coupon.

The last one is the difference between a good season and a frustrating one. In a game where a hard match pays more than an easy one, the skill is not being bold, it is choosing where to be.