Guide

Both Teams to Score: How Statistics Estimate It

Learn how expected goals and simple probability turn into a both teams to score percentage, with a worked example and the limits of the method.

5 min read17 November 2026by AI Football Predictor

It is the 80th minute. One side leads 1-0 and the other has barely threatened. A friend in the stand says, "Both teams always score in this fixture." Is that a real pattern, or just the memory of one or two wild games? Statistics can give a calmer answer.

What "both teams to score" actually means

Both teams to score (BTTS) is a yes or no question: will each side score at least one goal in the match, stoppage time included? Here is how it works in practice:

  • 1-1, 2-1 and 3-2 are all Yes.
  • 1-0, 0-0 and 4-0 are all No.

BTTS says nothing about who wins or how many goals are scored in total. A 5-0 and a 0-0 are both "No". That is why it helps to split the question in two: can the home team score, and can the away team score?

The core idea: two separate scoring chances

Goals are rare, discrete events, which makes them a good fit for the Poisson distribution, a classic tool for counting how often something happens in a fixed period. You give the model one number for each team, its expected goals for the match, and it returns the probability of 0, 1, 2, 3 goals and so on.

The part we need is simple. If a team's expected goals is written as 位, then:

  • The chance it scores zero goals is e raised to the power of minus 位.
  • The chance it scores at least one goal is therefore 1 minus that number.

If we treat the two teams as independent, the chance that both score is just the two probabilities multiplied together. That is the backbone of most BTTS estimates.

AIFP
These numbers, for every match todayProbabilities, expected goals and checked reliability, free, on the site.
See the probabilities

A worked example

Take an invented fixture, used only to show the method. The home team has an expected goals figure of 1.60 and the away team 1.10.

StepHome teamAway team
Expected goals (example)1.601.10
Chance of scoring 0 goalsabout 20.2%about 33.3%
Chance of scoring at least 1 goalabout 79.8%about 66.7%
Both teams score (79.8% x 66.7%)about 53.2%

So in this example, BTTS Yes is about 53% and BTTS No about 47%. Roughly one match in two like this would see both sides score. It is not a verdict on this particular game.

Now lower the expected goals to 1.00 and 0.80. The same steps give about 63% and 55% for each side scoring, and a BTTS figure of only about 35%. The lesson is that BTTS is very sensitive to the weaker attack. When one team is expected to score little, the whole percentage drops fast.

Where the expected goals numbers come from

The example above used round numbers, but real models build them from data. Typical ingredients include:

  • Attack and defence strength: how many goals, and how much xG, a team creates and concedes compared with the league average.
  • Home advantage: home sides generally score a little more and concede a little less.
  • Recent form, weighted sensibly: recent matches count more than old ones, but a single result should not swing the picture.
  • Quality of opposition: 2.0 xG against a top defence says more than 2.0 xG against a weak one.
  • Team news: a missing striker or goalkeeper can change both sides of the equation.

For context, top leagues typically average somewhere around 2.5 to 3 goals per match, and BTTS lands on Yes in roughly half of games, give or take depending on league and season. These are approximate figures, and individual teams can sit far from them.

Why the simple multiplication is not perfect

The independence assumption is convenient, but football does not always follow it. Good models adjust for several things:

  • Game state: once a team leads, it may sit deeper, which changes the chances for both sides. A trailing team often pushes forward and leaves space behind.
  • Low-score correlation: plain Poisson tends to misjudge results like 0-0 and 1-1. Refinements such as the Dixon-Coles adjustment correct this.
  • Red cards and injuries: they can break a match open or shut it down.
  • Tactical style: two open, aggressive sides tend to produce different matches from two cautious ones, even with similar average numbers.
  • Context: cup ties, dead-rubber games and weather can all shift how a match plays out.

None of this makes the basic method useless. It means a BTTS percentage is a well-founded estimate, not a measurement.

How to read a BTTS percentage

A figure such as 53% means that, across many similar matches, you would expect Yes about 53 times in 100. Any single match can still go either way. A 70% chance still fails three times in ten.

AI Football Predictor publishes these numbers for every match, including win probabilities, expected goals and BTTS chances, so you can see how the reasoning fits together rather than just a single headline figure. When you look at them, a few habits help:

  • Compare the BTTS figure with each team's chance of scoring at least once. The lower of the two usually drives the result.
  • Check the expected goals behind it. A high BTTS with two modest xG values is less convincing than one built on two strong attacks.
  • Look at team news before kick-off, since models can only use information available when they run.
  • Remember that probabilities describe likelihood, not certainty.

If you also follow betting markets, keep in mind that statistics are not guarantees and that gambling is for adults only and carries real risks. Use numbers to understand the game, not to chase outcomes.

Common mistakes when thinking about BTTS

  • Trusting a short streak. Five BTTS matches in a row for one team can easily be noise.
  • Ignoring the opponent. A leaky defence only concedes if the other side can create chances.
  • Confusing BTTS with high scoring. A 3-0 is high scoring but BTTS No, and 1-1 is low scoring but BTTS Yes.
  • Forgetting the weaker attack. The team least likely to score is the limiting factor.
  • Treating a percentage as a promise. Even 80% is not a certainty.

The takeaway

Estimating BTTS comes down to two questions: how likely is each team to score at least once, and how do those chances combine? Expected goals gives the starting point, simple probability does the arithmetic, and adjustments for game state and correlation sharpen the result. Used with a clear head, the figure helps you judge whether "both teams always score" is a genuine pattern or just a good story.

faq

FAQ

How do you calculate the chance that both teams score?

Estimate each team's expected goals, convert it to the chance of scoring at least one goal (1 minus e to the power of minus xG), then multiply the two probabilities. More advanced models add adjustments for game state and low scores.

Does a high BTTS percentage mean a high-scoring match?

Not necessarily. BTTS only asks whether both teams score at least once. A 1-1 draw is BTTS Yes with few goals, while a 4-0 win is BTTS No with many.

Is a BTTS probability a guarantee?

No. It is an estimate of likelihood based on past data and current information. Even a high percentage fails regularly. Gambling is for adults only and carries risks, so statistics should never be treated as certainty.

The probability of every match, free

Today's most likely matches, live, team sheets and proof of reliability, free on the site.