BTTS strategy: when to use “both teams to score”?

The “both teams to score” market (BTTS) does not ask who wins. It asks whether each side scores at least once. That indifference to the winner is precisely its point.

What has to be estimated

Under a Poisson model, writing λ₁ and λ₂ for each team’s expected goals, the probability that a team fails to score is e^(−λ). The probability that both score is therefore (1 − e^(−λ₁)) × (1 − e^(−λ₂)).

In practice: two teams expected to score 1.4 goals each give (1 − 0.247)² ≈ 57%. A lopsided match at 2.2 against 0.6 gives (1 − 0.111) × (1 − 0.549) ≈ 40%, even though the total of expected goals is identical. The total is not enough; the split is what decides.

When this market is interesting

• When both attacks are decent and both defences leaky: the 1X2 stays uncertain, BTTS much less so.
• When a clear favourite faces a team that rarely scores: the mirror case, where “no” carries the value.
• Rarely when the total of expected goals is very low: the market knows that too, and the price reflects it.

Beware the independence assumed between the two terms. An early red card, a scoreline that locks down, a team managing the end of a match: the two sides’ goals are not perfectly independent, and the product above slightly overstates the extreme cases.