How to Build a Correct Score Accumulator (and the Brutal Maths)

A correct score accumulator is a bet where every leg is an exact final score, and every one has to be right. The likeliest single scoreline in a football match lands only about 11% to 15% of the time, so a treble of correct scores wins roughly once in 500 attempts. The prices look huge because the chance is tiny, and the bookmaker's margin on these markets is much heavier than on a match result.

This guide shows the maths with a full scoreline grid, explains how to build one if you still want to, and sets out the alternatives that keep a much better strike rate.

How a scoreline probability is worked out#

A Poisson model treats each side's goals as independent random counts. If a team is expected to score an average of λ goals, the chance of exactly k goals is e^(−λ) × λ^k ÷ k!. Multiply the home and away figures for any scoreline, and you have its probability. Our Poisson research note goes through the idea and its limits.

Take a match with λ of 1.6 for the home side and 1.1 for the away side, the same example used in how ArtificialAcca's model works. Here is the scoreline grid, with home goals down the side and away goals across the top. All figures are percentages, computed with the formula above.

The favourite among all 121 scorelines up to 10 goals each is 1-1 at 11.8%, followed by 1-0 at 10.8% and 2-1 at 9.5%. The five likeliest scores together cover only 48.0% of the probability. The other 52% is scattered over everything else.

The most-likely-score figure moves with the type of match:

Real football history agrees. A Sports Trading Network article on correct score betting reports that even the most common scores have occurred less than 12% of the time across English league history. A single leg is therefore a 1-in-8 shot at best, and that is before the bookmaker's cut.

Two cautions on the Poisson numbers. The independence assumption tends to understate low-scoring draws, so 0-0 and 1-1 are probably a touch more likely than the grid says. And the λ inputs are estimates, so every figure here could be a point or two out either way.

What a correct score accumulator is really worth#

Pick the likeliest score in three matches and see what the treble looks like:

Multiply: 0.118 × 0.126 × 0.130 = 0.00193, or 0.19%. That is about 1 in 518, with fair odds of roughly 518.0 in decimal (517/1 in fractions). Choosing the single most likely score in every match is the best this bet can do, and it still wins about once in every five hundred attempts.

Add legs and it collapses. At 12% per leg:

With £1 staked every week on a four-fold of 12% scores, you would expect to place a winning bet about once in 93 years. The treble above is not much better. This is closer to a lottery ticket than a betting strategy, and it should be staked like one.

The margin makes it worse#

Fair odds are not the odds you are offered. Correct score is a market with dozens of possible outcomes, and bookmakers shade every one of them.

• In the Sports Trading Network article, Pinnacle's correct score margin for a League Cup semi-final was 5.3%, against 2.5% for the match result, and that covered scores only up to three goals per team. The same article says other bookmakers' correct score margins can be far larger, up to 60% in markets going to six or more goals.
• A SoccerNews guide to correct score betting puts the typical Premier League correct score overround at 15-20%, against 6-8% for the result. It gives no source for those figures, so treat them as an estimate.

Take an illustrative 18% overround, which sits inside that range. Dividing the fair price by 1.18 gives:

Those are not real quotes from a named bookmaker. They show what an 18% margin does to a fair price.

Now compound it across the treble. Each leg gets 1 ÷ 1.18 of its fair value, and three legs retain 1 ÷ 1.18³ = 60.9%. The treble's fair 518.0 becomes about 315 in decimal. Expected return per £1 staked is 61p, a loss of roughly 39p in the pound. Compare that with the 5% a typical result market takes per leg, covered in the hidden cost of accumulators. A three-leg acca at 5% per leg keeps about 86p in the pound.

How to build one anyway#

If you want a correct score accumulator for fun, there is a sensible way to go about it. You are buying a lottery ticket, so stake accordingly.

1. Keep it to two or three legs. A fourth leg turns 1 in 579 into 1 in 4,823.
2. Choose matches where one score dominates. A tight game at 1.2 v 0.9 gives 1-0 a 14.7% chance, a better single leg than a balanced match where 1-1 is 11.8%.
3. Compare the score probability with the price. Convert the odds you are offered to a chance (1 ÷ odds). At 7.16, 1-1 implies 14.0%, above the 11.8% in the grid, so that price is short. You want the offered implied chance to be at or below the model's.
4. Stake a fixed small amount. 50p or £1, an amount you would not miss.
5. Never chase. A losing run is the norm, not a signal.

Our guide to building an accumulator and the acca calculator cover the general arithmetic, and the staking guide explains why small, fixed stakes suit long odds.

Safer legs that use the same maths#

The same scoreline grid gives you the probabilities of every other goals market, and these are much easier to land. For 1.6 v 1.1:

The treble column assumes three matches with the same figures, which is a simplification, but the gap is the point. A treble of over 1.5 goals legs is hundreds of times more likely than a treble of exact scores, at much shorter odds. If you like the idea of a goals-based acca, look at BTTS accumulators, the over/under 2.5 goals guide, and the live pages for BTTS, over 2.5 goals and over 1.5 goals.

Shorter prices are not free either. Every leg still carries margin, and our leg-count article shows how fast the chances fall as you add them. But a 42% bet and a 0.17% bet are not the same kind of thing.

Why ArtificialAcca does not offer correct score legs#

The generator builds legs from match result, double chance, both teams to score and over/under 1.5, 2.5 and 3.5 goals. There is no correct score market, and that is deliberate.

Our method keeps the market in charge: the blended probability is at least 70% a de-vigged bookmaker price and at most 30% Poisson. A correct score market gives us a poor anchor. The prices are thin, the margin is large and uneven across scorelines, and removing it with proportional scaling is unreliable. We would end up leaning on the Poisson half, which is the weaker input, to invent edges the market does not support. We would rather not publish picks we cannot defend. The method is in full in how the model works, and the daily record, wins and losses, is on the results page.

If you want probabilities for markets we do cover, the generator builds a combination from them and shows each leg's reasoning.

The honest summary#

• The likeliest single score lands about 12% of the time.
• A correct score treble is about a 1-in-500 shot at fair odds, before margin.
• Margin on the market is higher than on the match result, which takes the expected return on an illustrative treble down to about 61p in the pound.
• If you play, stake a lottery-ticket amount and keep it to two or three legs.
• If you want an accumulator that wins reasonably often, use goals and BTTS legs instead.

18+ only. Correct score accumulators lose far more often than they win, and the figures above are probabilities, not predictions or promises. Gamble only with money you can afford to lose. Free, confidential support is available from BeGambleAware.org and the National Gambling Helpline (GamCare) on 0808 8020 133. You can block yourself from UK-licensed online gambling sites with GAMSTOP.