The Hidden Cost of Accumulators: How Bookmaker Margin Compounds

Bookmaker margin is paid once per leg, and an accumulator multiplies the legs together, so the margin multiplies too. At a 5% margin per leg, a single bet keeps 95.2% of its fair payout, a four-fold keeps 82.3%, and a six-fold keeps 74.6%. That is the hidden cost: not a fee you see, but a smaller price on every winning bet. This article works through the maths and puts numbers on it.

The tables are derived, not observed. They show what a given margin does under stated assumptions, not what any particular bookmaker charges.

Margin and overround#

In a bookmaker's market the implied probabilities of all outcomes add up to more than 100%. A 1X2 market priced 2.10 / 3.50 / 3.60 sums to 104.0%. That 4.0% excess is the overround, and it is a good measure of the margin in the market. The workings are in how to read implied probability and remove the margin.

If the margin is spread proportionally, a bet on any outcome pays you 1 ÷ 1.04 = 96.2% of the fair payout. Pinnacle, a bookmaker known for low margins, publishes an article on how margins affect betting profitability, a useful companion read to the maths below.

The compounding formula#

If every leg has an overround of m, the fair payout of an n-leg acca is reduced by a factor of

Retention = 1 ÷ (1 + m) to the power of n

The expected return per £1 staked, for a bettor with no edge, equals the retention.

Derivation: let a leg have a true probability p. The bookmaker's price has an implied probability p × (1 + m), so the odds are 1 ÷ [p(1 + m)]. For n independent legs the true chance of winning is p to the power of n, and the odds are 1 ÷ [p(1 + m)] to the power of n. Multiply them and the p terms cancel, leaving 1 ÷ (1 + m) to the power of n.

Table 1: share of the fair payout you keep#

At 5% and twelve legs you keep 55.7p of every pound in expectation. At 8% and ten legs, 46.3p.

Table 2: expected cost of a £10 stake#

These are long-run averages. On any single bet you either lose £10 or win a large amount, and the expected value is the weighted average of the two.

A worked example#

Six legs, each with a true probability of 72% (a fair price of 1.39). At 5% per leg the bookmaker's price is 1.32 (an implied 75.6%).

The chance of winning is identical. The payout is £18.20 smaller on a win, and about £2.54 smaller per bet in expectation. The smaller payout, not the failure rate, is the margin.

Margin per leg is not the same as the market's headline margin#

Two details change the picture.

1. You do not pay the market's whole overround on a given leg. In a three-way market with an overround of 104%, the proportional cost of any one outcome is 1 ÷ 1.04. The cost is the same for each outcome only if the bookmaker spreads the margin proportionally. If it loads the long shots, a favourite costs you less than that and an outsider costs you more.
2. Two-way and three-way markets differ. A BTTS or over/under market has two outcomes. A 1X2 has three, and double chance sums two of them. A bookmaker's margin on each can differ, so measure it for the specific market you want. Over 1.80 and Under 2.00, for example, sum to 105.6%.

What it takes to overcome the margin#

For a leg to have a positive expected return, your probability has to beat the de-vigged market probability by a relative 5% at a 5% margin. If the market's fair probability is 70%, you need 73.5% for break-even, and more for profit.

Across several legs, the requirement compounds in the same way: for six legs at 5%, the acca needs a combined chance 34.0% above the fair market chance just to break even. In practice that means a real edge on every leg.

What a boost has to be#

A boost on winnings must offset the compounded retention gap. The break-even uplift to total payout is 1.05 to the power of n minus 1:

The uplift in this table applies to the whole payout. Boosts normally apply only to the winnings (the profit), which raises the requirement: in the six-leg example above you need about 42% on the profit. That is the figure used in acca insurance and boost offers explained. Most boosts are smaller and capped, so they reduce the cost of the acca without removing it. That can be worth having, as long as it does not drive you towards more legs.

Accumulator versus the same money in singles#

Stake £10 as a six-fold at 5% per leg and the expected cost is £2.54. Split the £10 into six £1.67 singles at 5% and the expected cost is £0.48. The singles cost less because margin is paid once per bet, not six times on the same pound.

That is not an argument for singles. They pay far less when they win. It is a reminder that the extra payout of an accumulator is not free. You pay for it in margin, and that cost scales with the number of legs. See how many legs should an acca have for how this plays out in strike rates.

Shopping for price#

Margins differ by bookmaker and market. Take two bookmakers with 3% and 6% per leg. On six legs, the retention is 83.7% against 70.5%. The lower-margin price pays 18.8% more on every winning acca, for exactly the same selections. That is the cheapest improvement most bettors can make. Compare prices before you bet, and prefer well-covered markets, which is covered in best leagues for accumulators.

What this means for ArtificialAcca#

The generator uses the best price across bookmakers for each leg, which helps with the margin, and its leg limits keep legs within sensible odds. It cannot remove the margin. A methodology note lists this under what the model does not do. Our view is that an acca is an entertainment product with a known price, and the best response is to know the price. You can see each leg's odds and de-vigged probability in the generator and check how every daily acca settled on the results page.

18+ only. Margin means accumulators cost more than they appear to, and most lose. These tables show probabilities and averages, not what will happen on any one bet. If you need support, visit BeGambleAware.org or call 0808 8020 133.