dutching for bettors: equal-profit stakes across a shortlist

what dutching is

Dutching is one of the oldest staking tricks in the book: instead of putting your whole stake on one selection, you spread it across two or more in the same market, sized so that any of them winning returns the same profit. Two horses in the same race, home win or draw in a football match, two runners in a handicap you cannot split — the stake is divided so the result of the race does not decide whether you collect, only which ticket pays. It is pure insurance: you trade the chance of a bigger single win for the certainty of a smaller one.

Be precise about what dutching is not. It is not an accumulator, where every leg must land; a dutch wins if any one selection wins and the rest can lose. It is not a system that creates value out of nothing; it reshapes how your profit lands, and if the prices are poor it loses money in the long run like any other bet. Dutching converts a shortlist you believe in into a single position.

the maths: equal profit in one formula

For a dutch to pay the same profit on any winner, every stake must satisfy the same equation: stake × (decimal odds − 1) = profit. Decimal odds include your stake back, so profit per pound rises with the price, and the longer-priced selections need the smaller stakes. Solving that equation for every leg gives the rule every dutching calculator uses: stake fractions are proportional to 1 divided by the decimal odds, normalised so they add up to 1.

In symbols, for selections with decimal odds o1, o2 and so on, let S = (1 ÷ o1) + (1 ÷ o2) + …. Then the stake on selection i is total stake × (1 ÷ oi) ÷ S. That one number S is the combined implied probability of the dutch — the chance the prices say one of your selections wins — and the profit when any leg lands is total stake × (1 ÷ S − 1). If S is below 1 the dutch profits; if S is above 1 it loses, whichever leg wins. That is the whole arithmetic, and the dutching calculator works through it in seconds for two to six selections, showing the book % before you commit a penny.

worked example: two horses

Take a race where you fancy two of the six runners: horse A at decimal 2.50 (6/4) and horse B at 3.50 (5/2), with a total stake of £100. First the inverses: 1 ÷ 2.50 = 0.400 and 1 ÷ 3.50 = 0.286, so S = 0.686. Horse A takes 0.400 ÷ 0.686 of the stake — 58.3% — which is £58.33. Horse B takes 0.286 ÷ 0.686 — 41.7% — which is £41.67.

Now watch what happens when either lands. If horse A wins, £58.33 × 2.50 returns £145.83, a profit of £45.83. If horse B wins, £41.67 × 3.50 returns £145.83, a profit of £45.83. Same profit on either winner; that is the dutch working. (Rounding to the penny makes the real return wobble a few pence.) The combined implied probability S = 68.6% is the market's verdict, through those two prices, that one of your pair wins, and the guaranteed profit is £100 × (1 ÷ 0.686 − 1) = £45.83. Compare a straight £100 on horse A: £150 profit if it wins, the full £100 lost if it does not. The dutch gives up upside for certainty — and whether that trade is worth it depends on your probabilities, covered below.

worked example: a three-way football market

Now a full three-way match market, priced home 2.10, draw 3.40, away 3.60 — the same book used in the implied probability guide. The inverses are 1 ÷ 2.10 = 0.476, 1 ÷ 3.40 = 0.294 and 1 ÷ 3.60 = 0.278, so S = 1.048. The stakes on a £100 dutch: home £45.43, draw £28.06, away £26.50.

Now the uncomfortable bit. If the home side wins, £45.43 × 2.10 returns £95.40, a loss of £4.60. If it is a draw, £28.06 × 3.40 returns £95.40, a loss of £4.60. If the away side wins, £26.50 × 3.60 returns £95.40, a loss of £4.60. Every outcome loses, because the three prices sum to 104.8% and the dutch has bought the whole book. Dutching every outcome of a market is not a bet at all; it is paying the overround in full, on every race. Carry away the rule: only dutch a subset you genuinely believe in, and never dutch the full field at a book that takes a margin.

when dutching beats a straight bet

Dutching wins the comparison in three situations. First, when you cannot separate two selections: rate horse A at 40% and horse B at 38% and picking one is a coin flip dressed as a decision — dutching both pays you on whichever of the pair lands and only loses when both lose, a 22% miss instead of a 60% or 62% one. Second, when you want insurance on a strong opinion: backing the favourite and dutching the second favourite protects a short-priced pick you do not fully trust. Third, when the prices are generous enough that the dutch itself is a value bet — combined implied S below your combined true estimate — and the equal-profit structure is simply the tidiest way to collect.

The other side of the ledger matters too. A dutch never beats a straight bet on upside: the same £100 that returns £150 on a winning single returns £45.83 on a winning dutch, and if your single is genuinely the best value on the card, adding a second leg only dilutes it. A dutch with one bad leg is worse than no dutch at all — every leg carries its own margin and its own probability, and a −EV passenger drags the whole position down. Dutch when the shortlist is real; never dutch to avoid making a decision.

the overround is the tax

The risk in any dutch is the margin: you pay it on every leg. Each price carries the book's overround on top of the true chance, and the combined implied probability S is that tax in one number: S = 1.048 means the dutch loses 4.8% of the stake whichever outcome lands. Horse racing margins run fatter still — a 108% book dutched across three runners bleeds roughly 8p of every £1, every race. The mechanics of why books always sum past 100% are covered in the implied probability and overround guide; here the point is simpler. The margin comes out of your edge first, and a dutch multiplies the number of margins you pay. That is why the book % display on the dutching calculator is the first number to read: it is the fee on the position, stated up front.

dutching, implied probability and value

Everything above is implied probability in action. Each price is 1 ÷ odds as a percentage, S is the combined implied probability of your shortlist, and the value test writes itself: if your combined true probability — the sum of your honest per-leg estimates — beats S, the dutch has positive expected value; if it does not, you are spreading a losing bet thinner. Run the two-horse example: horse A at 40% and horse B at 30% combine to 70%, against S = 68.6%. The dutch returns £1.4583 per £1 staked when either wins, so the expectation is 0.70 × 1.4583 − 1 = +2.1% per pound, a genuine if thin edge. Notice both legs must earn their place: horse A at 2.50 against a true 40% is exactly fair, while horse B at 3.50 against 30% carries the whole +5% of the value — and checking each leg individually is the value betting habit applied to a shortlist instead of a single.

Two practical points follow. First, price shopping matters twice in a dutch: a tick of improvement on any leg raises the profit on the whole position, and the odds converter is the quick way to compare fractional, decimal and American quotes across books. Second, once the dutch passes the value gate, treat it as a single position and size it as one: the combined bet wins with probability equal to your combined estimate and pays 1 ÷ S to 1, which is the shape the staking calculator expects, so the whole dutch gets a proper Kelly stake rather than two gut-feel bets. The complete loop — convert the prices, strip the margin, compare your estimates, stake the edge — is the one described in the value betting guide, with the shortlist standing in for the single selection.

the dutching checklist

A short routine before any dutch, in order: write down a true probability for every leg and add them for your combined estimate; compute S by adding 1 ÷ odds for each leg (the dutching calculator shows it); pass the dutch only if your combined true probability beats S; check every leg on its own merits so no −EV passenger rides along; shop the best decimal price on each leg, because a tick moves the whole position; size the dutch as one Kelly-staked bet, not two hunches; and never dutch a full market — when S runs past 1, every outcome loses and you are simply paying the overround. Seven checks, none of them optional.

the honest endnote

Dutching is risk management, not edge creation. It cannot turn bad prices into good ones — it only decides how the profit lands; a dutch of two losing ideas loses twice as politely. Done with real probabilities, real value checks and real stakes, it becomes one of the safest shapes a bet can take. Bet within your means, treat the bankroll as money you can afford to lose, and if betting stops being enjoyable, stop. 18+ | BeGambleAware.