Dutching can feel like free money the first time you use it. Back three runners in a race, size the stakes correctly, and whichever one wins gives you the same return. What dutching cannot do is create value that was not already present in the prices. That is the part worth checking before you place the stakes.
The formula
For each selection, the stake is proportional to its implied probability:
a selection's stake = total stake × (1 ÷ its own odds) ÷ Σ(1 ÷ odds, across every selection)
The sum in the denominator, which we can call the book, is the combined implied probability of every selection you are covering. It is the most important number in the calculation, and also the one that is easiest to overlook when the focus is on getting the stakes right.
Example A: dutching without checking
Three runners at 2.00, 3.40 and 4.20, with a $300 total stake and no commission:
- Implied probabilities - 50.0%, 29.4%, 23.8%
- Book - 103.2%
- Stakes - $145.32 / $85.48 / $69.20
- Return, any winner - $290.64
- Result - a guaranteed −$9.36, whichever of the three wins
The stakes are calculated correctly. The split really does produce the same $290.64 return regardless of which runner wins. It is still a guaranteed loss, though, because the three prices add up to more than 100% of implied probability.
The book is 103.2%, meaning you are effectively paying more than 100 cents of implied probability for every 100 cents of stake. Dutching has done exactly what it is supposed to do: equalise the result. It does not guarantee that the result will be positive.
Example B: the same process, a book under 100%
The same three-runner setup, but with slightly better prices: 2.10, 3.60 and 4.60. The total stake is still $300 and there is no commission.
- Implied probabilities - 47.6%, 27.8%, 21.7%
- Book - 97.1%
- Stakes - $147.07 / $85.79 / $67.14
- Return, any winner - $308.85
- Result - a guaranteed +$8.85, whichever of the three wins
The process is exactly the same. The only thing that changed is that the three prices now combine to less than 100% implied probability.
That 2.9 percentage point gap is where the profit comes from. It was already present in the prices before you placed a single dollar. Dutching did not create the value. It simply distributes the result evenly across the three possible outcomes.
What commission does to a thin edge
The 2.9 percentage point gap in Example B is real, but it is also small. Exchange commission can eat into a thin edge surprisingly quickly.
Run the same 2.10, 3.60 and 4.60 prices through a 2% commission on the exchange's net winnings:
- Gross return, any winner - $308.85
- Commission (2% of the return) - $6.18
- Net return - $302.67
- Result - a guaranteed +$2.67, down from $8.85 without commission
The book was under 100% by 2.9 percentage points, but the commission removes roughly two-thirds of the resulting edge.
That is not a reason to ignore commission. It is a reason to include it when checking the book, rather than confirming that an edge exists and only afterwards discovering how much of it the commission takes away.
More selections, not fewer, is when the book check matters most
Dutching works in exactly the same way with four, six or ten selections. You stake in proportion to implied probability and use the same formula.
What changes as you add selections is how easy it becomes to lose track of the book. Three prices are easy enough to add mentally. Eight are not, and every additional selection creates another opportunity for the combined implied probability to move above 100% without being noticed.
The book check therefore becomes more important as the field gets larger. It is not just a useful habit for small fields. It matters even more when there are enough selections for intuition to become unreliable.
Where the edge in Example B actually comes from
A book below 100% rarely comes from taking every price from a single bookmaker's market. A bookmaker normally includes its own margin, which is what pushes the total above 100% in the first place, as in Example A.
Genuine sub-100% books are more commonly created by combining the best available price for each selection across several bookmakers. Each bookmaker may price the market slightly differently. The best price for one selection might come from Book X, while the best price for another comes from Book Y.
When those prices are combined, the total book can fall below 100% even though none of the individual bookmakers offers a sub-100% market.
That cross-shopping is where many real dutching opportunities come from. The formula determines how to distribute the stakes, but finding a book below 100% is a pricing exercise first.
Sum the book before you size a single stake
The practical rule is simple: add up 1 ÷ odds for every selection you are about to dutch before calculating any individual stake.
If the total is below 100%, you have a genuine edge and the stake split determines how that edge is distributed across the outcomes.
If it is above 100%, you are paying for the insurance of covering all the outcomes. That can still be a reasonable decision if your own probability estimate gives one of the selections significantly more value than the market does, but it is not a free result and the staking formula cannot make it one.
Three ways to leak value anyway
- Staking by conviction instead of by price. If you like one of the three more than the others, that belief belongs in your probability estimate before you dutch, not as a manual override to the formula afterwards - nudging one stake up "because it feels safer" breaks the equal-return property and leaves you exposed on the outcome you nudged away from.
- Rounding stakes carelessly. Small selections in a large field carry small stakes; round them off and the returns stop matching across outcomes, by exactly the amount you rounded.
- Ignoring commission on exchanges. If you are dutching against exchange prices rather than fixed-odds prices, commission comes off the return before profit is calculated — it changes the book check itself, not just the final number.
Run the actual prices in front of you through the dutching calculator. It shows the book percentage before showing you a single stake, because that is the number that tells you whether you are collecting value or paying for it.