Betting exchange vs bookmaker odds: which price is closer to the truth?
By FootInsights · Published · 5 min read
Anyone comparing betting exchange vs bookmaker odds notices the same thing within about thirty seconds: the exchange price is bigger. Same match, same outcome, a number that is often five or six percent longer. The obvious conclusion — that the exchange is simply more generous — is wrong, and the real explanation is more useful than the myth. The two prices are not the same kind of object. One is a product with a fee built into it. The other is what two strangers agreed on, with the fee charged separately at the end. Understanding which is which changes how you read every probability you are ever quoted.
The structural difference in one sentence
A bookmaker takes the other side of your bet and sets the price it is willing to do that at. An exchange takes no position at all: it matches you against another customer who wants the opposite side, and charges commission on winnings for running the matching.
That single difference produces everything else. Because the bookmaker carries the risk, it must be paid for carrying it, and the payment is baked into the odds themselves. Because the exchange carries no risk, its prices are whatever supply and demand settle on, and it bills you afterwards.
The consequence is that an exchange lets you be the bookmaker. Backing an outcome is the familiar bet. Laying it is the other side: you accept someone else’s stake and pay out if the outcome happens. Lay £10 at odds of 4.0 and you are risking (4.0 − 1) × £10 = £30 to win £10. That is exactly a bookmaker’s position on that selection, in miniature. The margin has not disappeared from the market; it has become something you negotiate rather than something handed to you.
What the two books actually sum to
The clean way to compare any two prices is to convert them to implied probabilities and add them up — the method in our guide to converting odds to implied probability. Take a three-way match market.
A bookmaker’s card might read 2.30 / 3.40 / 3.20. The implied probabilities are 43.5%, 29.4% and 31.3%, summing to 104.2%. That surplus is the margin: the fee for the service, visible to anyone who can divide.
The same market on a liquid exchange might show best back prices of 2.42 / 3.55 / 3.35 — implying 41.3%, 28.2% and 29.9%, a total of 99.4%. It looks like free money, and it is not, because those are the prices you can take right now from other customers, and every one of them is a fraction better than the level at which the market genuinely clears. On the lay side the same market might sum to a little over 100%. The truth sits inside the gap between the two, and that gap — the back-lay spread — is the exchange’s real width. In a heavily traded market it is narrow enough to be almost invisible.
Where commission puts the fee back
Commission is charged on net winnings, not on turnover, and that changes the arithmetic in a way people routinely get wrong. If you back at decimal odds O with commission rate c, your profit on a winner is (O − 1) × (1 − c), so the price you effectively took is:
Effective odds = 1 + (O − 1) × (1 − c)
Back at 3.00 with 5% commission and your effective price is 1 + 2 × 0.95 = 2.90. A bookmaker offering 2.95 on the same outcome was the better price all along, despite the smaller headline number. Back at 1.50 with the same commission and you get 1 + 0.5 × 0.95 = 1.475 — the deduction is small in absolute terms because the winnings were small.
That asymmetry matters. Commission takes a fixed share of profit, so it costs you most on long shots and least on short favourites. A bookmaker’s margin does the opposite in most books: it is loaded disproportionately onto the outsiders, which is one of the mechanics behind how bookmakers set football odds and the favourite-longshot bias that comes with it. Two different fee structures, two different distortions. Neither price is raw probability.
Why exchange prices are the better probability estimate anyway
Once you strip both fees, exchange prices usually win the accuracy contest, for a reason that has nothing to do with generosity.
A bookmaker’s price reflects its opinion, its margin and its liability — it will shade a line because too much money has arrived on one side, not because anything about the match changed. An exchange price has no opinion and no book to balance. It is the level at which people who disagree stopped disagreeing, updated continuously by anyone willing to put money behind a view. In a well-traded market, that is the closest thing to a public, fee-free consensus probability that exists.
This is why serious evaluation of any forecaster uses closing market prices as the yardstick, and exchange closing prices where liquidity allows. It is a benchmark nobody controls and everybody can look up afterwards.
The limits worth stating plainly
Exchange prices are only as good as the money behind them, and that money is unevenly distributed.
- Liquidity is concentrated. Match-result markets in major leagues are deep. Minor leagues, obscure markets and anything more than a few days from kick-off can be thin, and a price with almost nothing matched at it is one person’s opinion wearing the costume of a market.
- The spread is the real cost in thin markets. A quoted back price means little if the lay price sits far away from it.
- Commission tiers vary, and some exchanges charge winning accounts more. A strategy that survives at one rate can die at another.
- Small edges do not survive fees. An edge of two percent against a bookmaker is not an edge at all once commission is applied to the exchange equivalent.
What to do with the comparison
The practical routine is short. Convert every price you are shown to implied probability. Strip the fee — margin for a bookmaker, commission for an exchange — so you are comparing forecasts rather than products. Then treat the exchange closing price as the number to beat, and be honest with yourself about how rarely anything beats it.
The same discipline applies to prediction sites, including this one. A published probability is a claim, and claims are checkable: score them against outcomes with a proper scoring rule, over enough matches for the answer to mean something. That is why our public track record lists settled predictions rather than a headline win rate — a win rate can be produced by picking favourites, while a probability that survives scoring against the market cannot.
Understanding the fee structure of a price does not make betting profitable. It only makes you harder to mislead. Stake only what you can afford to lose.