Why Do the Bookmakers Always Seem to Win?

Because they are not trying to beat you at picking horses.

That is the part most people have upside down, and it is worth saying plainly before anything else. A bookmaker does not need to know which horse will win. They need the money to land across the runners roughly in line with the prices they set, and they have built those prices so that when it does, they keep a slice whatever happens.

You can see it in our own results. We run a computer model that prices every runner in every British and Irish race, and over 9,568 races it has been a genuinely decent judge: its top pick wins 24.3% of the time. Backing that top pick every day has still returned −14.0%.

Being right often is not the same as being profitable, because the charge sits on every bet whether it wins or loses. That is the whole mechanism, and the rest of this piece is just showing it working.

A three-horse race, priced two ways

Say three horses each have exactly a one-in-three chance. A fair book prices them all at 2/1. Turn those into chances and you get a third apiece, adding to exactly one whole. Back all three for a tenner each and one of them wins, returning your thirty pounds. You finish level. Nobody has an edge.

Now price them at 6/4 instead. That is two chances in five apiece, and the three together come to a fifth more than a whole race. Back all three for a tenner and you still stake thirty pounds, but the winner returns twenty-five. You are five pounds down before a horse has left the stalls, and it did not matter which one won.

Five pounds lost on thirty staked is a takeout of about a sixth, which sits right in the middle of what we actually measured across the festival cards this summer.

Nothing in that example required the bookmaker to know anything about the horses. They only had to shade every price by the same amount and take bets on all three.

Real racing is messier. Money does not spread itself evenly, prices move as it arrives, and a bookmaker who badly misjudges a race can lose a lot on the day. But across thousands of races the structure holds, and the structure is the point.

What this changes about how to think

Three consequences follow, and they are more useful than any tip.

The first is that the bookmaker's edge does not depend on you being bad at picking. It is charged on every bet, including your good ones. A punter who is a genuinely fine judge of a horse still pays it.

The second is that where you bet matters as much as what you back. We have measured the charge in the biggest fields running to nearly three times what it costs in the smallest. Nothing about your selection method changes as reliably as that.

The third is the uncomfortable one, and we would rather say it than dance round it. We have tested 24 different ways of betting across 29,286 races. The number that made money is 0. The best of them still lost -5p in the pound.

That is not a reason to give up on racing. It is a reason to treat a betting bank as the price of an afternoon rather than an investment, and to notice that the size of the charge is one of the few things you can actually control.

Later this month we look at what that means for the once-a-year punter, who is charged the same rate as everyone else.

A word on all of this

None of these pages is a tip, and none of them describes a way to win. They describe what betting costs, which is a different and more reliable subject. If your betting has stopped being fun, BeGambleAware has free, confidential help, and the National Gambling Helpline is on 0808 8020 133.

Common questions

Do bookmakers lose money on races?

Regularly. The margin is built to work across many races and many bets, not to guarantee a profit on any single one. A well-backed favourite winning a big handicap can be an expensive afternoon for a layer.

If bookmakers always win, how do they lose to some punters?

The edge is an average across everything they take. Individual punters can and do beat it, particularly those who take a price before the market moves. Firms tend to restrict those accounts, which tells you they know exactly where the leak is.

Would I do better on an exchange?

An exchange charges commission on winnings rather than building a margin into the price, so the headline cost is usually lower. It is not free, and it takes more work. We exclude exchange prices from our margin measurements because they are not comparable.

Is any of this hidden or dishonest?

No. It is arithmetic on a printed racecard, and anyone can do it. It is simply that the price you see does not come with a label saying what share of it is the charge.

Does the model beat this?

No. Across 9,568 races it has returned −14.0%, and we publish that figure precisely so nobody can mistake the model for a way round the maths.