Do jumps favourites make money?

Of all the folklore in betting, this is the one that refuses to die: the favourite over jumps is the one bet that actually makes money. You will have heard it in the pub, read it in a tipping column, or seen it quoted as gospel on a forum. The logic sounds watertight. Flat racing, the story goes, gets turned over by some 33/1 chancer every other afternoon, but jumping is a proper test of stamina and class, so the best horse holds up and the form is reliable. Back every favourite over the sticks, let the cream rise, and the prices are fair enough that you come out ahead.

It is the most natural bet in racing dressed up as a clever one. The favourite is the horse the whole market trusts, so backing it feels like siding with the wisdom of the crowd rather than gambling, and over jumps that feeling gets an extra layer of comfort. National Hunt horses are older, more exposed, more thoroughly tested over a longer trip, so people assume there are fewer nasty surprises than on the Flat. Add the half-remembered claim that this was the one profitable angle anyone ever found, and you have a belief that feels both sensible and secretly canny.

We wanted to know if it holds up, because if any favourite bet was ever going to beat the bookmaker, this was the one with the reputation for it. So we did not argue about it. We took every jump race in a large sample of real British racing and backed the favourite in all of them, flat stakes, settled at the official Starting Price. Then we counted the result honestly, including the part of the story the folklore quietly leaves out: the horses that never finish.

The appeal works on two levels, and both are seductive. The first is the everyday pull of the favourite itself. It wins more often than any other single runner, about a third of all races, so you are constantly collecting. That drumbeat of small wins feels like proof the method works, and it quietly hides the truth that the wins are too small and too infrequent to cover the losses. People mistake a high strike rate for profit, when how often a bet lands tells you nothing on its own about whether the price was big enough to pay.

The second level is specific to jumps, and it is where the real trap sits. Pascal, our deadpan racehorse, puts it the way most punters do: Flat favourites get mugged by some no-hoper every other race, but jumping is a proper test of stamina and class, so the best horse holds up, the form is reliable, the prices are fair. This is the safe one. It sounds like hard-won wisdom. It is the opposite.

On top of that sits the folklore. A lot of people half-remember reading that the jumps favourite was the one bet that turned a profit, and that memory does a lot of heavy lifting. It makes the bet feel not just safe but smart, a rare edge the bookmaker somehow left on the table. The problem is that the claim was real for a while only because the sum behind it was wrong. Once you fix the sum, the comfort evaporates. What Pascal's reasoning ignores is the one thing that defines jump racing, and it is the same thing that turns this supposed winner into the worst favourite bet on the page.

The thing Pascal's reasoning ignores is that jumpers fall. Over hurdles and fences your horse can be going best of all, travelling sweetly, full of running, and still hit the deck three out or be eased to a halt at the second-last. When that happens your bet is gone. A faller is a losing bet, your full stake handed over, exactly like a beaten favourite that trailed in last. And jump racing serves those up far more often than the Flat does.

That is the whole story of why this loses more than its gentle reputation suggests, and it runs on two engines at once. The first is the same one that beats every backing line in racing: the bookmaker's built-in margin, the overround. Add up the chances implied by every price in a race and they come to more than 100%, roughly 12% on a typical field and climbing towards 30% in a big-field handicap chase. That surplus is the house cut, and you pay it on every single bet, win or lose. The market prices the jumps favourite almost right and then charges you the overround for the privilege.

The second engine is the one Pascal forgets: the races where the favourite simply does not complete. A horse pulled up lame, a faller at the open ditch, an unseated rider at the water, all cost you the full stake and none of them ever appears in the win column. Stack those lost stakes on top of the margin you pay on every bet, and the leak is bigger over jumps than on the Flat, not smaller. The crowd is not stupid and the favourite is the right horse to fancy. The price is just never quite big enough to cover the beaten favourites, and the fallers make sure of it. It is a steady, grinding loss, not a run of bad luck, and it does not turn.

We did not run a clever model or a fancy filter. The whole point of this experiment is its bluntness: back the favourite in every jump race, flat stakes, and add up what happens. The sample is 27,909 real British races in total, of which 10,862 were run over jumps, so this is not a curated highlights reel or a backtest tuned to look good. It is the whole population of races we hold, the good days and the bad, the small fields and the cavalry charges.

Every bet is settled at the industry Starting Price, the official odds returned as the race goes off. That matters, because SP is the kind figure. It assumes you got the official return with no commission, no early-price shading, no account restrictions. A real punter taking real prices in the real world bleeds a touch faster than the number you are about to see, so if anything our test flatters the bet.

The single most important rule is how we settle the runners that do not finish. Fallers, unseated riders and pulled-up horses are counted as the losing bets they are, because that is exactly what they are: your stake is gone whether the horse falls at the first or is eased down the back straight. This is the rule the old folklore broke. We also split joint-favourites rather than letting a tie quietly leak a winning result into the count. Where two horses share favouritism we treat the bet honestly rather than picking the one that happened to win.

That is the entire method. No selection, no staking trick, no hindsight. Pick the shortest price in every jump race, stake the same each time, settle at SP, and count the non-finishers as losses. Then read the number off the bottom of the column.

Here is the number. Backing the favourite in every jump race, flat stakes to Starting Price, returns -14.39% across 10,862 real British races. For every 100 pounds you put through, you get about 85.61 pounds back over the long run. The favourite won 35% of these races, a bit more than one in three, which is a perfectly healthy strike rate and exactly what makes the bet feel safe. It is not enough. The 95% confidence range runs from -16.8% to -11.9%, so even on the kindest reading of the data this is a clear, durable loss, not a quirk of variance you could wait out.

Now put that figure where it belongs. The jumps favourite is the worst favourite bet we measured. It loses more than the all-races favourite at -12.5%, more than the Flat favourite at -11.3%, and far more than the odds-on favourite at -7%, which is the least-bad backing line on the whole page. Every step of the comparison points the same way: the more fallers in the mix, the deeper the hole. The Flat, with fewer non-finishers, bleeds least among the all-comers favourites. Jumps, with the most, bleeds worst.

The contrast with the folklore is the headline. The old claim was that this bet made +4.6%. It did not. That figure came from dropping the fallers and pulled-up horses from the sum, which deleted a pile of losing stakes and flipped a real loss into a fake profit. Put those non-finishers back where they belong, in the losing column, and the +4.6% mirage collapses straight through break-even to -14.39%. The horses that hit the deck were always there. The old number just refused to count them, and counting them is the difference between a story and the truth.

So, do jumps favourites make money? No. They lose about 14p in every pound, forever, and far from being the one bet that beats the bookmaker, the jumps favourite is the worst favourite bet of the lot. The reputation had it exactly backwards. The very thing people trust about jump racing, the idea that a proper test of stamina and class protects the best horse, is the thing that sinks the bet, because over hurdles and fences the best horse falls, unseats and pulls up often enough to turn a near-miss of break-even into the steepest favourite loss we measured.

The famous +4.6% was never real. It was a sum with the fallers wrongly left out, and once you count them as the losing bets they are, the supposed edge does not just shrink, it vanishes and turns sharply negative. That is worth sitting with, because it is the cleanest example we have of how a flattering number gets made: not by lying about the winners, but by quietly forgetting the losers. We publish the honest figure, losses included, the same way we would publish a win.

This suits nobody as a way to make money. Past performance is not future returns, this is measured to SP with no commission so the real world is worse again, and no staking plan laid on top of a negative edge can fix it. If you want to size bets sensibly the honest tools are the dull ones: a flat stake you can afford to lose, or, only if you genuinely hold an edge, a fraction of the Kelly criterion. The jumps favourite is not that edge. It is proof of how completely the overround and the risk of not finishing beat you, which is exactly why we treat the market as a lens, never a tip.