Last updated: 21 May 2026
The idea sounds clean. An odds-on horse loses, on average, about four times out of ten. Lay it on Betfair and you collect the stake every time it doesn’t win. Stack a few of those wins together and you have a steady income. The maths is simple, the exchange does most of the work, and it doesn’t feel like betting in the same way a hopeful punt at 8.0 does.
Almost every backer has run that thought experiment. Race Advisor’s analysis of thirteen years of UK racing says, very politely, that the maths is right and the strategy still doesn’t work.
Can you lay odds-on horses for profit?
No, not as a blanket strategy. Across UK racing between 2013 and 2026, laying every odds-on runner at Betfair SP loses money after commission. Narrower filters can flip the edge for short bursts, but the market closes those gaps quickly and the variance is brutal. The realistic conclusion is that the odds-on lay market is one of the most efficient corners of British racing and the edge does not sit there.
What “odds-on” actually means and why people want to lay it
An odds-on horse is one priced shorter than evens. A Betfair Starting Price below 2.0. The horse needs to win less than half the time to make sense as a back bet, and the implied probability the market is selling sits between 50% and 100%. Across 22,762 odds-on UK runners between 2013 and March 2026, the actual win rate is 61.3%. Bet them all to level stakes and you finish slightly down after commission, which is the market doing its job.
The lay-side appeal is straightforward. If you lay every one of those 22,762 horses at Betfair SP, you collect the stake on 38.7% of them. The remaining 61.3% cost you a smaller amount each because the prices are short. You are trading high strike rate for low payout, which is the opposite of how most backers bet, and it appeals to anyone who has been ground down by chasing winners. The thinking goes: catch a four-out-of-ten loser at 1.65, then keep doing it.
The trouble is that the people setting those prices have done exactly the same calculation. The Betfair SP forms because thousands of people, including professional teams, are pricing the same race. The question is not whether the maths is correct. The question is whether the maths leaves anything on the table.
What thirteen years of data show about laying every odds-on horse
Lay one unit on every Betfair-SP favourite priced below 2.0 between July 2013 and March 2026. Apply the standard 2% commission on net winning markets. Track the result across 22,752 races.
The horse wins 61.3% of the time. The back-side A/E ratio sits at 0.997, which is as close to a perfectly priced population as you will see anywhere in racing. The lay-side total profit and loss across the full thirteen years is minus 99.4 units. That is a return on investment per bet of minus 0.44%.
To frame it: nearly 23,000 bets, almost half a unit of loss per hundred. Commission is doing most of the damage. The Betfair SP on these horses is set just slightly below what the long-run results would justify on the lay side. Strip out commission and you have a market that breaks even to four decimal places.
The win rate inside the odds-on band varies sharply with price, and the market segments it cleanly. Horses priced 1.01 to 1.30 win 82.9% of the time. Horses priced 1.30 to 1.50 win 71.3%. Horses priced 1.50 to 1.70 win 61.6%. Horses priced 1.70 to 2.00 win 53.7%. Each step down the price ladder maps almost exactly to a step down in actual strike rate, which is the cleanest signature of an efficient market you can ask for.
Where the lay-side edge briefly appears (and why it doesn’t last)
The temptation, once you see the overall result, is to narrow the filter. There are always slices of the data where the back side overpays by a fraction, and laying those slices looks like a strategy.
Flat Turf odds-on favourites, across 6,652 runners, give a lay ROI of plus 0.28% per bet. Flat AW gives plus 0.26% across 4,433 runners. Class 6 racing gives plus 1.36% across 1,711 runners. Standard going (essentially all-weather) gives plus 1.01% across 3,615 runners. Read those numbers in isolation and the lay-side edge sounds real.
Then look at the offsetting slices. Chase Turf odds-on favourites lose 1.96% per lay across 2,995 runners. Class 3 racing loses 2.11% across 1,689 runners. The 11-15 runner field band loses 2.28% across 3,050 runners. Heavier going on Good gives a lay ROI of minus 1.36% across 6,001 runners.
The pattern is that the small lay edges in one slice are paid for by larger lay losses in another. Combine them and you return to the headline minus 0.44% per bet. That is exactly what an efficient market should produce: small wobbles around the true price, no durable edge in any direction.
Time windows show the same story. Across 2013 to 2018, lay ROI was minus 0.62% per bet. Across 2018 to 2023, minus 0.56%. Across the most recent five years to March 2026, minus 0.85%. Look at the last two years alone and the number flips to plus 0.87% across 3,768 races, which is the kind of number that gets a backtester excited until the next two years arrive. The variance is enormous because lay losses are larger than lay wins by definition, and a single 1.20 BSP horse breaking favourite trust costs you what twenty straight successes paid.
Why progressive staking doesn’t rescue the strategy
The other instinct is to fix the small loss with progressive staking. If laying flat loses 0.44% per bet, increase the stake after every losing lay and recover when the next loser arrives. The maths is the same as a Martingale on a coin flip, applied to an event that pays out at uneven prices.
The arithmetic falls over quickly. The average BSP across odds-on favourites is 1.65. After commission, a losing lay at 1.65 costs you 0.65 per unit; a winning lay returns 0.98. To recover one full unit of loss, you need a winning lay to cover the previous loss and produce a new unit on top, which means each stake step has to grow by a factor of around 2.7. Five steps in, you are staking 53 units to make 1 back. Six steps in, you are at 143. The longest losing run in the odds-on favourite data is greater than four in every five-year window observed. The bankroll required to survive a real losing streak is many times what the strategy could plausibly earn even if it worked.
This is the deeper problem with progressive staking on short prices generally. The maths is constructed to convert a bankroll into a small income, which sounds attractive until you ask how often the bankroll fails to convert. Whenever the losing run exceeds the planned ladder, the entire stack disappears in a single race. The trade you are making is small frequent wins against rare catastrophic losses, and the long-run expectation is identical to flat staking with worse variance.
What the data does suggest you do with odds-on horses
The honest takeaway from the numbers is not “lay odds-on horses”. It is “the odds-on favourite market is the most efficient corner of UK racing, so look elsewhere for an edge”.
There are three corollaries. The first is that you can safely ignore most of the odds-on favourites in your daily races for the purposes of finding value bets. The market has them priced. Looking inside that 22,752-race population for a hidden lay angle is statistically harder than looking elsewhere for a back angle. The second is that the same market efficiency tells you something useful about odds-on horses on the back side: backing them blindly returns you a loss equal to the bookmaker margin or, on the exchange, a result close to break-even. The price is honest. There is just no edge to be extracted from it.
The third corollary is the one that matters for value-hunting. The interesting prices are not in the 1.01-2.00 band. They are in the longer prices the market is less rigorously efficient about. Race Advisor’s wider analysis consistently finds the back-side edge sitting at Betfair SPs of 4.0 and above, where the market has more disagreement and less consensus, and where multi-rating analysis can find horses the market is pricing as 8.0 that the data thinks should be 6.5. That is where the edge is. The odds-on lay market is, statistically, a dead end.
The honest conclusion: market efficiency is the lesson, not the loss
If the answer to “can you lay odds-on horses for profit” had been a clean yes, every professional team in the world would have been laying odds-on horses for the last twenty years. They are not, because they ran this analysis a long time ago and concluded the same thing the 2013-2026 data confirms: the market sees the question and prices it correctly, give or take a fraction of a per cent that commission absorbs.
The useful output of that finding is not a strategy. It is information about where the efficient frontier sits. Knowing that the lay-side edge does not exist on short-priced favourites tells you something about every other price you look at. When you see a horse at 1.45 in a Class 4 handicap, the data says the market has it about right. When you see a horse at 7.5 in the same race, the question of whether the price is honest is genuinely open. That is where the work is.
The 2014 version of this article reached the same conclusion using a single year of data. Thirteen subsequent years have done nothing but reinforce it. The edge on short-priced favourites does not exist, and a strategy built on its absence is not a strategy at all. The honest move is to spend your analytical time on the part of the card the market has not priced as cleanly. The Race Predictor on today’s cards is the place to start that work.
