Every race at Epsom in the OddsLens results database from 22 April 2025 to 31 August 2026: 127 flat races, 1,070 runners at official starting price. What the numbers say, recomputed every week and never edited by hand.
Across 127 races the favourite won 27.1% of the time (36 of 133 bets, joint favourites counted separately) and £1 on every one lost 26.1p in the pound. The same bet at every UK and Irish course on the flat over the same window won 32.9% and lost 11.7p in the pound. Favourites here have been a worse bet than the national picture.
| Ground | Races | Strike rate | £1 return |
|---|---|---|---|
| Good or faster | 77 | 26.3% | −31.0% |
| Soft or heavy | 50 | 28.3% | −18.7% |
Flat races with eight or more runners, each field split into thirds by stall number, and the winners counted. A third that wins more than its share of the runners is where the stalls have helped.
| Stalls | Share of runners | Share of winners |
|---|---|---|
| Lowest third | 36.5% | 35.5% |
| Middle third | 33.9% | 29.0% |
| Highest third | 29.6% | 35.5% |
Over seven furlongs to a mile the thirds have won roughly in proportion to their numbers: no draw bias worth the name. (31 races.) Thirds are by stall number, so the lowest third holds the spare runner when a field does not divide by three; that is why the runner shares are unequal.
Trainers and jockeys with at least 15 runners and four winners here in the window. Strike rate is wins over runs; the return is £1 on every runner at SP. Thirty runs is a small sample, and the yards with the best strike rates are usually the yards with the best horses everywhere.
| Trainer | Wins / runs | Strike rate | £1 return |
|---|---|---|---|
| Andrew Balding | 5 / 38 | 13.2% | −36.8% |
| Gary & Josh Moore | 5 / 40 | 12.5% | −52.9% |
| A P O'Brien | 4 / 19 | 21.1% | +33.3% |
| Eve Johnson Houghton | 4 / 26 | 15.4% | −53.7% |
| Ralph Beckett | 4 / 32 | 12.5% | −45.9% |
| Jim Boyle | 4 / 53 | 7.5% | −15.6% |
| Jockey | Wins / runs | Strike rate | £1 return |
|---|---|---|---|
| Oisin Murphy | 11 / 47 | 23.4% | −6.1% |
| Rossa Ryan | 7 / 32 | 21.9% | −18.0% |
| Kieran Shoemark | 5 / 26 | 19.2% | −3.4% |
| Hector Crouch | 5 / 33 | 15.2% | −19.1% |
Since 28 April 2026 the OddsLens model has made 51 calls at Epsom, each logged before the off with the best price at that moment. 10 won, a 19.6% strike rate, for a −30.1% return at the logged prices and an average closing line value of +14.9%. That makes Epsom one of the model's weaker tracks so far; every one of those calls, and every call to come, sits on the public record, wins and losses alike.
Method. Every settled race at Epsom in the OddsLens results database from 22 April 2025 to 31 August 2026 (official returns, starting price, non-runners excluded). "Favourite" means the shortest starting price in the race; joint favourites are each backed for £1. Ground is the official going for the meeting: soft, heavy and their variants against everything else. The draw study uses flat races with at least eight drawn runners, in two bands (five to six and a half furlongs; seven furlongs to a mile and half a furlong), each field split into thirds by stall number. Big handicaps are races with "handicap" in the title and 14 or more runners at SP. The OddsLens record is the public ledger only (no shadow models), settled calls, at the prices logged before the off. Sections with too few races to say anything true are left out. Recomputed weekly; this page was built on 10 September 2026.