Racing is full of "human" signals: the little things a trainer does that are supposed to tip their hand. They make great pub talk. The question we care about is simpler: when you actually count them, do they win, or does the betting market already know?
We measure every angle the same way, with A/E (actual wins divided by expected wins from the Starting Price). It is the fairest yardstick there is. An A/E of about 1.00 means the market has it priced to perfection and the angle gives you nothing extra. Below 1.00 means you are overpaying. Above 1.00 means the angle beat the market. We are not selling a system here, and nothing below is a tip. This is a look at how much of the folklore is real.
01When a trainer runs more than one
The classic puzzle: a yard saddles two or three in the same race, so which is the one they fancy? Across 36,146 of these multi-runner situations, the answer is clear, and it is the boring one. The stable's shortest-priced runner, the market's "stable elect", is where almost all the winning happens.
| Which stable runner | Runs | Win % | A/E |
|---|---|---|---|
| Shortest-priced of the stable | 36,146 | 13.8% | 0.87 |
| The stablemates (the rest) | 41,192 | 4.5% | 0.77 |
UK and Ireland, 2021 to date. When a yard with multiple runners won, it was the shortest-priced one 73% of the time.
So the romantic idea, that the stable's unconsidered second string is secretly the plot horse, is a money-loser (A/E 0.77). When a yard's runner wins, it is the one the market made favourite of theirs nearly three times in four. Useful to know, but not an edge: even the stable elect sits below 1.00, because the market is already watching the same thing you are.
02Following the gamble
This is the one that surprised us most, and it is worth tattooing somewhere. We split every runner by how its price moved from the morning to the off. "Steamed" means heavily backed (SP shortened to 0.7 of the morning price or less). "Drifted" means the price went out.
| Market move | Win % | A/E | Return at SP |
|---|---|---|---|
| Steamed (heavily backed) | 8.2% | 0.70 | -40% |
| Shortened a little | 14.3% | 0.87 | -19% |
| Steady | 12.4% | 0.89 | -19% |
| Drifted (price went out) | 10.1% | 1.06 | -7% |
UK and Ireland, 2022 to date, morning price versus Starting Price.
Backing the big gamble is the single worst value on the card: an A/E of 0.70 and a 40% loss at SP. The money has already shortened a price that was generous in the morning, and it tends to overshoot. The drifters, the horses the market cooled on, are the only group that actually beat their SP (A/E 1.06). "Follow the smart money" is, at Starting Price, close to the opposite of good advice. We dug into this one on its own, over 71,873 runners, here: do market movers win?
03The fresh horse off a break
Punters are wary of a horse returning from a layoff, and trainers know it. Does the caution go too far? Grouped by the days since a horse last ran, value quietly improves the longer it has been away.
| Days since last run | Win % | A/E |
|---|---|---|
| 1 to 7 days (quick back-up) | 14.8% | 0.89 |
| 8 to 28 days | 11.7% | 0.88 |
| 29 to 56 days | 12.1% | 0.90 |
| 57 to 120 days | 10.8% | 0.93 |
| 121 to 250 days | 10.9% | 0.92 |
| Debut (first ever run) | 9.5% | 0.87 |
UK and Ireland, 2022 to date. A gentle rise in A/E with the length of the break.
The market slightly over-discounts the horse coming back from a rest: the longer-break groups post the best A/E figures (still below 1.00, so no free money, but better than the fresh-and-fit runner the crowd piles on). The honest footnote: when we checked this against our own ratings rather than the market, we found our model goes the other way and over-rates the long-absence horse. That is exactly the sort of thing our public record is for, and we are correcting for it.
04First-time blinkers, and other tweaks
A trainer fitting blinkers for the first time is the most talked-about "we are trying something" signal. The data matches the old hands' view almost exactly: first-time blinkers win fewer races than a horse with no change, but offer slightly better value, and the real edge is yard-specific, some trainers fire with the switch, most do not.
| Headgear | Win % | A/E | Return at SP |
|---|---|---|---|
| First-time blinkers | 9.9% | 0.91 | -13% |
| No headgear | 11.6% | 0.89 | -22% |
Two more quiet positives from the same stable of ideas. A horse having its first run for a new yard beats a horse staying put (A/E 0.92 versus 0.89): a trainer who has just taken on a recruit often has a plan. And a horse dropping in class is a shade under-priced (A/E 0.91), placed somewhere easier for a reason. A change of jockey, on its own, tells you nothing: it is usually just who was available.
05Horses for courses
Does "this yard just goes well here" hold up? We flagged every trainer-and-track pair that had a real record in the past, then followed only their runners at that course in a later, unseen period. The specialism partly survived: those runners returned an A/E of 0.92 against 0.88 for the average runner. A genuine, small signal, not a profit.
What it all means
Put together, the human trends are real but modest. The crowd's instincts are mostly right (the stable elect is the one; the yard with a plan does a little better), the market is efficient enough to price almost all of it, and the one widely-held belief that is flat wrong is "follow the money", which at SP is a reliable way to lose faster.
We do not turn any of this into tips, and we are not claiming an edge over the market, because honestly we do not have one. What these trends do is feed the one thing we are good at: giving every runner a fair, honest chance of winning, and publishing whether those chances hold up. That is the whole product.
All figures from The Overlay's database of UK and Irish results, measured as A/E versus Starting Price (actual wins over expected wins). Retrospective analysis for information, not a tip and not a betting system. 18+ · BeGambleAware.org.