How Often Do Horses Trading at 1.10 In-Running Get Beaten?
Across 1,260,853 Australian and New Zealand runners, I test how often a horse trading odds-on in-running still gets beaten, and where the price gets it wrong.
Every regular on the betting exchange has watched a horse trade at 1.10 halfway up the straight and then get swallowed in the last fifty metres. It feels as though it happens all the time. I wanted to know whether it really does, or whether memory simply keeps the painful ones. The same questions about tempo and finishing speed sit behind my horse racing predictions page, so this one was hard to resist.
The headline answer is calmer than the folklore. Of 128,893 runners that touched 1.10 or shorter in-running, 8.9% were beaten, against the 9.1% that price implies. At that level the market is almost exactly right. The interesting part is at the very bottom of the ladder, where horses that traded at 1.01 lost 1.5% of the time, against 1.0% implied.
1Defeat rates climb steadily as the lowest price lengthens
This chart groups each horse by the shortest price it was matched at in-running, then shows how many of them lost. A horse that touched 1.03 counts in every band from 1.05 upwards, so each bar includes the horses in the bars before it.
The shape is orderly. From 1.5% at 1.01 it rises through 2.6% at 1.02 and 5.5% at 1.05 to the highlighted 8.9% at 1.10, then 16.8% at 1.20 and 48.4% at 2.00. Nothing about it suggests short-priced horses collapse more often than they should; it reads like a market doing its job.
View the data: beaten after trading at or below each price
| Share of runners that lost, by lowest in-running price matched | |
|---|---|
| 1.01 or shorter | 1.5% |
| 1.02 or shorter | 2.6% |
| 1.05 or shorter | 5.5% |
| 1.10 or shorter | 8.9% |
| 1.20 or shorter | 16.8% |
| 1.30 or shorter | 22.8% |
| 1.50 or shorter | 32.2% |
| 2.00 or shorter | 48.4% |
What I take from this is perspective. Roughly one in eleven horses at 1.10 gets beaten, which is enough to see it most weekends without it being a flaw in the market.
2The very shortest prices lose more often than implied
Here each row sets the defeat chance the price implies against how often those horses actually lost, with 95% ranges shown. Where the two dots sit apart beyond those ranges, the price has been consistently wrong.
At 1.02 the implied figure is 2% and the actual is 2.6%; at 1.05 it is 4.8% implied against 5.5% actual. Those horses were beaten more often than their price allowed. From 1.10 outward the gap closes or flips slightly, with 1.50 showing 32.2% lost against 33.3% implied.
Implied defeat chance at price X = 1 โ 1/X. 95% ranges shown.
View the data: chance of defeat: price-implied vs actual
| Implied by that price | Actually lost | 95% range | |
|---|---|---|---|
| Traded 1.02 (110,073 runners) | 2.0% | 2.6% | ยฑ0.1 |
| Traded 1.05 (120,436 runners) | 4.8% | 5.5% | ยฑ0.1 |
| Traded 1.10 (128,893 runners) | 9.1% | 8.9% | ยฑ0.2 |
| Traded 1.20 (145,448 runners) | 16.7% | 16.8% | ยฑ0.2 |
| Traded 1.30 (159,244 runners) | 23.1% | 22.8% | ยฑ0.2 |
| Traded 1.50 (184,022 runners) | 33.3% | 32.2% | ยฑ0.2 |
| Traded 2.00 (245,251 runners) | 50.0% | 48.4% | ยฑ0.2 |
My reading is that the crowd overrates near-certainty. Once a horse looks home, people pile in at 1.02 and 1.03, but a stumble, a late swoop or a horse that idles in front still happens slightly more often than those prices admit.
3Laying pays at 1.02 and 1.05, then turns negative
This model stakes one unit at exactly each threshold price the first time it traded, either backing it gross or laying it with 5% commission taken from winning lays. Lay returns are shown as a share of liability.
Laying at 1.05 returned +9.4% of liability, and at 1.02 the figure is larger still at 24.1. By 1.10 the picture reverses: laying returned โ6.7% while backing returned +0.2%. At 1.30, laying returned โ4.8% and backing +0.3%.
Assumes a bet was matched at exactly that price the first time it traded.
View the data: backing or laying at that price
| Back at the price (gross) | Lay at the price (after commission) | |
|---|---|---|
| At 1.02 | โ0.6% | +24.1% |
| At 1.05 | โ0.7% | +9.4% |
| At 1.10 | +0.2% | โ6.7% |
| At 1.20 | โ0.1% | โ3.6% |
| At 1.30 | +0.3% | โ4.8% |
| At 1.50 | +1.7% | โ6.6% |
| At 2.00 | +3.1% | โ5.6% |
The caveat matters more here than anywhere else. Prices at this level move in fractions of a second, many viewers watch a delayed picture, and a horse can jump straight past 1.05 without anyone being matched there, so this is a model of what was available on paper, not a system anyone can run.
4Trip makes only a small difference to late defeats
I expected distance to matter, because staying races give a leader longer to be run down. The heatmap splits each price band by race distance.
The effect is there but faint. At 1.10, sprints up to 1,100 m saw 8.5% beaten against 9.2% over 1,801 m or more, and at 1.50 the figures were 31.8% and 32.9%. The middle distances wobble: at 1.05, horses lost 5.6% over 1,101-1,400 m and 5.4% over 1,401-1,800 m.
View the data: beaten from a short in-running price, by distance
| โค1.05 | โค1.10 | โค1.20 | โค1.50 | |
|---|---|---|---|---|
| Up to 1,100 m | 5.1% | 8.5% | 16.3% | 31.8% |
| 1,101-1,400 m | 5.6% | 9.1% | 16.8% | 32.0% |
| 1,401-1,800 m | 5.4% | 8.5% | 17.0% | 32.7% |
| 1,801 m+ | 5.8% | 9.2% | 17.1% | 32.9% |
So the longer trips do produce slightly more late reversals, which fits what I see in staying races where the tempo builds late. But the gap is small enough that I would not build anything on trip alone.
Betting notes
- Respect 1.10 as fair. Horses at 1.10 lost 8.9% of the time against 9.1% implied, so neither side has a meaningful edge there. Treat a reversal at that price as normal racing, not a market error.
- Question the near-certainties. At 1.02 and 1.05 horses lost more often than their prices implied, 2.6% and 5.5% respectively. If you hold a lay position, those are the levels where the crowd tends to overpay.
- Allow for the picture delay. The +9.4% lay return at 1.05 assumes being matched at exactly that price, which a delayed feed often makes impossible. Count only the prices you can realistically get.
Quick answers
How often does a horse trading at 1.10 in-running lose?
In this sample of Australian and New Zealand races, 8.9% of horses that traded at 1.10 or shorter were beaten. That is slightly less often than the 9.1% the price implies.
Do horses trading at 1.01 ever get beaten?
Yes. Of 103,936 runners that traded at 1.01, 1.5% still lost, against the 1.0% that price implies.
Is laying short-priced horses in-running profitable?
In the model, laying at 1.05 returned +9.4% of liability after commission, but laying at 1.10 returned โ6.7%. Getting matched at exactly those prices is far from guaranteed, so treat it as an illustration rather than a system.
I'm Aisha Rahman, based in Dubai, and I write the horse racing tips at bettips.vip โ built on the going, the draw and the pace, not the price of the favourite.
All of Aisha Rahman's work โ
