The AI Lab · Silicon Tipster League
Stablebet Model: horse racing tips, tested
Stablebet Model · StablebetCurrently #7 of 8
The Stablebet Model is our racing engine's straight answer to the only question most punters ask: who wins? In every UK and Irish race it backs the runner the model rates most likely to win, logged before the off and settled at SP. Its stablemate, Stablebet Edge, reads the same numbers and backs the biggest disagreement with the market instead. One engine, two rules, both on the record.
Research, not tips. Every pick logged before the off and settled at industry Starting Price, wins and losses alike. 18+ · please gamble responsibly.
Stablebet Model's scorecard
38 picks settledStaked (£1 stakes)
£38.00
Returned
£11.63
Profit
−£26.37
ROI
-69.4%
Value Test
−£26.37(-69.4%)Sees the market by construction. Backs the runner the model rates most likely to win, one £1 pick per race.
- Staked
- £38.00
- Returned
- £11.63
- Bets
- 38
- Win%
- 13%
No Form Test entry
Our model is trained on market-derived data, so the market's view is baked into every rating it makes. A 'blind' version would be a fiction, which is why it competes in the Value Test only, against the AIs that can also see the market.
Went its own way: 3 races where Stablebet Model was the lone dissenter while the rest of the field agreed on a different horse.
Stablebet Model's running profit
One £1 pick per race, settled to Starting Price. The whole record, wins and losses alike.
Stablebet Model's recent calls
The actual picks and one-line reasons Stablebet Model logged before each off, newest first. Read the logic and judge it for yourself. The board settles the argument at Starting Price.
| Race | Arm | Pick | Conf. | Reason | Result |
|---|---|---|---|---|---|
| Doncaster21:00 | Value Test | Pergamon | Medium | Model's top-rated runner on the card (32.4% chance) | pending |
| Nottingham20:43 | Value Test | Revelio | Medium | Model's top-rated runner on the card (21.7% chance) | pending |
| Doncaster20:25 | Value Test | Electric Lightning | Low | Model's top-rated runner on the card (12.4% chance) | pending |
| Nottingham20:08 | Value Test | Thursday Girl | Medium | Model's top-rated runner on the card (26.7% chance) | pending |
| Doncaster19:50 | Value Test | Tai Hang Pegasus | Medium | Model's top-rated runner on the card (23.3% chance) | pending |
| Nottingham19:33 | Value Test | Del Corso | High | Model's top-rated runner on the card (42.5% chance) | pending |
| Doncaster19:15 | Value Test | Big Apple Jack | Low | Model's top-rated runner on the card (14.8% chance) | pending |
| Nottingham19:00 | Value Test | Control Room | Low | Model's top-rated runner on the card (17.2% chance) | pending |
| Doncaster18:45 | Value Test | First Time | Low | Model's top-rated runner on the card (14.7% chance) | pending |
| Nottingham18:30 | Value Test | Whiskey Bent | Medium | Model's top-rated runner on the card (27.4% chance) | pending |
| Doncaster18:15 | Value Test | Sanbona Warrior | Low | Model's top-rated runner on the card (15.2% chance) | pending |
| Nottingham17:55 | Value Test | No Knee Never | Medium | Model's top-rated runner on the card (20.2% chance) | pending |
About Stablebet Model
This entry is the prediction engine behind our AI Race Predictor doing exactly what people assume it does: naming its most likely winner. The engine is a gradient-boosted ensemble trained on years of UK and Irish results, and its probabilities are well calibrated, meaning a horse it gives 25% wins about a quarter of the time. When this row backs a horse, that is the engine's genuine best guess at the winner, taken automatically with no human override.
Calibration is not profit, and we say so on every page that shows these numbers. A model can be right about the percentages and still lose money at SP, because the market's prices already carry most of the same knowledge plus the bookmaker's margin. Our published track record shows exactly that pattern, and this entry puts it to the test in the most public way possible.
Why enter the same engine twice? Because the two rules answer different questions. This row tests the model as a tipster: can its most likely winner beat the chatbots and the favourite? Stablebet Edge tests it as a value hunter: are its disagreements with the market worth anything? A widely-held suspicion in racing modelling is that a model's biggest claimed edges sit where it is least reliable. Running both rules on the same races, live, is how we find out.
How Stablebet Model reads a race
No prose, no gut feel. Each runner reaches the model as numbers: career form figures, official rating against the field, course and distance record, the going, the class, the weather. Out comes a win probability for every horse, and this entry simply takes the highest one. Expect its pick to be the market favourite a lot of the time; the model and the market are reading the same races, and most favourites are favourites for good reasons. The interesting races are the ones where this row quietly steps off the favourite, because those are the cards where the numbers and the crowd genuinely disagree about the most likely winner.
Why it runs in The Value Test only
Like its stablemate, the Stablebet Model runs in the Value Test only. The engine is trained on market-derived data, so the market's view is stitched into every probability it produces and a blind entry would be a fiction. It sits in the Value Test field beside the chatbots that are shown the market's implied chances, which keeps the comparison honest: everyone in this test can see what the crowd thinks, and the question is what each of them does with it.
What to watch on Stablebet Model's board
- How often its top-rated pick is simply the market favourite, and what happens on the days it steps off the favourite. Those disagreement races are where a specialist proves itself or doesn't
- The head-to-head with Stablebet Edge: same engine, same races, win rule against edge rule. If the edge rule keeps finding losers while this row grinds, that says something real about where models are weakest
- Strike rate versus profit: this row should win far more often than Stablebet Edge and can still lose money at SP. The gap between winning often and winning money is the bookmaker's margin, live
- Whether the purpose-built engine's best guesses beat the chatbots that were merely asked for value
The rest of the field
Stablebet Model is one of 8 on the board. See how the others are reading the same races:
Questions about Stablebet Model tips
What does the Stablebet Model back in the league?
The runner our racing engine rates most likely to win, in every UK and Irish race on the day's card: one pick per race at £1 level stakes, logged before the off and settled at industry Starting Price. The pick is read straight from the model's published probabilities for the day, with no human override.
How is this different from Stablebet Edge?
Same engine, different rule. This entry backs the horse with the highest win probability; Stablebet Edge backs the horse whose price sits furthest below what the model thinks it should be, which usually means a longer-priced runner. This row will win much more often; whether either wins money is what the board exists to show.
Won't it just back the favourite every time?
Often, yes, and that is informative rather than embarrassing. The model and the market study the same races, so their most likely winners frequently agree. The board's 'Backed the fav' column tracks exactly how often, and the races where the model steps off the favourite are the ones that reveal whether it knows something the crowd doesn't.
Can I make money following the Stablebet Model's picks?
No. Our track record page shows the model picking winners at a consistent rate and still losing at SP, because the bookmaker's margin outweighs its accuracy. This entry is research and entertainment, not a tipping service. 18+, please gamble responsibly, and never stake more than you can afford to lose.
This page sits inside the AI Lab, where we test whether any betting system makes money (across 26,000+ races, none of them do), and ask the bigger question in does following an AI tipster work?
Gamble responsibly.This page is research and entertainment, not betting advice. No AI here beats the bookmaker's margin, and nothing on it is a signal to stake. Betting should never be a way to make money. If it is affecting you or someone you know, free and confidential support is at BeGambleAware.org. 18+.