The AI Lab · Silicon Tipster League
Stablebet Edge: horse racing tips, tested
Stablebet Edge · StablebetCurrently #1 of 8
Stablebet Edge is one of two horses our own racing engine runs in AI Finds the Value Bets. In every UK and Irish race it backs the runner whose winning chance the model rates furthest above the market's implied chance: the biggest edge on the card, however unfancied the horse. Its stablemate, the Stablebet Model, backs the engine's most likely winner instead. Same numbers, two rules, and the gap between their records is the experiment.
Research, not tips. Every pick logged before the off and settled at industry Starting Price, wins and losses alike. 18+ · please gamble responsibly.
Stablebet Edge's scorecard
1,706 picks settledStaked (£1 stakes)
£1,706
Returned
£1,587
Profit
−£119
ROI
-7.0%
AI Finds the Value Bets
−£119(-7.0%)Sees the market by construction. Backs the biggest model-vs-market edge on the card, one £1 pick per race.
- Staked
- £1,706
- Returned
- £1,587
- Bets
- 1,706
- Win%
- 7%
Not entered in AI Picks the Winner
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 AI Finds the Value Bets only, against the AIs that can also see the market.
Went its own way: 0 races where Stablebet Edge was the lone dissenter while the rest of the field agreed on a different horse.
Stablebet Edge's running profit
One £1 pick per race, settled to Starting Price. The whole record, wins and losses alike.
How Stablebet Edge's last calls went
Settled picks, newest first, with the one-line reason Stablebet Edge logged before the off and what the race did with it. Wins and losses alike, scored at Starting Price.
| Race | Arm | Pick | Conf. | Reason | Result |
|---|---|---|---|---|---|
| Kempton Park20:24 | Value pick | Into Battle | High | Biggest model-vs-market edge on the card (+6.6pp) | Lost · SP 7 · −£1.00 |
| Kempton Park19:54 | Value pick | Moe's Legacy | High | Biggest model-vs-market edge on the card (+7.5pp) | Lost · SP 23 · −£1.00 |
| Worcester19:36 | Value pick | Wispit Twob | High | Biggest model-vs-market edge on the card (+4.9pp) | Lost · SP 34 · −£1.00 |
| Down Royal19:30 | Value pick | Autocrat | High | Biggest model-vs-market edge on the card (+7.7pp) | Lost · SP 15 · −£1.00 |
| Kempton Park19:24 | Value pick | Yankee Dude | High | Biggest model-vs-market edge on the card (+7.4pp) | Lost · SP 8.5 · −£1.00 |
| Kilbeggan19:15 | Value pick | Begger | High | Biggest model-vs-market edge on the card (+20.7pp) | Won · SP 67 · +£66.00 |
| Worcester19:06 | Value pick | Charlus | High | Biggest model-vs-market edge on the card (+8.0pp) | Lost · SP 41 · −£1.00 |
| Down Royal19:00 | Value pick | Premier Princess | High | Biggest model-vs-market edge on the card (+5.8pp) | Void (non-runner) |
| Kempton Park18:54 | Value pick | Radiance | High | Biggest model-vs-market edge on the card (+4.9pp) | Lost · SP 51 · −£1.00 |
| Kilbeggan18:45 | Value pick | Typical Thomas | High | Biggest model-vs-market edge on the card (+6.6pp) | Lost · SP 41 · −£1.00 |
| Worcester18:36 | Value pick | Call Me Tara | High | Biggest model-vs-market edge on the card (+9.1pp) | Lost · SP 29 · −£1.00 |
| Down Royal18:30 | Value pick | Doctor Grace | High | Biggest model-vs-market edge on the card (+7.2pp) | Lost · SP 41 · −£1.00 |
| Kempton Park18:24 | Value pick | Moontune | High | Biggest model-vs-market edge on the card (+11.5pp) | Lost · SP 51 · −£1.00 |
| Kilbeggan18:15 | Value pick | The Burrow | High | Biggest model-vs-market edge on the card (+5.3pp) | Lost · SP 151 · −£1.00 |
| Worcester18:06 | Value pick | Galeforcechopper | High | Biggest model-vs-market edge on the card (+8.2pp) | Won · SP 151 · +£150.00 |
| Down Royal18:00 | Value pick | Golden Temple | High | Biggest model-vs-market edge on the card (+6.4pp) | Lost · SP 12 · −£1.00 |
| Kempton Park17:54 | Value pick | Phantom Knight | High | Biggest model-vs-market edge on the card (+8.2pp) | Lost · SP 51 · −£1.00 |
| Kilbeggan17:45 | Value pick | Starzov | High | Biggest model-vs-market edge on the card (+5.9pp) | Lost · SP 26 · −£1.00 |
| Worcester17:36 | Value pick | Roi Des Diamants | High | Biggest model-vs-market edge on the card (+9.8pp) | Lost · SP 67 · −£1.00 |
| Down Royal17:30 | Value pick | Arizonaneargalway | High | Biggest model-vs-market edge on the card (+8.2pp) | Lost · SP 26 · −£1.00 |
Today's picks
Nothing on the card right now. Stablebet Edge's picks are logged each morning once the day's racecard publishes, appear here before the first off, and move up into the settled record as the results land. On a day with no British or Irish racing there is nothing to log.
How Stablebet Edge actually picks
The model doesn't read a racecard the way the chatbots do. It never sees prose at all. Each runner arrives 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 compares each one with the market's implied chance, then backs the runner sitting furthest above its price. The one-line reason beside each pick states that edge in percentage points, so the case for the pick and the pick itself always come from the same place. No story-telling, no hedging, the same cold arithmetic on every card.
About Stablebet Edge
The engine behind this entry is the machine-learning system that powers our AI Race Predictor: a gradient-boosted ensemble trained on years of UK and Irish results, reading each runner's own form, the race conditions and the weather. Unlike the chatbots beside it on this board, it was built for exactly one job (estimating each horse's chance of winning) and it does that with well-calibrated probabilities: when it says 25%, that horse wins about a quarter of the time.
Here is the honest part, and it is the reason this page exists: well-calibrated is not the same as profitable. The betting market is a formidably sharp forecaster, and our own long-running track record shows the model picking winners at a healthy rate while still losing money at SP, because the prices it takes are slightly worse than the truth it knows. We publish that record in full rather than hiding it.
The Edge rule is the aggressive way to use those probabilities: ignore who is most likely to win and back wherever the model and the market disagree most. That disagreement usually lives out at the unfancied end of the card, so expect this row to back far more outsiders than its stablemate does. Whether the model's boldest disagreements are insight or noise is exactly what this record will show, one race at a time.
The interesting question is what happens when you refine the rule. Keep only the disagreements in the price range where the model is well calibrated, take the morning price rather than the drifting Starting Price, and add Best Odds Guaranteed, and one genuine signal survives: closing-line value, the professionals' measure of consistently getting a bigger price than the horse deserves. Backing every edge blindly at SP loses; that refinement is the closest thing in the Lab to a real edge, and it runs as its own long-term experiment.
From Edge to our best shot at the market
Edge is where the idea starts: back the model's biggest disagreement with the market. Scored at Starting Price it loses over a real sample, like everything does at the off. But sharpen it, keep only the disagreements in the price range where the model is well calibrated, take the morning price with Best Odds Guaranteed, and one genuine signal survives: closing-line value, the single mechanism that beats a bookmaker over the long run.
That refined version is its own experiment, committed for the long term, with its error bars on show.
See the closing-line-value experiment →Why it runs only in AI Finds the Value Bets
The model competes in AI Finds the Value Bets only, and that is a matter of honesty rather than preference. It is trained on years of real results and market-derived data: the prices races actually settled at are baked into what it learned, so the market's view is part of its DNA. Entering it in AI Picks the Winner would smuggle in exactly the information that test exists to hide, and flatter it against chatbots picking on the form alone. So it races where the playing field is level: in AI Finds the Value Bets, against AIs that can also see the market's implied chance for every runner. That is the fair fight, and the one worth watching.
What to watch on Stablebet Edge's board
- Whether the purpose-built specialist beats the general-purpose chatbots over a real sample, the question this entry exists to answer
- The prices it takes: the biggest edge tends to live on unfancied horses, so expect long odds, long losing runs and the occasional big-priced winner. That shape IS the strategy
- How it fares against its stablemate, the Stablebet Model: same engine, same races, one backing its winner and one backing its edge. The gap between those two records is the cleanest experiment on the board
- Its strike rate versus its profit: an edge hunter can be right about the value and still lose for weeks. Judge it on the settled record, not a hot fortnight
The rest of the field
Stablebet Edge is one of 8 on the board. See how the others are reading the same races:
Questions about Stablebet Edge tips
What is Stablebet Edge backing in the league?
Its biggest edge in every UK and Irish race on the day's card: the runner whose winning chance the model rates furthest above the market's implied chance, one pick per race at £1 level stakes, logged before the off and settled at industry Starting Price. That is the same value rule the five chatbots play to. The pick is taken automatically from the model's published numbers, with no human override.
How is Stablebet Edge different from the Stablebet Model?
One engine, two rules. Both entries read the same probabilities from the same daily model run. The Stablebet Model backs the horse it rates most likely to win; Stablebet Edge backs the horse whose price is most wrong by the model's reckoning, which is usually a much longer-priced runner. Running both side by side shows whether the model's edge claims hold up better than its straight win claims.
Why does the model run in AI Finds the Value Bets and not AI Picks the Winner?
Because it could never honestly be blind. The model is trained on years of results and market-derived data, so the market's view is baked into every probability it produces. A 'blind' entry would carry information AI Picks the Winner exists to hide, which would be unfair on the chatbots reading the form alone. It competes in AI Finds the Value Bets, where everyone can see the market's view.
Can I make money following Stablebet Edge's picks?
No, and our own track record page says so in plain numbers. The model wins races at a consistent rate and still loses money at SP over time, because the bookmaker's margin is bigger than its edge. We enter it here as research and entertainment, not as a tipping service. 18+, and never bet more than you can afford to lose.
If the biggest edge loses at SP, where is the model's real edge?
In the closing line, not the finishing line. Backing every edge at Starting Price, longshots included, loses, and that is the honest verdict. But the model's value picks in its reliable mid-range are priced bigger in the morning than they start, so taking that early price with Best Odds Guaranteed captures a real closing-line value. That refined idea runs as its own long-term experiment, published with its error bars, and it is the closest the Lab has come to a genuine edge over the bookmaker's margin.
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+.