Top 10 AI horse racing predictors (UK & Ireland), rated
We rate 25 AI horse racing prediction tools covering UK and Irish racing. 13 publish enough of a level-stakes record for a full Performance score and 1 a provisional one; the strongest published record is Turfs Edge Tipping (TeT) (Performance 59/100, Transparency 40/100, recomputed by us). 6 of the 13 are in profit at level stakes and 7 are self-reported, so check the evidence line under each one. Below are the top 10, ranked by Performance, then Transparency score, then settled bets.
Can AI predict horse racing?
AI is good at ranking horses: racing models look for signal in form, sectional times and market data, and they can rank runners far better than chance. The harder test is the price: the starting price and Betfair markets already rank runners well, so what matters is a record of bets at the prices you could get, over hundreds of races.
The promising part is that some tools publish records you can check. Of the 25 AI horse racing tools we rate, 13 publish enough for a full Performance score and 6 of those are in profit at level stakes (median ROI -1.40%); 11 don't publish enough to score yet. 7 of the 13 are self-reported and 2 independently proofed, so check the evidence line under each one.
Independent evidence is still early: there is little yet showing AI beating the racing market's prices. We haven't found an independent, peer-reviewed test of a UK racing tool's betting returns; the studies below use Japanese racing and measure ranking, not profit. Past results don't predict future returns.
What independent research says (2 studies)
- Leakage-aware pre-event machine learning evaluation for horse-race prediction under temporal validation (Sugiura, 2026; Out-of-sample test, 4,556 races)
A machine-learning model trained on Japanese (JRA) flat races from 2015 to 2022 was tested on 4,556 later races (January 2025 to May 2026). It ranked winners with a ROC AUC of 0.754 (1.0 is perfect, 0.5 is chance). AUC measures how well runners are ranked, not whether bets at the available odds would make money. - Leakage-aware race-level upset-risk diagnostics for prospective horse-race prediction (Sugiura, 2026; Out-of-sample test, 4,556 races)
A model built only from pre-race information flagged races likely to produce an upset (ROC AUC 0.646 on the same 2025–26 test races): in the 10% of races it rated most upset-prone, 68% produced an upset, against 52% overall.
Summaries checked against each source on 19 Oct 2026: the full paper where we could read it, otherwise the published abstract. Studies use their own samples and methods, so they aren't directly comparable with the records we score, and none tests the specific tools we list.
How this ranking works
- Ranked by Performance score, then Transparency score, then settled bets.
- A full Performance score needs a published level-stakes record of 200+ settled bets; 50 to 199 is provisional and not ranked.
- Every record is labelled self-reported, recomputed by us or independently proofed; self-reported records count half their bets.
Top 10: tools with a full Performance score
Order: Performance score (high to low), then Transparency score, then settled bets. Full scores need 200+ settled bets at level stakes. Of these 10, 5 are self-reported and 2 independently proofed: check the evidence line under each. Tiers explained.
- 1
Turfs Edge Tipping (TeT)
+15.23% ROI over 419 level-stakes bets (13 Aug 2026 – results page read 2 Nov 2026), recomputed by us. Price basis: Published (advised) odds; win only.
↻ Recomputed by us⚑ Advised odds, may be hard to matchRecord period: 13 Aug 2026 – results page read 2 Nov 2026Checked 2 Nov · next re-check 9 Nov59PerformanceProfitable record40TransparencyMedium transparency - 2
RaceBot UK
+6.70% ROI over 3,475 level-stakes bets (November 2025 – 17 Oct 2026), independently proofed. Price basis: Advised bookmaker odds at time of posting.
✓ Independently proofed⚑ Advised odds, may be hard to matchRecord period: November 2025 – 17 Oct 2026 (Tipstrr stats page, horse racing, all time)Checked 1 Nov · next re-check 8 Nov57PerformanceProfitable record90TransparencyHigh transparency - 3
ProBetAI
+9.60% ROI over 709 level-stakes bets (Tipstrr last-12-months view, to 4 Nov 2026), independently proofed. Price basis: Advised bookmaker odds at time of posting.
✓ Independently proofed⚑ Advised odds, may be hard to matchRecord period: Tipstrr last-12-months view, to 4 Nov 2026Checked 4 Nov · next re-check 11 Nov57PerformanceProfitable record80TransparencyHigh transparency - 4
FormPick
+1.78% ROI over 12,955 level-stakes bets (28 Oct 2025 – 28 Oct 2026), recomputed by us. Price basis: SP (decimal).
↻ Recomputed by usRecord period: 28 Oct 2025 – 28 Oct 2026 (365 days)Checked 29 Oct · next re-check 5 Nov52PerformanceProfitable record50TransparencyMedium transparency - 5
EdgeDigger
+4.90% ROI over 516 level-stakes bets (1 Aug – 2 Nov 2026), self-reported. Price basis: Betfair Exchange price taken from 10:30 on race morning, before commission.
• Self-reported⚑ Advised odds, may be hard to matchRecord period: 1 Aug – 2 Nov 2026 (Betfair-price months)Checked 3 Nov · next re-check 10 Nov52PerformanceProfitable record50TransparencyMedium transparency - 6
RaceChat AI
+1.53% ROI over 249 level-stakes bets (27 days to 25 Oct 2026), self-reported. Price basis: Exchange SP less 5% commission.
• Self-reportedRecord period: 27 days to 25 Oct 2026Checked 26 Oct · re-check overdue50PerformanceBreak-even record75TransparencyHigh transparency - 7
OddsLens
-1.40% ROI over 7,443 level-stakes bets (9 Apr 2026 – 4 Nov 2026), self-reported. Price basis: Best price available when each pick was logged before the off, £1 level stakes.
• Self-reported⚑ Advised odds, may be hard to matchRecord period: 9 Apr 2026 – 4 Nov 2026 (all time)Checked 4 Nov · next re-check 11 Nov48PerformanceLosing record100TransparencyHigh transparency - 8
Race-AI
-2.73% ROI over 1,347 level-stakes bets (11 Sep – 16 Oct 2026), self-reported. Price basis: Betfair SP (commission treatment not stated).
• Self-reportedRecord period: 11 Sep – 16 Oct 2026Checked 3 Nov · next re-check 10 Nov48PerformanceLosing record0TransparencyLow transparency - 9
realtime.tips
-2.24% ROI over 38,192 level-stakes bets (6 Mar – 17 Oct 2026), self-reported. Price basis: Betfair SP with SP fallback, no commission.
• Self-reportedRecord period: 6 Mar – 17 Oct 2026Checked 19 Oct · re-check overdue47PerformanceLosing record100TransparencyHigh transparency - 10
HorsePicker
-7.70% ROI over 275 level-stakes bets (18 Aug – 16 Oct 2026), recomputed by us. Price basis: SP, win only.
↻ Recomputed by usRecord period: 18 Aug – 16 Oct 2026 (live since rule adopted)Checked 3 Nov · next re-check 10 Nov47PerformanceLosing record50TransparencyMedium transparency
The rest, in brief
- 3 more with a full Performance score: RacingOracle, Juvenile Race Analytics, StableBet AI Race Predictor.
- 1 with a small sample (provisional, 50 to 199 settled bets, not ranked): The Overlay.
- 11 don't publish enough for a Performance score yet: see the not enough data yet list.
Every tool, with both scores and the evidence behind them: all 122 AI tool ratings · the scoreboard.
With a free option (14)
Tools whose own pricing page lists a free tier, free tips or free use. Free doesn't mean better or worse: check the scores. Prices are quoted on each entry page.
| Tool | What's free | Performance |
|---|---|---|
| Turfs Edge Tipping (Te | Free tip of the day | 59 Profitable record |
| Race-AI | Free (3 selections/day) | 48 Losing record |
| realtime.tips | Free tips and public record | 47 Losing record |
| Horse | Free to use | 47 Losing record |
| Racing | Free to use | 39 Losing record |
| Stable | Free to use | 34 Losing record |
| The Overlay | Free to use | 40 Provisional (small sample) |
| At The Races Predictor | Free to use | Not enough data yet |
| Hoofs Analysis | Free to use | Not enough data yet |
| Horse | Free to use | Not enough data yet |
| Outstrider | Free option (see entry) | Not enough data yet |
| Post | Free to use | Not enough data yet |
| Probatrix | Free rankings | Not enough data yet |
| The Racing Bot | Free (scores hidden) | Not enough data yet |
Accuracy and calibration, as stated (4)
Some horse racing tools publish how often they pick the right result, or how well calibrated their probabilities are, sometimes next to the betting market's own figure. We quote these as stated by each tool, unverified. They don't feed the Performance score: accuracy alone can't show whether bets at the available odds make money.
| Tool | Measure | As stated | Market figure (as stated) | Sample | Source |
|---|---|---|---|---|---|
| EdgeDigger | Calibration, predicted v actual win rate | ±0.4pp average gap | No market comparison stated | 39,687 runners, 131 racing days | Source checked 20 Oct |
| Juvenile Race Analytics | Log loss, all races (model v market) | 2.11 v 1.93 | Market more accurate (as stated) | 1,439 races | Source checked 25 Oct |
| StableBet AI Race Predictor | Brier score, all races (model) | 0.0964 | Lower than the model’s on the same races (as stated) | 10,852 races | Source checked 19 Oct |
| HK Racing Model PRO | Out-of-sample test, top-rated blend at level stakes (final HKJC odds) | −15.1% (favourite −20.9%, model alone −17.8%) | Market log loss 2.050 v model 2.269 (market better, as stated) | 854 races, Sep 2025 – Jul 2026 | Source checked 19 Oct |
Brier score: lower is better; it measures how close forecast probabilities were to what happened. A tool beats the market on this measure only if its Brier score is lower than the market's on the same matches.
How we check the records
The transparency checklist sets out what counts as evidence for each Transparency criterion, the same for every listing. The AI tipping transparency report 2026 shows how many of the AI tools and tipsters we rate meet each criterion.