Results & Performance
The edge, validated on real Tote dividends. Updated 2026-06-21 12:04.
Combo Tricasts — The Edge
Real Tote dividends • 2026-02-18 to 2026-06-20 • the filter we actually bet (chase handicaps + small-field flat non-handicaps).
| Segment | Bets | Strike | Tricast P&L | ROI |
|---|---|---|---|---|
| Chase handicap 7-8 | 45 | 20% | +£479 | +177.4% |
| Flat non-hcap 5-6 | 110 | 15.5% | +£517 | +78.3% |
| Chase handicap 5-6 | 55 | 12.7% | +£11 | +3.2% |
Returns are heavy-tailed — a few big dividends carry the edge, with long losing runs in between. Stake small and bankroll for variance.
Yesterday by Rank — 2026-06-19
How the model's rank 1/2/3 picks finished — flat £1 win at SP, settled on results (single day, high variance). W = won, P = placed.
Combo Tricast P&L by Month
Real Tote dividends, the filter we bet. Winning months are big; losing months are bounded — this is the heavy tail in numbers, not a smooth grind.
| Month | Bets | Strike | P&L | ROI |
|---|---|---|---|---|
| 2026-02 | 29 | 17.2% | +£241 | +138.3% |
| 2026-03 | 57 | 17.5% | +£758 | +221.7% |
| 2026-04 | 34 | 8.8% | £-71 | -34.8% |
| 2026-05 | 66 | 21.2% | +£195 | +49.2% |
| 2026-06 | 24 | 4.2% | £-116 | -80.7% |
Singles Strategy Tracking
Rank-1 singles strategies settled daily at SP. The two marked LIVE are in the betting portfolio (WIN only, Tote pool); the rest are tracked baselines.
| Strategy | Bets | Strike | Staked | P&L | ROI |
|---|---|---|---|---|---|
| R1 Chase + JT Form ≥ 0.2LIVE | 78 | 35.9% | £350.00 | +£57.55 | +16.4% |
| R1 Chase + Gap ≥ 0.3LIVE | 124 | 31.5% | £515.00 | +£74.07 | +14.4% |
| R1 Chase + Odds 2-6 | 257 | 24.9% | £1262.50 | £-67.89 | -5.4% |
| R1 Chase | 486 | 21% | £2015.00 | £-234.09 | -11.6% |
| R1 Field ≤ 4 Runners | 158 | 33.5% | £735.00 | £-118.15 | -16.1% |
| Month — live strategies | Bets | P&L | ROI |
|---|---|---|---|
| 2026-02 — R1 Chase + JT Form ≥ 0.2 | 9 | +£28.50 | +76% |
| 2026-03 — R1 Chase + JT Form ≥ 0.2 | 29 | +£28.75 | +21.3% |
| 2026-04 — R1 Chase + JT Form ≥ 0.2 | 18 | £-0.75 | -1% |
| 2026-05 — R1 Chase + JT Form ≥ 0.2 | 16 | +£9.00 | +12% |
| 2026-06 — R1 Chase + JT Form ≥ 0.2 | 6 | £-7.95 | -26.5% |
| 2026-02 — R1 Chase + Gap ≥ 0.3 | 22 | +£22.44 | +24.9% |
| 2026-03 — R1 Chase + Gap ≥ 0.3 | 60 | +£37.75 | +15.6% |
| 2026-04 — R1 Chase + Gap ≥ 0.3 | 15 | +£31.78 | +55.3% |
| 2026-05 — R1 Chase + Gap ≥ 0.3 | 21 | £-24.44 | -25.1% |
| 2026-06 — R1 Chase + Gap ≥ 0.3 | 6 | +£6.55 | +23.8% |
Nightly Simulated Tracker
Every selection settled on real results / real Tote dividends, no money required. 2026-02-02 to 2026-06-20. Singles: £5 WIN (dec odds < 6.0) / £2.50 E/W (>= 6.0). Combos: 10p/line · 3 picks · 6 perms = 60p per race, on every race.
| Strategy | Staked | P&L | ROI |
|---|---|---|---|
| Rank 1 single | £14085.00 | £-1993.14 | -14.2% |
| Rank 2 single | £13305.00 | £-2940.86 | -22.1% |
| Rank 3 single | £12977.50 | £-2899.98 | -22.3% |
| Combo Exacta (10p) | £2695.20 | £-1279.04 | -47.5% |
| Combo Trifecta (10p) | £2633.40 | £-1379.38 | -52.4% |
These combos are unfiltered (every race) — they lose, as expected. The profitable edge is the filtered combo set shown above (chase handicaps + small-field flat non-handicaps). Straight singles lose at SP too. This table is the honest full-coverage baseline.
US Model — Live Tracker
US-specific LightGBM ranker, trained on 1,623 races (2023–2025). OOS backtest showed +10% ROI in the 8–9 runner sweet spot. Live picks settled flat £1 win at SP from 18 picks across 6 races.
| Cohort | Wins/Bets | Strike | P&L | ROI |
|---|---|---|---|---|
| Rank 1 | 2/5 | 40% | +£3.75 | +75% |
| Rank 2 | 1/5 | 20% | +£0.30 | +6% |
| Rank 3 | 0/5 | 0% | £-5.00 | -100% |
| Sweet spot (R1, 8–9 runners) | 2/5 | 40% | +£3.75 | +75% |
Live sample — small until 3+ months of forward data accumulate. Combos (US Tote exacta/trifecta) are being collected; not yet enough sample to report.
Methodology
- Combo edge is settled on real Tote Exacta/Trifecta dividends pulled from the Tote API, across 2,000+ races over both jumps and flat seasons — not an SP approximation.
- We only bet the two segments that survive that test: chase handicaps and small-field (≤6 runner) flat non-handicaps. Hurdles, flat handicaps, large/tiny fields and short-priced favourites all lose and are dropped.
- Yesterday by rank shows the straight win/EW result of the model's rank 1/2/3 picks for the latest settled day. Straight win betting is not the edge — it's shown for transparency.
- Tricast P&L is heavy-tailed; size stakes for a multi-month dry spell.