One horse. One bet. Every day. Sly Man’s read on the day’s card — the shortlist, the price, and the case for the pick.
Browse the logbookA steamer with a 25% trainer in the last fortnight of a sprint series, at odds that still carry real value — that is the story of today's tip. The scan ran across 38 races and 291 runners declared on the day. From there the filter pipeline applied three successive cuts: first, the race-state check, which dropped nothing (all races live, none already resulted); second, the odds band — 2.0 to 50.0 decimal, stripping odds-on runners and any shot beyond 49/1; third, field size, keeping only races with 6–14 declared runners, removing the noise of tiny fields and the chaos of cavalry charges. Feature weights loaded are: market_move at 1.6 (the heaviest signal), speed at 1.5, form at 1.2, class_change at 1.0, trainer_strike and going both at 0.8, draw reserved for v2 and unavailable throughout.
The day started with 291 runners across 38 races. The race-state pass left all 291 intact. The odds-band cut removed 27 — either odds-on or priced beyond the band — leaving 264. The field-size filter then cleared the field down to 208 runners as the working universe. Those 208 went into the composite scoring engine, which ranked every survivor by weighted feature score.
The top of the board threw up six names. Darkened Edge led on a composite of 6.85/10. War Gaming sat second at 6.67/10 but immediately drew a flag — market-implied probability of 40.6% against an engine estimate of 34.0% gives a negative edge of -25.2%; the model thinks the market has overshot on this one. Baila Conmigo was third at 6.61/10, a strong market signal at Killarney but running over 1m6f80y on good going with no prior runs on that surface. Meriden sat fourth at 6.51/10 — remarkable form string, but no market move whatsoever and a cold trainer. Private Peaceful posted 6.40/10 but carries a trainer on 0-from-10 in the last 30 days and no form string on file; the edge figure of +587.1% is eye-catching for the wrong reason — it usually signals a data gap, not a genuine misprice. Highland Harvey matched Private Peaceful on 6.40/10 but is 0-from-4 on going. The separation at the top was driven by market_move combined with a speed advantage and a traininer running hot — all pointing the same direction for Darkened Edge.
The read I keep coming back to is that the market and the model are saying the same thing. Darkened Edge is a confirmed steamer — a strict majority of bookies shortened, none drifted — and that market_move score of +2.00 against a feature weight of 1.6 is the single largest contributor to the composite. When the money moves that clearly on a 10/3 shot in a Class 6 five-furlong handicap at Bath, it deserves a hard look rather than a dismissal. The speed figure reinforces it: an SR of 63 against a field mean of 55.2 gives a z-score of +0.91 and a weighted contribution of +0.33 — comfortably the best raw speed in the race. The form string '58-341' decays to a weighted average of +0.78, good for a form score of +1.17 — that latest win is recent enough to carry real weight in the decay model. Trainer Christopher Mason has sent out 4 winners from 16 runners in the last 30 days, a 25% strike rate the engine calls hot, contributing +1.08 to the trainer score. And the going reads correctly — 2-from-6 on firm or adjacent, with five placed, meaning this horse does not disappear on fast ground. The engine edges the implied market probability (20.6%) by +51.1%, landing an estimated true win chance of 34.9% at odds of 10/3. That is not a massive overlay, but it is a clean, honest one — and at this price it is worth backing.
The composite here is 6.85/10 — solid, not strong. There is no dominant feature pushing this into the 8-plus bracket; the case is built on the accumulation of several solid signals rather than one overwhelming one. The draw data layer is unavailable (v2 feature), and in a five-furlong Bath sprint that is a meaningful gap — stall position can matter materially at this track and distance. The class is Class 6 with no class-change uplift (same grade as last time), so there is no bounce from a drop in the weights. The nearest challenger, War Gaming, has a higher composite of 6.67 and an SR lead in its own field, though the negative edge there makes it a pass for me today. If the going eases from firm by post-time the going score of +1.00 weakens — worth checking the morning going report before placing.
Darkened Edge · 10/3 · 1 unit · MEDIUM
The Fox’s Wire · Bet of the Day
The algorithm pulled five contenders from today’s card. One earned the ticket — the rest are what it beat. When each race settles, the finishes fill in so you can judge the shortlist, not just the selection.
Position only — payout-neutral. A process check, not a P/L. Click any horse to read its profile.
