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 logbookEvery runner today was checked against the market before the form book — and the money tells a clear story that points straight to one horse. The engine scanned all 20 races and 226 runners declared on 23 August 2026, running each through a four-stage filter pipeline. Stage one cleared abandoned or already-resulted races — no runners dropped. Stage two applied the odds band: decimal 2.0–50.0, cutting out odds-on runners and anything north of 49/1, trimming the field to 194. Stage three enforced a field-size window of 6–14 declared runners — small fields are noisy, big fields are chaos — leaving 110. Stage four removed one runner that collides with the Lay of the Day selection, leaving 109 qualified runners. Feature weights applied were: market_move 1.6 (highest), speed 1.5, form 1.2, class_change 1.0, trainer_strike 0.8, going 0.8, draw 0.5 (reserved for v2 — unavailable throughout today).
The day opened with 226 runners across 20 races. The odds filter cut that to 194. Field-size enforcement took it further to 110. One Lay of the Day collision dropped a further runner. 109 runners survived into scoring. The shortlist you see here represents the top six composites from that pool — all sitting on good ground today, spread across Worcester and Naas.
The top six composites are Spanish Darling at 7.88/10, At The Oche and Kernie d'Airy joint second at 7.15/10, Soul Love at 7.00/10, Axel Bleue at 6.96/10, and Dancing Saxon at 6.88/10. The gap between first and second is not enormous — 0.73 points — so I looked hard at what separated them. The answer is market_move and form combining at a level none of the chasing pack can match: Spanish Darling posts market_move +1.24 as its weighted contribution and form +1.03, both of which are the largest weighted outputs in the shortlist at those features. At The Oche has the superior SR and a hot trainer in Dan Skelton, but at 2.25 decimal the model's implied edge is -18.3% — the market has already priced it tighter than the model rates it. That alone pushes it down the order for a value-focused read. Kernie d'Airy is interesting — Jamie Snowden's 30-day strike rate of 40% is the standout trainer number of the card — but the form score is softer and the composite ties with At The Oche rather than leading. The top score belongs to Spanish Darling, and it earned it.
The read I keep coming back to is the market edge. The model estimates Spanish Darling's win probability at 39.8%; the market is implying 9.4% at 17/2. That is an edge of +278.3% — the largest positive edge in today's shortlist by a distance and the primary reason the composite sits 7.88/10, which I class as a strong score. The market_move feature contributes a weighted +1.24 to that composite: the engine flags a steamer — a strict majority of bookmakers shortened, none drifted. In a National Hunt Flat at Worcester that is not nothing. Money moving into a field that is by definition less exposed is a meaningful signal, because there are fewer excuses for confident bookmaker positioning. When the money comes in anyway, I pay attention. The speed figure supports it: SR 102 against a field mean of 85.3, a z-score of +1.06, weighted contribution +0.61 — solid, not spectacular, but comfortably ahead of the pack in this Class 5 NHF. The form string '251-1' with a decay-weighted average of +1.47 on the latest win is what I call a returning-winner pattern — the horse has been placed, gone away, come back, and won. The weighted form contribution comes in at +1.03. Three of the five scored features are pointing the same direction, and the one that matters most by weight — market_move — is screaming.
I want to be straight about the gaps in this profile. Three features are UNAVAILABLE: class_change (no prior recognised-class races to compare), trainer_strike (only 4 runners in 30 days, below the noise threshold of 5), and going (no career runs on good or adjacent going recorded). That last one is the risk I sit with most. Good ground today at Worcester, and I have no data on how Spanish Darling handles it — the going score contributes nothing to the composite either way. The composite of 7.88/10 is built on three features rather than six. A horse with a fuller evidence base could be scored more reliably; here the gap between the model's implied probability (39.8%) and the market (9.4%) is doing a lot of the heavy lifting. If the going plays a role the horse does not suit — and we simply do not know — that edge narrows fast. The runner-up pair at 7.15/10 are not far behind, and Kernie d'Airy in particular has a trainer number that catches the eye. Eyes open.
Spanish Darling · 17/2 · 1 unit · HIGH
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.
