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 logbookThe money told me something today, and it pointed the same direction the model did — which is exactly when I pay attention. The engine scanned all 203 declared runners across 25 races on the card for 11 August 2026, running them through a four-stage filter pipeline. First, the odds band: only runners priced between 2.0 and 50.0 decimal survive — nothing odds-on, nothing at the each-way fringe above 49/1. Second, field size: races with fewer than 6 or more than 14 declared runners were cut — small fields are too volatile, big fields are chaos where the model's edge dissolves. Third, a lay-of-day collision check removed any runner the system is actively opposing today. That pipeline trimmed 203 down to 170 qualifying runners, and the feature-weighted composite — led by market_move at weight 1.6, speed at 1.5, and form at 1.2 — ranked what remained.
The day opened with 25 races and 203 runners. The odds-band filter removed 20 runners priced odds-on or above the 50.0 decimal ceiling, leaving 183. The field-size gate then trimmed a further 13, leaving 170 runners in scope. No runners were dropped in the lay-of-day collision check. Those 170 runners across eligible races fed into the composite scorer.
Five horses broke 6.40/10 on the composite. Wrist Art topped the table at 6.69/10, a single point clear of Asia Force at 6.68/10 and Influencer Man at 6.67/10 — genuinely tight, those three. Hope Rising came in at 6.65/10 and Patrick Street at 6.42/10. What separated Wrist Art at the top was the combination of a class drop and a market move score of +2.00 landing on a horse also posting +0.92 on trainer strike rate — none of the others stacked those three pillars simultaneously. Asia Force is the runner-up I want to flag honestly: it carries a higher SR (86 vs 55) and stronger speed z-score, but it is 6/4 into a market that already has it at 35.7% implied — the model's estimated edge is -14.8%, meaning the market has already done the work. That negative edge was the deciding factor.
The read I keep coming back to with Wrist Art is the market move against the price. At 7/2, this horse is sitting at a 19.8% market-implied probability, yet the model's estimated win probability is 34.1% — an edge of +53.5%. That is not a rounding error; that is a bookmaker community that has not fully priced a steamer. The market_move feature scored +2.00 — the maximum signal the engine issues — meaning a strict majority of bookmakers shortened, none drifted. I do not ignore that. At 7/2 with that divergence, the value case is concrete.
Beyond the market, the class drop confirms the logic. Wrist Art runs today at Class 6, stepping down from a Class 5 last time out — the class_change feature scored +0.90, reflecting a delta of plus one rank. Dropping into a Lingfield AW Classified Stakes at 1m4f is exactly the kind of placement a trainer uses when they want to give a horse a winnable opportunity. And the trainer, James Owen, is running at 23% over the last 30 days — 24 wins from 104 runners — which the model rates as hot, scoring +0.92 on trainer strike rate. Speed is solid too: an SR of 55 against a field mean of 37.0 gives a z-score of +1.04, sitting comfortably above the field.
The composite of 6.69/10 is solid rather than strong — I will not dress it up as dominant. But when speed, class change, trainer form, and the highest-weighted feature all point the same way on a horse still available at 7/2, I am not looking for an excuse to walk past it.
The going score is the honest caveat: Wrist Art is 0-from-6 on good or adjacent going, with only three placings to show from those six attempts — the feature scored +0.00, a flat zero. All of this horse's market confidence may be resting on its all-weather form at Lingfield rather than turf-good form, and the model cannot fully disentangle that from the going feature as currently built. Form reads '5U3452' — there is an unseated rider in there, and the decay-weighted average of +0.32 is modest. Draw data is also unavailable — the draw-bias layer is reserved for v2, so on a track where draw can matter at 1m4f on the AW, that is a gap in the picture. The composite lead over Asia Force and Influencer Man is just 0.01 and 0.02 points respectively — this is a close score-off, and on another day the rankings could swap with minimal input change. I am backing the edge and the move; I am not pretending this is a certainty.
Wrist Art · 7/2 · 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.
