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 market screamed at me today, and the form backed it up — that is the story of the 19:48 at Kempton. The engine scanned all 33 races and 280 runners declared for 8th July 2026, working through a three-stage filter pipeline: first stripping out any abandoned or already-resulted races (no change, 280 → 280), then cutting to runners priced within the 2.0 – 50.0 decimal odds band — no odds-on, no longshot chaos — which dropped the field to 226, and finally enforcing a 6–14 runner field-size rule to avoid both the lottery of small fields and the noise of big ones, leaving 198 qualified runners to score. Feature weights across the engine ran at speed ×1.5, market move ×1.6 (the highest single weight), form ×1.2, class change ×1.0, trainer strike and going both ×0.8, and draw flagged as unavailable pending v2.
The day opened with 33 races and 280 runners. After the odds-band filter removed odds-on and 50/1-plus runners, that became 226. The field-size filter then trimmed to 198 runners across races with between six and fourteen declared starters — the working universe the composite scores were built from.
Five horses separated themselves from the pack. Fallacious Promise topped the board at 7.10/10, with Speed Nation close behind at 6.98/10 and Warby at 6.87/10. Crayons In The Sky posted 6.81/10 and Mr Tony 6.80/10, with Argentarius just outside at 6.77/10. The gap between first and second — 0.12 points — is real but not wide, so I looked hard at what actually tipped it. For Fallacious Promise, the combination of the highest weighted market-move score, a speed rating well clear of the field mean, and a hot trainer in form was enough to hold off Speed Nation, who has gaps in her profile that the engine cannot fully score.
The read I keep coming back to is the market move. Fallacious Promise is a steamer — a strict majority of bookmakers shortened, none drifted — and that delivered the maximum +2.00 market-move contribution on the highest-weighted feature in the model (×1.6). That alone is significant. Someone, somewhere, decided this race is winnable at odds that were already relatively short, and they pushed the price in regardless. That is the kind of signal I take seriously.
Behind the market move, the speed figure reinforces the case. An SR of 73 against a field mean of 54.1 — a z-score of +1.81 — puts Fallacious Promise clearly ahead of this field on raw ability. The form string '-92231' with a decay-weighted score of +1.07 shows a horse trending in the right direction, the recent win carrying meaningful weight. And trainer John Butler is running at 26% over the last 30 days (11/43), which the engine rates as hot — contributing +1.13 on the trainer-strike feature.
I had a long look at Speed Nation as a challenger. The model likes her edge — the implied win probability of 35.5% against a market implied 27.5% gives a positive edge of +15.4%, which is the best edge number on the shortlist. But there is no form string on file, no going data, and no class-change context. Three unavailable features on a fillies' novice at 9/4 is a meaningful hole in the evidence base. Fallacious Promise has the deeper data trail and the stronger composite.
The going data for Fallacious Promise is thin — 0-from-1 on good to soft or adjacent ground, with zero placed runs in those conditions, scoring -0.10 on a feature weighted at ×0.8. That is not catastrophic, but if the Kempton surface plays softer than expected this evening it is the one feature pointing the wrong way. The market implied probability of 37.6% is actually slightly above the engine's estimated win probability of 36.1%, giving a negative edge of -14.3% — meaning at 11/8 the model does not see a quantifiable overlay. The bet here is being made on the composite strength and the market-move signal rather than a strict value edge, and that distinction is worth owning. The draw data layer is unavailable for v2, so any draw bias at Kempton over one mile is unquantified. Speed Nation at 6.98/10 is breathing closely enough that I would not argue hard if someone preferred the Crisford filly, but the missing data there is the bigger unknown.
Fallacious Promise · 11/8 · 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.
