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 logbookGood ground across two countries today, and the money on one horse moved with a conviction the speed figures actually justify. The engine scanned all 53 races declared on 28 August 2026, covering 538 runners at the start. Three filters were applied in sequence: an odds band of 2.0–50.0 decimal (no odds-on, nothing beyond 49/1), a field-size window of 6–14 runners to strip out the noise of small fields and the chaos of big ones, and a lay-of-day collision check. The market-move feature carries the highest weighting in the model at 1.6, followed by speed at 1.5 and form at 1.2 — so when a horse steams in the morning market and also posts a credible speed figure, the model takes that seriously.
The race-state check left all 538 runners untouched. The odds band cut that to 413, removing odds-on favourites and outsiders beyond the band. Field-size pruning brought it down further to 302 qualifying runners. The lay-of-day collision check dropped zero additional runners — four horse IDs were excluded on the lay side but none of those collided with the surviving 302. That is the pool the composite scores were built from.
Six horses cleared the filters with composites worth discussing. Kasper Hauser led on 7.56/10, driven by a market-move score of +2.00 and a speed z-score of +1.85 — SR 111 against a field mean of 87.1. Martin's Denuo came second on 7.45/10 with the single largest edge in the shortlist (+221.0% over implied odds) but a trainer stone cold at 6% over 30 days and no going data on file. Down After Dark scored 7.20/10 on a strong speed figure and a +174.2% market edge, but form is entirely absent from the record and the trainer is running at 4% — thin foundations under a decent score. Faerie Tales sits in the same race at 7.10/10 with the highest raw SR of the day (150 vs the same field mean of 130.3), but the going record — zero wins from three on good or adjacent — was the tiebreaker that kept it behind its stablemate. Drymee and Thursday Girl both scored in the mid-sixes, both steamers in the same Sandown mile handicap, but their composites (6.97 and 6.91) and near-breakeven market edges set them apart from the top of the list. Kasper Hauser won the score-off because it is the only runner that combines a top-three composite with a meaningful speed advantage, a live form string, and a confirmed going profile — three pillars rather than two.
The read I keep coming back to is the combination of the steam and the speed number. Kasper Hauser is a confirmed steamer — a strict majority of bookies shortened, none drifted — and that market-move contribution of +2.00 at a weight of 1.6 is the highest-weighted driver in the model firing at its ceiling. That alone does not make a bet. What makes it a bet is that the SR of 111 sits +1.85 standard deviations above the field mean of 87.1. The speed feature, weighted at 1.5, is the second-heaviest in the model, and it is doing real work here, not just covering for a weak price. The form string '2-5311' adds further weight: the decay-weighted average of +1.39 on the latest win reflects recent, genuine winning form rather than a distant highlight. The going data gives me confidence on the conditions — 2-from-5 on good or adjacent going with three placed, so this horse has been tested on the ground and handled it. The model's estimated win probability is 38.3% against a market-implied 19.8%, an edge of +72.3%. I do not expect the edge to be that clean in reality, but a number that size is pointing in a clear direction. At 7/2 decimal 4.5, the price has shortened but still represents value against a model that puts him closer to 5/2 on the numbers.
The honest risks start with the trainer. Kevin Thomas Coleman is running at 10% over the last 30 days — on par, not hot — and the trainer-strike feature came in slightly negative at -0.17. That is not a red flag but it is not a tailwind either. The class-change feature is unavailable because today's race class is unrecognised by the engine, so I have no model read on whether this is a step up or a drop in grade — that is a genuine gap. Draw data is reserved for version 2 of the model and is simply not available for any runner today. The race is a steeplechase over 2m3f100y on good ground at Wexford in the evening, and late going changes on an Irish summer evening are always a possibility — if it softens meaningfully before the off, the going profile (2-from-5 on good or adjacent) becomes less informative. Finally, Martin's Denuo has a composite only 0.11 behind at 7.45/10, and its raw market edge of +221.0% is extraordinary — if you think the Fontwell market is sharper than the Wexford market tonight, that is a credible counter-argument.
Kasper Hauser · 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.
