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 bookies have been hammering this horse in all morning, and that kind of sustained shortening does not happen by accident. Today I ran the full engine across Saturday's card — odds band set between decimal 2.0 and 50.0 (no odds-on shots, nothing drifting past 49/1), field size capped at 6 to 14 runners to cut out the lottery fields, and five market lay candidates removed after a Lay of the Day collision check. Feature weights lean hardest on market movement at 1.6, then speed at 1.5, form at 1.2, with class change and going both at 0.8 and trainer strike rate at 0.8. Draw bias remains unbuilt and sits out again.
The day opened with 31 races and 265 runners declared. The odds filter trimmed that to 239 — removing the odds-on favourites and the exposed longshots. Field size then cut to 208, keeping only races between 6 and 14 runners. The Lay of the Day collision check dropped a further 5 runners, leaving 203 qualified runners to score. Of those, six made it onto the shortlist with composites of 7.0/10 or above.
The shortlist came out in a tight cluster. Early Dawning topped it at 7.51/10, with its speed z-score and a clean steamer read doing the heavy lifting. Brushstroke came second at 7.42/10 — a huge implied-edge number (+313.8%) but no form string on file and no going data at all, which makes that edge hard to trust at face value. Luker's Tipple was third at 7.38/10 with the best form score on the board, but Gavin Cromwell's trainer strike rate coming in at just 5% over 30 days (-0.56 on a weighted feature) was the drag that cost it the top spot. Mr Badger and Kasper Hauser both posted solid composites (7.08 and 7.04 respectively) but neither separated themselves from the pack on any individual feature. Green Sky at 7.01/10 showed the strongest form string of the entire shortlist — form '/45111' is the kind of sequence you notice — but rising in class (delta -1) capped her ceiling here.
The read I keep coming back to with Early Dawning is the market move. A +2.00 score on the market_move feature — the highest possible signal on that metric, weighted at 1.6 — means a strict majority of bookies shortened this horse and not one of them drifted it. That is not noise. At 15/8 (decimal 2.875) the engine estimates a true win probability of 38.0% against a market-implied 31.1%, giving an edge of +9.4%. That is a meaningful gap at a relatively short price, and it is the edge figure I put most weight on when the market feature is singing this clearly.
The speed score backs it up. An SR of 153 against a field mean of 118.4 produces a z-score of +1.74, which at a weight of 1.5 contributes +0.74 to the composite — a genuinely strong reading. The recent form string of '3/4-1' is compact but the decay-weighted latest win carries a +1.09 average, enough for a +0.55 contribution. Willie Mullins' trainer strike rate of 23% over 30 days from 22 winners from 96 runners is comfortably in hot territory (+0.91 on that feature). And the one career run on good or adjacent going ended in a win — thin sample, but pointing the right direction.
The composite of 7.51/10 is the top score of the day, and it is not marginal — it sits 0.09 clear of second-placed Brushstroke, with Brushstroke carrying significant data gaps that make its number less reliable. Class change is unavailable for today's race (unrecognised class format), and draw bias remains a v2 feature, so two inputs are dark. But the three features that are lit — speed, form, and market — are all pointing in the same direction at meaningful scores.
Class change is unavailable, which is a genuine blind spot — I cannot tell whether this is a step up or step down in class, and that matters in a novice hurdle where the field quality can vary sharply. The going score is built on a single run (1-from-1) on good ground, which is directionally fine but statistically thin; a slight softening of the surface could move into untested territory. The price is 15/8, which means the market is already doing a lot of the work — there is limited cushion if the race unfolds awkwardly. And Luker's Tipple — running on the same card at Tramore in an earlier race — holds a form score of +2.10, the highest on the shortlist; if Cromwell's string is about to warm up, that 5% trainer rate becomes a lagging indicator rather than a current truth.
Early Dawning · 15/8 · 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.
