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 moved, and I followed it — that is the read that keeps pulling me back to Leopardstown tonight. The engine scanned 48 races and 404 runners declared for Saturday. From that universe, the odds filter stripped everything priced outside 2.0–50.0 decimal — no odds-on, nothing at 50/1 or beyond — leaving 327 runners. The field-size filter then removed anything running in fewer than six or more than fourteen declared runners, on the grounds that small fields are noisy and big fields are chaos. That brought the working universe down to 234 runners across viable races. Feature weights were applied in full: market_move carried the heaviest load at 1.6, speed at 1.5, form at 1.2, class_change and going both at 0.8 with trainer strike rate also 0.8, and draw at 0.5 — though draw-bias data is not yet built and returned unavailable across all runners today.
The day opened with 48 races and 404 runners. The odds-band filter reduced that to 327 runners, and the field-size cut brought the live universe to 234 runners in qualifying races. Six runners made the composite shortlist, all scoring above 6.6/10. The top of the board was tight — just 0.23 points separated first from sixth — which means today's conviction is honest rather than emphatic.
The top six were How'sthebai (6.90), Geryon (6.83), Trucial Pearl (6.78), Global Skies (6.75), Ballyskeagh (6.69), and Marmeladova (6.67). The feature that separated the top two was tight: How'sthebai and Geryon both posted a market_move score of +2.00 — both steamers — but How'sthebai edged it on the speed z-score (+1.51 vs +1.42) and on form weight (+1.26 vs +0.71). Global Skies is the most intriguing runner in the six on raw edge — a model-implied win probability of 34.4% against a market-implied 23.8% gives it a +29.0% edge, the second largest in the group — but its market_move returned +0.00, no clear movement either way, and that silence is meaningful when the weight on that feature is 1.6. The horses with the strongest composite scores all share one thing: the bookies moved them in.
The read I keep coming back to on How'sthebai is the combination of speed and market movement in the same runner. The SR of 104 against a field mean of 83.3 produces a z-score of +1.51 — that is not a marginal speed edge, that is a runner operating at a measurably different level to most of what it is facing tonight in the Dublin To Bahrain Handicap at Leopardstown 18:35. The speed feature alone, weighted at 1.5, is doing real work in this composite.
The form score of +1.26 is solid. The recent string '61-414' is not a wall of wins, but the decay-weighted average of +0.84 on the latest first position tells me the engine is giving credit for a recency that matters — the last win is not a distant memory. The market_move score is the headline number: +2.00, the maximum the engine assigns, meaning a strict majority of bookies shortened and none drifted. At 13/8 that is not a horse being ignored. The model puts its win probability at 35.1% against a market-implied 34.0% — the edge is marginal at -7.8%, I will not pretend otherwise, but the direction of the market and the speed superiority are doing enough to make this the day's selection.
Trainer G M Lyons returns a 30-day strike rate of 6/56 = 11% — on par, no boost, no drag worth flagging. The going score is a mild negative: -0.20 on a record of 0-from-2 with no placed runs on good or adjacent ground. That is a flag, not a veto, but it is worth naming.
The going record is the clearest risk on the data. How'sthebai is 0-from-2 with zero placings on good or adjacent ground — the -0.20 going score is a genuine red mark, and if the ground is riding on the faster end of Good tonight that concern tightens. Class_change is unavailable — today's race class is unrecognised — so I cannot confirm whether this represents a step up, step down, or lateral move in grade; that is a blind spot I am holding. The model edge of -7.8% means the market is almost exactly in line with the model's own probability — there is no value gap here in the traditional sense, and anyone who needs a mis-priced favourite to justify their bet should look elsewhere. Geryon at 6.83 is only 0.07 points off the top score and also steamered; it is a live danger and I am not dismissing it. The composite conviction is MEDIUM and the engine agrees — this is a narrow call, not a banker.
How'sthebai · 13/8 · 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.
