The July Festival’s big handicaps, through the lens of Smartform runner insights

First, what this is. These aren’t tips. At Royal Ascot we read the meeting through the breeding tables. Across the July Festival we are doing something different: taking one of the meeting’s top handicaps each day and asking what the Smartform runner insights tables make of it — which of these races can actually be read, and which cannot.

The tables doing the work are daily_runners_insights and its historical twin historic_runners_insights. Between them they carry a few hundred leak-free columns describing every runner exactly as it stood on the morning of its race: how its official mark has moved across its last three starts, what proportion of the field it beat each time, whether it stepped up in class, how long since it ran. Alongside those we use trainer_bin_stats_daily for a yard’s current form.

The method is deliberately slow. For each of the festival’s big handicaps we built a profile of the typical winner using the runnings up to 2023, wrote that profile down, and only then tested it against 2024 and 2025 — two years we had held back and not looked at. Most ideas died on that test. Speed figures died. Clever theories about horses well handicapped against their old peak died. Two of the festival’s handicaps came through it with something worth having.

One of them is on today’s card.

3:00 — Betway Handicap (Heritage Handicap)

Six furlongs, Class 2, three-year-olds, £100,000.

The trend you will see quoted this morning concerns the draw: nine of the last eleven winners came from stall eight or higher. We have looked at the draw over Newmarket’s summer sprint trips and have not found it to be significant. Eleven runnings is a thin sample, and when we tested a draw bias over six furlongs here against renewals we had held back, it did not survive. That is not to say there is nothing in it — only that we cannot find it, and we would not move a horse up or down on it.

What does matter is the direction of the handicapper’s pen. The insights tables carry each horse’s official rating not just for today but for each of its last three starts, so you can watch the assessor forming his opinion in public.

Look at the recent winners. Dancing Star came here in 2016 rated 75, then 80, then 89 — and won off 96. Quinault won in 2023 having gone 74, 74, 85, 90. Sergeant Wilko in 2024: 78, 85, 84, 87. Prince Of India last year: 77, 80, 84, 87. Every one was being put up, run after run. Every one won anyway.

That is the whole thing. The rising mark is not a warning — it is the evidence. This race is won by the three-year-old improving faster than the handicapper can raise him, and it works in July for a reason: young horses improve between runs quicker than any assessor can react, and by midsummer there is enough form on the page for the gap to become visible. In an all-aged handicap full of exposed horses, a mark going up means the handicapper has caught up. In this one, it usually means he has not.

Advertised is a non-runner. Marks read antepenultimate → penultimate → last time out → today; a dash means the horse had not yet been rated. Forecast prices in decimal.

Reading the columns. Mark is the official rating the horse carried at each of its last three starts, and the one it carries today — the handicapper’s opinion, revised in public. Rivals beaten is the percentage of the field the horse finished in front of last time and the time before; a fairer measure of a run than the bare finishing position, since eighth of twenty-eight is a very different afternoon from eighth of nine. Yard PRB is how the stable’s runners have been performing lately on that same rivals-beaten scale, with the number of runs behind the figure in brackets — 50 is par, and a small sample tells you little.

Sea Cookie is the closest thing to the archetype. Four starts in his life. Won last time, stepping up two classes out of a Class 4, and put up five pounds for it. He has beaten every rival in each of his last two runs, has two wins in the last ninety days, and already owns a course-and-distance win. Lay him alongside Sergeant Wilko and Prince Of India and it is very nearly the same horse. The reservation is his yard’s figure, which reads well but on only seventy-odd runs — too few to lean on.

Thunder Call is the same shape and makes no secret of it. Won last time, raised six pounds, every rival beaten in each of his last two, four starts, and the strongest yard in the race behind him on a properly powered sample. The profile likes him and so does the market. Worth saying plainly rather than pretending otherwise: when the data and the betting agree, you are not being paid for the insight.

Red Spells Danger has the steepest curve on the card — 79, 79, 89, 93, fourteen pounds across three runs, with 95% and then 100% of his rivals beaten. On the form line alone he is arguably the most improving horse here. The pause is his yard, whose current figure is the coldest of the fancied runners and on a big enough sample to mean something. He is the one the form book loves and the numbers behind him do not.

Then the other side of the ledger. Ten Carat Harry is top-rated on 101, a course-and-distance winner, seven wins at the trip from fourteen starts. He may well run a fine race. But he is precisely the horse the handicapper has already read — the mark has not moved in three runs because there is nothing left to learn. That is the type this race has consistently overrated. Ghost Mode and Mo Of Cairo are travelling the other way entirely, drifting down the weights, which is what happens to horses that are not improving.

The profile points to Sea Cookie at a price and Thunder Call at a short one; Red Spells Danger is the improver with a question behind him. On the draw, we have nothing to add.

That is today’s “trends” pick. The mark paths, the rivals-beaten figures and the yard form all come straight from the Smartform data — run the same columns yourself in daily_runners_insights and trainer_bin_stats_daily and you will see exactly the picture we did, including the parts that did not work.