Reading the Track
When the surface picks sides
Some days the track itself has an opinion. The rail dries out and front-runners on the inside never get caught — a speed-favoring, rail-biased day. Other days the surface is deep up front, the early pace melts, and closers swallow the field. A horse can run its best race and still lose because it was fighting the track. Miss the bias and you'll back beaten favorites all afternoon.
Flip the surface below and watch which running styles get a tailwind and which hit a wall:
Front-runners carry it home; closers run into a wall.
Same five horses, different surface. A closer on a speed-favoring day is fighting the track; a rail speedster on a rail-bias day gets a free upgrade. The model folds this into zBiasFit — a tailwind nudges its number up, a headwind nudges it down.
The two things bias cares about
Running style — is this horse early speed (E), a presser (EP/P), or a closer (S)? — and post position — inside or outside. A speed bias rewards E types; a closer bias rewards S types; a rail bias rewards low posts. Match a horse to the day's bias and it's running downhill; mismatch it and it's running into wind.
How the model reads it
The Brisnet card carries a track-bias block — recent win rates by style and post. We parse it, compare each horse's style and post to the day's tendency, and produce zBiasFit: positive for a tailwind, negative for a headwind. In the v4 feature set it joins speed/form/pace in the signal at a modest weight (~0.18) — present, but never the whole story.
How to use it at the windows
Two moves. Upgrade an overlay whose style fits the bias — the market may not have caught the tailwind yet, which is where the value hides. And fade a tempting favorite that's fighting the track — a lone closer on a speed-favoring rail day is a trap, however good it looks on paper.