How We Read a Race
Every card follows the same repeatable process — from raw program data to a probability-weighted read on the field.
Data Ingestion
Full past-performance programs are parsed race by race — every runner, post position, jockey, trainer, morning line, running line and workout history from the official card.
Pace Projection
Each field is mapped by running style to project early pressure, contested fractions and the shape of the stretch run — the single biggest driver of how a race unfolds.
Merit Scoring
Runners are graded across weighted pillars covering recent form, class, speed figures, surface and distance fit, connections and trip context.
Simulation
The field is run through large-scale Monte Carlo simulation to convert merit and pace into win, place and show probability distributions rather than single-point guesses.
Expert Review
Model output is reviewed against the tape and the card. Scratches, track bias and late information trigger a re-run before anything is published.
Grading
Once races are official, selections are graded against charted results so the track record reflects what actually happened.
What the numbers do and don't mean
Probabilities describe how often an outcome is expected across many runnings of the same race — not a guarantee for any single running. Horse racing carries inherent variance, and our analysis is published for informational purposes only. No wagers are placed or accepted on this site.
See it applied in the Paddock