Prediction Intelligence
See the model’s projected finishing order and relative confidence.
Racing intelligenceThe intelligence behind every race
Thousands of connected details become one composed forecast—built to deepen understanding without hiding the evidence behind it.
The app experience
Every screen is designed for calm, rapid understanding—from the first race card to the completed-race review.

More than a prediction
Advanced prediction should be explainable. Every FormSignal forecast is supported by visible evidence, performance context, model confidence and an enduring post-race record.
See the model’s projected finishing order and relative confidence.
Understand likely pace, positioning and the evolving shape of the contest.
Discover the conditions that could strengthen or weaken each horse’s performance.
Interactive race visualiser
Horse intelligence profiles
A living performance profile brings together suitability, consistency, trajectory and competitive strength.
“A progressive profile with strong pace compatibility and a clear preference for a sustained final approach.”Model commentary
Built for the serious racing mind
Created for knowledgeable enthusiasts, owners, breeders, syndicates, analysts and followers who appreciate racing at a deeper level.
Exclusive membership
Access is deliberately limited to preserve the quality and exclusivity of the experience.
Private memberIndividual review · Private access · Member-first releases
Methodology & credibility
FormSignal’s research archive contains more than 1.5 million historical runner records across almost 188,000 races. Its model laboratory evaluates XGBoost, LightGBM and CatBoost—three of the world’s most respected machine-learning frameworks—then promotes only an approved model into live production.
More than 1.5 million runner records across almost 188,000 races support large-scale racing research and model development.
XGBoost, LightGBM and CatBoost are trained and compared as controlled candidates; only an approved model is promoted to production.
Past races are used to test model behaviour before new methods are trusted in production.
Every forecast communicates the model’s measured certainty instead of presenting all selections as equal.
Eligible forecasts are timestamped and sealed before the scheduled start, then revealed publicly after a 24-hour member-protection period.
Every completed race becomes new evidence—revealing both where the model succeeded and where it must improve.
Forecasts represent modelled probabilities, not guaranteed outcomes.
Public verification ledger
Trust is earned race by race. Forecasts received before the scheduled start are timestamped, fingerprinted and permanently retained. Each sealed entry proves a forecast existed before the start without exposing the member selection. Forecast details and outcomes become public 24 hours later, preserving member value while keeping the record accountable.
Earlier PC history has not been backfilled. Eligible forecasts will appear automatically when received by the cloud before upcoming races.
Forecasts represent modelled probabilities, not guaranteed outcomes. Confirmed non-runners and abandoned races are excluded from completed-race rates.
Questions, answered
FormSignal is a private race-forecasting and thoroughbred performance-intelligence platform. It brings complex racing evidence into one clear, explainable view.
Applications are reviewed individually. Early membership is intentionally measured so the experience, support and intelligence remain exceptional.
Structured racing datasets are evaluated through proprietary predictive models, then presented with evidence, context and confidence.
No. Forecasts represent modelled probabilities, not guaranteed outcomes. Racing remains uncertain by nature, and that uncertainty is reflected openly.
Your private invitation
Join a private circle of racing enthusiasts using a more intelligent way to read the field.