About AI Horse Racing Prediction Lab

Test probability statements against outcomes rather than judging a model by its name.

Our focus

AI horse racing tips use model outputs to inform a racing opinion. AI betting tips can involve ratings, probabilities or ranked selections, so the meaning of the output matters. This publication focuses on probability calibration and forecast evaluation. A prediction of 30% can lose without being disproved and win without being validated. Evaluate many predictions using a declared target and an untouched test period. Calibration asks whether predicted frequencies align with observed frequencies; discrimination asks whether higher scores tend to belong to the eventual outcome. Both can matter, and neither removes uncertainty in one race.

Our publisher

Part of Past The Post Technolgies Pty Ltd

AI Horse Racing Prediction Lab belongs to the collection of racing, AI and technology publications described in our website directory. Each publication focuses on a specific research question, practical method or product requirement.

Racing research

For horse racing ratings, form and related racing resources, visit PTPTips.

Editorial approach

Our guides distinguish recorded facts from interpretation, explain their assumptions and use clearly identified hypothetical scenarios for worked examples. Changing race information and commercial product details should be checked at their source.

For a question or correction, use the contact options on PTPTips and include the publication name and article URL.