Turfs Edge predictions are generated by a data pipeline that analyses large volumes of historical and current racing data. This page explains the approach conceptually.
Recent finishing positions and performance trends.
The class level of the race and how it compares to each runner's recent level.
Course characteristics and distance suitability for each runner.
Ground conditions and how they suit each runner's profile.
Recent trainer and jockey performance and win rates.
Published prices and market movements where available.
Draw position and weight allocation where relevant.
Large-scale historical racing data for calibration.
The 1–5 star confidence rating represents the relative strength of the predictive signal — how strongly the model's inputs agree on a selection. It is not a guarantee of winning. See confidence ratings for details.
Published prices can change after a selection is posted. Non-runners or late withdrawals can occur. TeT preserves original published odds and timestamps immutably, and records results separately.
No prediction system can guarantee winners. Racing involves significant chance and risk. TeT's confidence ratings reflect model signal strength, not certainty.
Every published selection is permanently recorded with its original timestamp, odds and reasoning. Results are appended separately and never silently rewrite historic selections. This dataset is the foundation of TeT's track record.