Our tennis bot's betting record

256 settled bets since 2026-07-08. Win rate 39.1%, ROI -12.8%, beating the fair closing line 44.7% of the time.

By market

By tour and surface

How this record is kept

What is CLV, and why does this page lead with it?

Closing-line value compares the odds a bet was struck at against the same market's closing price. Because the closing line is the market's best estimate, beating it is evidence of edge that shows up in far fewer bets than win/loss does. A small sample of wins proves little; a consistent CLV edge is the leading indicator.

How are bets graded?

From a snapshot frozen at the moment the bet is placed - the pick, market, line, odds and stake. Grading reads that snapshot, never the live model, so a later re-run cannot retroactively improve a result. Moneyline, handicap and totals are graded on their own terms; alternative markets void on a retirement.

Are these results forward-tested or backtested?

Forward only. Every bet on this page was published before the match started. Backtest results are kept in a separate simulated pool and never mixed into these figures.

Does the model beat the market?

On raw accuracy, no - and we published the measurement. Against the devigged closing market our blend scores lower on both accuracy and Brier score. The bot is not trying to out-predict every match; it passes most of them and stakes only where it reads a priced discrepancy.

We also publish the measurement where our model loses to the market: how accurate are tennis predictions? Full method notes in the Lab.