Articles

A price that shifts between the first observation and the latest one shows where money and opinion have gone. It is evidence of a change, not proof of a winner. BETSPY records the first price it saw, every price since, and never invents a value in between.

One book shortening a price can be a local correction. Several books moving the same way in the same window is a stronger signal, because it is much harder to explain away as one trader's adjustment.

Every posted price includes the bookmaker's built-in margin. Stripping that margin out of both sides of a market gives a fairer estimate of the true chance, which is what any comparison of value has to start from.

Fixed stake keeps every bet identical. Percentage of bankroll scales with what you hold. Capped recovery raises after a loss but is hard-limited. Edge-based staking sizes in fixed bands by the validated edge and refuses the bet below a 3% edge. Same picks, four very different curves.

Recovery staking is popular and is included so it can be measured honestly, but with a firm ceiling and a reset after any win. Without a cap, a short losing run ends the bankroll rather than testing the method.

Thirty settled bets is the minimum before a win rate or return means much. Below that the figures are shown, but flagged as an insufficient sample — an early hot streak is noise, not proof.

Every observation, pick, stake and settlement is written once and locked. Losing bets cannot be quietly removed and past results cannot be re-run under a plan chosen later. A record that can be edited is not a record.