Methodology

How forecasts are graded

The public record is a match by match archive, not an aggregate performance claim. This page explains the rules used to turn a published selection into won, lost or void.

View graded archive

Grading rules

  1. 01

    Published before kickoff

    The selection and its model snapshot are stored before the match begins.

  2. 02

    Settled from the final result

    A grade is written only after a final score and an accepted full time status are persisted.

  3. 03

    Won, lost or void

    Every archived shortlist selection keeps one explicit grade. Void is not counted as a win or loss.

  4. 04

    Snapshot stays frozen

    The archived matrix is the pre match model output. The final result is never used to recalculate it.

Settled record by market

Counts, not a claim. These are the selections we published and later graded, not every forecast the model made. A void returns the stake and is left out of the settled total.

MarketWonLostSettledNo verdict
Over/Under Goals28,0429,60837,650
Double Chance5,0282,7737,801
Asian Handicap4,4772,2406,7174
Team Total Goals3,6801,6935,373
1X2 / Match Winner2,9871,4734,460
BTTS6883761,064
Result + Total Goals19974273
Total settled45,10118,23763,3384

Download the settled record (CSV)Every settled selection of the last 365 days, one row each, with the price and the model probability we published.

What is included

Only shortlisted selections with a complete frozen snapshot, direct answer, signals and final grade can enter the public archive.

What is not claimed

The archive does not guarantee future results. It avoids headline win-rate and ROI claims that can hide sample and selection effects.