Signal research
What the current score gets right, gets wrong and has not yet proved. Betting performance lives only on the Results page.
Public evidence archive
Research evidence — not a betting performance record
Every row is one fixed pre-match probability vector selected under the same rule. It measures what the signal got right and wrong; it is not a published pick and does not support a profit or ROI claim.
Every eligible pre-match decision
The nearest complete probability vector recorded at least one hour before kickoff is preserved. Repeated scans cannot add extra decisions for the same fixture and market.
“Evaluated” means the final result is confirmed and scored. “Provisional final” is only the short state before a final score is independently reconfirmed.
Evaluation rule
One decision per fixture and market, chosen before the result is known. This archive evaluates signal quality only. Verified prices and returns belong exclusively to published Gems.