Browse › Reject graveyard: signals killed by the leak-free gate
Analytics module · Date not published.
Reject graveyard: signals killed by the leak-free gate
A REJECT or DEFER is honest market-efficiency evidence.
confirmednull (a finding)not testabledescriptivepending
Reject Graveyard
How to read this figure
The raw counts double-count retested signals; the distinct-signal view (latest_per_signal_graveyard_count) shows 68 unique (sport, signal) pairs currently sitting on a reject verdict.
What it means
The ledger has logged 804 total verdict rows across all reruns; 628 of them are rejects and just 95 signals ever earned a SHIP. The single largest reason for rejection (510 rows) was the as-of reclaim sweep, signals that looked promising until leakage was removed and then died.
Caveats and confounds
The 804 total rows include the same signal retested many times plus rows whose latest verdict later flipped to SHIP, so raw counts overstate the number of distinct dead signals (68 distinct remain).