Data check of XAU-GH-L-N (1-hour golden cross, gold, long only, no stop). Run 2026-10-08T17:16:43.328Z; 2099 suspect or unclear gap windows (data_holes_xau.json)
published run: 387 trades, development t 2.65, gates PPPPPP -> Survivor
winners are held 484 hours on average, losers 142 hours: long trades are more likely to cross a gap
(1) strict, as the rule is written: 159 trades open across a gap removed -> development t -3.57, period trades 31/86/111, gates FFFPFP -> Buried
(2) targeted: 26 trades whose signal bar, entry or exit touches a suspect gap removed -> development t 2.29, gates PPFPPP -> Buried
(3) for scale: removing 26 trades at random (2,000 draws): mean development t 2.56, 36% of draws below 2.5; the targeted result 2.29 is at the 9th percentile of the draws
-> rule 3: the verdict changes without those trades, so XAU-GH-L-N becomes Inconclusive (data quality). Owner decision 2026-10-08: apply the rule (option A).