Opening gap fill self-test, 2026-10-22T12:13:24.081Z

1. Hand-worked days (Monday 8 and Tuesday 9 January 2024)
A. close 4800, open 4810 (+0.21%): short at 4810, target 4800, stop 4820; bar 600 trades down to 4799 -> out at 4800
  [ok] GF-1X: 469.24 dollars
  [ok] reference identical (3 variants)
B. open 4790 (-0.21%): long, target 4800, stop 4780; bar 600 reaches 4779 and 4801 -> the stop first (4780)
  [ok] GF-1X stopped at 4780; GF-NS (no stop) reaches the target in the same minute; reference identical
C. size bounds: a +0.60% gap is taken by GF-1X and GF-NS, not by GF-SMALL; a +0.05% gap by none
  [ok] bounds respected; reference identical
D. no level reached: out at the session close (16:00) at the last close
  [ok] closed at 4815 at 16:00
E. after an early close (Friday 29 Nov 2024, 13:15) the previous close is the 13:15 bar; Monday 2 Dec gaps from it
  [ok] previous close 6000, gap 12
F. a day without its 09:31 bar is not a trading day; G. a previous trading day 8 days earlier (a hole) gives no trade
  [ok] no trade in either; reference identical

2. Real data: engine against the independent reference
  2378027 bars 2020-01-01 18:01 .. 2026-09-25 16:59 New York; 1737 trading days, 1736 with a previous close
  [ok] GF-1X: 1391 trades, reference 1391, identical
  [ok] GF-SMALL: 907 trades, reference 907, identical
  [ok] GF-NS: 1391 trades, reference 1391, identical

3. Control at offset 0
  [ok] GF-1X: 1391 control trades identical to the real ones
  [ok] GF-SMALL: 907 control trades identical to the real ones
  [ok] GF-NS: 1391 control trades identical to the real ones

PASSED