Stochastic oscillator 80/20 (14, 3, 3) on GBPUSD, 15-minute bars (spec stoch_spec.md). Run 2026-11-08T10:17:44.819Z; 500 control runs per variant
data: 8297247 minutes 2004-01-01 00:05 .. 2026-09-28 15:46 New York; days and bars: {"forex days":5907,"days with a bar in the window":5903,"bars":563363}; cost 1.5 pips a round trip

=== STO-ZONE ===
  DEV      n= 5051 net -5004.30 pips  avg -0.991  PF 0.93 t -1.98
  VAL      n= 4216 net -6726.80 pips  avg -1.596  PF 0.86 t -4.13
  HOLDOUT  n= 4314 net -6220.20 pips  avg -1.442  PF 0.86 t -4.13
  ALL      n=13581; net -17951.30 pips, avg -1.322 (gross 0.178, cost 1.5); PF 0.89 t -5.34; ended by {"signal":5833,"stop":7227,"close":521}
  in R: median R 20.20 pips; gross 0.009 R, net -0.077 R a trade; cost 0.086 R a trade
  cost scenarios (not gates): 1 pips: avg -0.822, net -11160.80 | 2.5 pips: avg -2.322, net -31532.30
  win rate 5822/13581 = 42.9% (95% Wilson 42.0-43.7%); the 10 best trades 1957.10; drawdown on closed trades -18451.30
  longs n=6939 net -8379.30 gross 2029.20 | shorts n=6642 net -9572.00 gross 391.00
  by year (n, net pips): 2004 537 -920 | 2005 516 -959 | 2006 526 -1,146 | 2007 523 -836 | 2008 548 590 | 2009 617 -856 | 2010 603 -192 | 2011 592 260 | 2012 589 -945 | 2013 600 -507 | 2014 594 -1,191 | 2015 584 -1,199 | 2016 605 248 | 2017 601 -1,274 | 2018 620 -1,349 | 2019 612 -1,455 | 2020 636 -849 | 2021 661 -1,678 | 2022 625 -459 | 2023 655 -1,068 | 2024 637 -302 | 2025 633 -1,311 | 2026 467 -553
  control (the same trades moved to other days at the same London time, 500 runs, 6 s): mean trades 11,666, median avg -1.513, 5-95% -1.894 .. -1.168; real avg -1.322 at percentile 80.00
  [FAIL] 1 PF >= 1.2 in validation: VAL PF 0.86
  [PASS] 2 >= 50 trades in each period: 5051 / 4216 / 4314
  [FAIL] 3 t >= 2.5 on development: DEV t -1.98
  [FAIL] 4 above 95th pct of the control: percentile 80.00
  [FAIL] 5 positive without top-10 trades: -19908.40 pips
  [FAIL] 6 holdout net positive (already used): HOLDOUT net -6220.20 pips PF 0.86
  VERDICT Buried (1/6 gates)
  rule 3: 11 trades open across a suspect gap; without them avg -1.332, gates passed 1/6 -> Buried (verdict unchanged)
  trades per traded day: mean 2.34, max 6; per day with a bar in the window 2.30

=== STO-KD ===
  DEV      n= 5511 net -9788.80 pips  avg -1.776  PF 0.87 t -4.08
  VAL      n= 4380 net -7819.10 pips  avg -1.785  PF 0.83 t -5.10
  HOLDOUT  n= 4564 net -6016.90 pips  avg -1.318  PF 0.86 t -4.11
  ALL      n=14455; net -23624.80 pips, avg -1.634 (gross -0.134, cost 1.5); PF 0.85 t -7.38; ended by {"stop":9663,"signal":4318,"close":474}
  in R: median R 14.10 pips; gross -0.006 R, net -0.140 R a trade; cost 0.133 R a trade
  cost scenarios (not gates): 1 pips: avg -1.134, net -16397.30 | 2.5 pips: avg -2.634, net -38079.80
  win rate 4514/14455 = 31.2% (95% Wilson 30.5-32.0%); the 10 best trades 1938.80; drawdown on closed trades -23707.30
  longs n=7384 net -12348.10 gross -1272.10 | shorts n=7071 net -11276.70 gross -670.20
  by year (n, net pips): 2004 579 -1,812 | 2005 540 -1,651 | 2006 543 -999 | 2007 561 -1,074 | 2008 647 -1,627 | 2009 686 -1,851 | 2010 704 -951 | 2011 651 1,199 | 2012 600 -1,023 | 2013 655 -1,355 | 2014 594 -1,265 | 2015 622 -1,178 | 2016 599 788 | 2017 599 -1,783 | 2018 680 -1,402 | 2019 631 -1,623 | 2020 682 -810 | 2021 676 -1,528 | 2022 672 -1,067 | 2023 715 -627 | 2024 651 -60 | 2025 682 -1,441 | 2026 486 -484
  control (the same trades moved to other days at the same London time, 500 runs, 6 s): mean trades 12,474, median avg -1.504, 5-95% -1.815 .. -1.223; real avg -1.634 at percentile 23.00
  [FAIL] 1 PF >= 1.2 in validation: VAL PF 0.83
  [PASS] 2 >= 50 trades in each period: 5511 / 4380 / 4564
  [FAIL] 3 t >= 2.5 on development: DEV t -4.08
  [FAIL] 4 above 95th pct of the control: percentile 23.00
  [FAIL] 5 positive without top-10 trades: -25563.60 pips
  [FAIL] 6 holdout net positive (already used): HOLDOUT net -6016.90 pips PF 0.86
  VERDICT Buried (1/6 gates)
  rule 3: 9 trades open across a suspect gap; without them avg -1.647, gates passed 1/6 -> Buried (verdict unchanged)
  trades per traded day: mean 2.55, max 7; per day with a bar in the window 2.45

=== STO-2R ===
  DEV      n= 4312 net -3559.90 pips  avg -0.826  PF 0.95 t -1.33
  VAL      n= 3590 net -4781.60 pips  avg -1.332  PF 0.89 t -2.84
  HOLDOUT  n= 3564 net -7141.80 pips  avg -2.004  PF 0.84 t -4.46
  ALL      n=11466; net -15483.30 pips, avg -1.350 (gross 0.150, cost 1.5); PF 0.90 t -4.37; ended by {"target":2907,"stop":6533,"close":2026}
  in R: median R 20.70 pips; gross 0.002 R, net -0.084 R a trade; cost 0.085 R a trade
  cost scenarios (not gates): 1 pips: avg -0.850, net -9750.30 | 2.5 pips: avg -2.350, net -26949.30
  win rate 4236/11466 = 36.9% (95% Wilson 36.1-37.8%); the 10 best trades 2127.00; drawdown on closed trades -15919.30
  longs n=5921 net -7952.80 gross 928.70 | shorts n=5545 net -7530.50 gross 787.00
  by year (n, net pips): 2004 480 -1,275 | 2005 467 -832 | 2006 469 -849 | 2007 445 -185 | 2008 460 1,108 | 2009 532 -1,423 | 2010 484 -48 | 2011 482 1,074 | 2012 493 -1,129 | 2013 514 -879 | 2014 500 -610 | 2015 512 -1,556 | 2016 505 100 | 2017 520 -454 | 2018 515 18 | 2019 524 -1,401 | 2020 525 -855 | 2021 543 -2,037 | 2022 510 -253 | 2023 548 -868 | 2024 524 -575 | 2025 528 -1,351 | 2026 386 -1,204
  control (the same trades moved to other days at the same London time, 500 runs, 7 s): mean trades 9,598, median avg -1.472, 5-95% -1.935 .. -1.036; real avg -1.350 at percentile 67.00
  [FAIL] 1 PF >= 1.2 in validation: VAL PF 0.89
  [PASS] 2 >= 50 trades in each period: 4312 / 3590 / 3564
  [FAIL] 3 t >= 2.5 on development: DEV t -1.33
  [FAIL] 4 above 95th pct of the control: percentile 67.00
  [FAIL] 5 positive without top-10 trades: -17610.30 pips
  [FAIL] 6 holdout net positive (already used): HOLDOUT net -7141.80 pips PF 0.84
  VERDICT Buried (1/6 gates)
  rule 3: 11 trades open across a suspect gap; without them avg -1.353, gates passed 1/6 -> Buried (verdict unchanged)
  trades per traded day: mean 1.98, max 6; per day with a bar in the window 1.94