Classic pivot points on GBPUSD, 15-minute bars (spec piv_spec.md). Run 2026-11-07T17:18:05.852Z; 500 control runs per variant data: 8297247 minutes 2004-01-01 00:05 .. 2026-09-28 15:46 New York; forex days 5907: {"no previous day":1,"short previous day":6,"traded day":5862,"no 08:15 London minute":38}; cost 1.5 pips a round trip previous-day range (H - L), median pips by year: 2004 168 | 2005 133 | 2006 131 | 2007 120 | 2008 185 | 2009 201 | 2010 154 | 2011 131 | 2012 90 | 2013 101 | 2014 79 | 2015 113 | 2016 128 | 2017 92 | 2018 101 | 2019 85 | 2020 116 | 2021 86 | 2022 114 | 2023 93 | 2024 73 | 2025 84 | 2026 75 === PIV-REV === DEV n= 1294 net 218.13 pips avg 0.169 PF 1.01 t 0.11 VAL n= 974 net -1625.03 pips avg -1.668 PF 0.90 t -1.30 HOLDOUT n= 973 net -2447.40 pips avg -2.515 PF 0.85 t -2.13 ALL n=3241; net -3854.30 pips, avg -1.189 (gross 0.311, cost 1.5); PF 0.94 t -1.44; ended by {"stop":842,"target":1160,"close":1239} in R: median R 54.97 pips; gross 0.004 R, net -0.026 R a trade; cost 0.030 R a trade cost scenarios (not gates): 1 pips: avg -0.689, net -2233.80 | 2.5 pips: avg -2.189, net -7095.30 win rate 1725/3241 = 53.2% (95% Wilson 51.5-54.9%); the 10 best trades 1918.77; drawdown on closed trades -5380.00 longs n=1599 net -2219.73 gross 178.77 | shorts n=1642 net -1634.57 gross 828.43 by year (n, net pips): 2004 128 -95 | 2005 159 -1,048 | 2006 127 590 | 2007 134 -21 | 2008 143 192 | 2009 154 1,397 | 2010 151 -49 | 2011 143 -144 | 2012 155 -604 | 2013 135 -604 | 2014 129 -394 | 2015 146 -490 | 2016 132 95 | 2017 146 241 | 2018 146 -246 | 2019 140 -227 | 2020 156 -221 | 2021 149 -199 | 2022 143 -428 | 2023 153 -767 | 2024 139 -273 | 2025 132 -289 | 2026 101 -271 control (the same trades moved to other days at the same London time, 500 runs, 76 s): mean trades 3,241, median avg -1.539, 5-95% -2.546 .. -0.300; real avg -1.189 at percentile 66.40 [FAIL] 1 PF >= 1.2 in validation: VAL PF 0.90 [PASS] 2 >= 50 trades in each period: 1294 / 974 / 973 [FAIL] 3 t >= 2.5 on development: DEV t 0.11 [FAIL] 4 above 95th pct of the control: percentile 66.40 [FAIL] 5 positive without top-10 trades: -5773.07 pips [FAIL] 6 holdout net positive (already used): HOLDOUT net -2447.40 pips PF 0.85 VERDICT Buried (1/6 gates) rule 3: 10 trades open across a suspect gap; without them avg -1.100, gates passed 1/6 -> Buried (verdict unchanged) === PIV-BRK === DEV n= 1380 net -3616.20 pips avg -2.620 PF 0.89 t -1.73 VAL n= 1031 net 236.13 pips avg 0.229 PF 1.02 t 0.20 HOLDOUT n= 1031 net -1391.70 pips avg -1.350 PF 0.92 t -1.17 ALL n=3442; net -4771.77 pips, avg -1.386 (gross 0.114, cost 1.5); PF 0.93 t -1.78; ended by {"target":1525,"close":1181,"stop":736} in R: median R 61.00 pips; gross 0.009 R, net -0.018 R a trade; cost 0.027 R a trade cost scenarios (not gates): 1 pips: avg -0.886, net -3050.77 | 2.5 pips: avg -2.386, net -8213.77 win rate 1912/3442 = 55.5% (95% Wilson 53.9-57.2%); the 10 best trades 1549.17; drawdown on closed trades -6417.97 longs n=1726 net -1327.97 gross 1261.03 | shorts n=1716 net -3443.80 gross -869.80 by year (n, net pips): 2004 151 315 | 2005 152 378 | 2006 149 255 | 2007 139 -184 | 2008 145 -886 | 2009 154 -1,876 | 2010 158 -925 | 2011 168 -378 | 2012 164 -314 | 2013 154 843 | 2014 132 167 | 2015 157 333 | 2016 133 147 | 2017 142 -710 | 2018 160 -225 | 2019 153 -319 | 2020 173 -620 | 2021 161 -223 | 2022 145 -198 | 2023 166 7 | 2024 139 -378 | 2025 140 317 | 2026 107 -298 control (the same trades moved to other days at the same London time, 500 runs, 102 s): mean trades 3,442, median avg -1.685, 5-95% -2.784 .. -0.648; real avg -1.386 at percentile 65.00 [FAIL] 1 PF >= 1.2 in validation: VAL PF 1.02 [PASS] 2 >= 50 trades in each period: 1380 / 1031 / 1031 [FAIL] 3 t >= 2.5 on development: DEV t -1.73 [FAIL] 4 above 95th pct of the control: percentile 65.00 [FAIL] 5 positive without top-10 trades: -6320.93 pips [FAIL] 6 holdout net positive (already used): HOLDOUT net -1391.70 pips PF 0.92 VERDICT Buried (1/6 gates) rule 3: 7 trades open across a suspect gap; without them avg -1.412, gates passed 1/6 -> Buried (verdict unchanged) === PIV-LDN === DEV n= 1846 net -4657.87 pips avg -2.523 PF 0.87 t -2.23 VAL n= 1592 net -3136.92 pips avg -1.970 PF 0.87 t -2.27 HOLDOUT n= 1499 net -1920.18 pips avg -1.281 PF 0.91 t -1.47 ALL n=4937; net -9714.97 pips, avg -1.968 (gross -0.468, cost 1.5); PF 0.88 t -3.44; ended by {"target+break-even":1197,"stop+stop":2256,"target+close":431,"target+target":552,"close+close":501} in R: median R 33.30 pips; gross 0.009 R, net -0.046 R a trade; cost 0.055 R a trade cost scenarios (not gates): 1 pips: avg -1.468, net -7246.47 | 2.5 pips: avg -2.968, net -14651.97 win rate 2322/4937 = 47.0% (95% Wilson 45.6-48.4%); the 10 best trades 2355.38; drawdown on closed trades -10671.48 longs n=2542 net -5258.97 gross -1445.97 | shorts n=2395 net -4456.00 gross -863.50 by year (n, net pips): 2004 215 -321 | 2005 209 199 | 2006 211 -334 | 2007 192 219 | 2008 202 -807 | 2009 183 -928 | 2010 203 -1,955 | 2011 211 -339 | 2012 220 -392 | 2013 221 -101 | 2014 220 -1,500 | 2015 240 87 | 2016 227 12 | 2017 226 -551 | 2018 231 -522 | 2019 227 -560 | 2020 221 470 | 2021 218 -726 | 2022 221 -512 | 2023 232 -106 | 2024 229 -888 | 2025 209 217 | 2026 169 -376 control (the same trades moved to other days at the same London time, 500 runs, 97 s): mean trades 4,937, median avg -1.264, 5-95% -2.101 .. -0.410; real avg -1.968 at percentile 9.80 [FAIL] 1 PF >= 1.2 in validation: VAL PF 0.87 [PASS] 2 >= 50 trades in each period: 1846 / 1592 / 1499 [FAIL] 3 t >= 2.5 on development: DEV t -2.23 [FAIL] 4 above 95th pct of the control: percentile 9.80 [FAIL] 5 positive without top-10 trades: -12070.35 pips [FAIL] 6 holdout net positive (already used): HOLDOUT net -1920.18 pips PF 0.91 VERDICT Buried (1/6 gates) rule 3: 9 trades open across a suspect gap; without them avg -2.002, gates passed 1/6 -> Buried (verdict unchanged)