RSI divergence control check (not part of the frozen protocol). Run 2026-10-07T18:23:28.103Z

=== DV-LS-S ===  real: n=622 net 32,255 avg 51.9 $/trade
  real LONG entries: 276; RSI < 50 at entry: 146 (avg 326.5 $/trade, avg bars 9.5); RSI >= 50 at entry: 130 (avg -1.6 $/trade, avg bars 1.8)
  real SHORT entries: 346; RSI > 50 at entry: 222 (avg -95.1); RSI <= 50: 124 (avg 47.6)
  C0 published (long only below RSI 50; short above 50): trades mean 619 (real 622); random median 2.1, 5th -99.8, 95th 99.3 $/trade; real 51.9 -> percentile 78.2
  C1 any RSI (random bar, long; or random side when shorts allowed): trades mean 626 (real 622); random median -14.0, 5th -90.9, 95th 69.0 $/trade; real 51.9 -> percentile 90.6
  C2 RSI-matched (same distribution of RSI at entry as the real signal): trades mean 632 (real 622); random median -16.7, 5th -101.9, 95th 64.3 $/trade; real 51.9 -> percentile 90.6

=== DV-L-S ===  real: n=281 net 48,731 avg 173.4 $/trade
  real LONG entries: 281; RSI < 50 at entry: 147 (avg 328.2 $/trade, avg bars 9.5); RSI >= 50 at entry: 134 (avg 3.7 $/trade, avg bars 1.8)
  C0 published (long only below RSI 50; short above 50): trades mean 277 (real 281); random median 101.0, 5th -75.6, 95th 237.7 $/trade; real 173.4 -> percentile 76.2
  C1 any RSI (random bar, long; or random side when shorts allowed): trades mean 285 (real 281); random median 37.2, 5th -99.8, 95th 146.9 $/trade; real 173.4 -> percentile 98.8
  C2 RSI-matched (same distribution of RSI at entry as the real signal): trades mean 291 (real 281); random median 35.3, 5th -91.3, 95th 136.6 $/trade; real 173.4 -> percentile 99.2

=== DV-LS-F ===  real: n=588 net -43,014 avg -73.2 $/trade
  real LONG entries: 251; RSI < 50 at entry: 141 (avg -58.4 $/trade, avg bars 14.9); RSI >= 50 at entry: 110 (avg -38.2 $/trade, avg bars 15.2)
  real SHORT entries: 337; RSI > 50 at entry: 225 (avg -178.7); RSI <= 50: 112 (avg 86.0)
  C0 published (long only below RSI 50; short above 50): trades mean 584 (real 588); random median -22.6, 5th -140.2, 95th 90.8 $/trade; real -73.2 -> percentile 24.0
  C1 any RSI (random bar, long; or random side when shorts allowed): trades mean 581 (real 588); random median -39.1, 5th -152.1, 95th 84.8 $/trade; real -73.2 -> percentile 32.0
  C2 RSI-matched (same distribution of RSI at entry as the real signal): trades mean 584 (real 588); random median -34.0, 5th -155.2, 95th 98.5 $/trade; real -73.2 -> percentile 29.8

=== DV-L-F ===  real: n=271 net -6,904 avg -25.5 $/trade
  real LONG entries: 271; RSI < 50 at entry: 150 (avg -31.1 $/trade, avg bars 14.9); RSI >= 50 at entry: 121 (avg -18.5 $/trade, avg bars 15.3)
  C0 published (long only below RSI 50; short above 50): trades mean 284 (real 271); random median 48.8, 5th -118.8, 95th 221.8 $/trade; real -25.5 -> percentile 26.0
  C1 any RSI (random bar, long; or random side when shorts allowed): trades mean 273 (real 271); random median 11.5, 5th -138.9, 95th 184.4 $/trade; real -25.5 -> percentile 32.8
  C2 RSI-matched (same distribution of RSI at entry as the real signal): trades mean 278 (real 271); random median 27.2, 5th -146.7, 95th 209.6 $/trade; real -25.5 -> percentile 30.6

[exploratory] DV-L-S: real signals with RSI < 50 at entry: n=147, avg 328.2 $/trade (DEV n=91 avg 241, VAL n=27 avg 203, HOLDOUT n=29 avg 720); random entries below RSI 50, same exit, 134 trades: median 98.9, 5th -143.5, 95th 335.6 -> percentile 94.2

[exploratory] DV-L-F: real signals with RSI < 50 at entry: n=150, avg -31.1 $/trade (DEV n=93 avg -178, VAL n=27 avg 63, HOLDOUT n=30 avg 341); random entries below RSI 50, same exit, 144 trades: median 39.9, 5th -228.7, 95th 322.0 -> percentile 34.0