MACD control check (not part of the frozen protocol). Run 2026-10-07T18:58:26.078Z

=== MC-LS-S ===  real: n=3249 net -95852 avg -29.5 $/trade  (engine copy reproduces the published trades: yes)
  C1 state-shift (the real MACD state series, circularly shifted): trades mean 3242 (real 3249); random median -31.9, 5th -85.2, 95th 22.6 $/trade; real -29.5 -> percentile 53.6

=== MC-L-S ===  real: n=1629 net 64167 avg 39.4 $/trade  (engine copy reproduces the published trades: yes)
  C1 state-shift (the real MACD state series, circularly shifted): trades mean 1619 (real 1629); random median 37.3, 5th -16.0, 95th 91.7 $/trade; real 39.4 -> percentile 52.8

=== MC-LS-F ===  real: n=1606 net -45897 avg -28.6 $/trade  (engine copy reproduces the published trades: yes)
  C1 state-matched (random entries while MACD is above its signal / below it): trades mean 1607 (real 1606); random median -52.6, 5th -113.6, 95th 6.1 $/trade; real -28.6 -> percentile 75.8
  C2 event-shift (the real cross events, circularly shifted): trades mean 1448 (real 1606); random median -42.9, 5th -132.6, 95th 37.9 $/trade; real -28.6 -> percentile 61.2

=== MC-L-F ===  real: n=1145 net 13698 avg 12.0 $/trade  (engine copy reproduces the published trades: yes)
  C1 state-matched (random entries while MACD is above its signal): trades mean 1139 (real 1145); random median 14.7, 5th -35.0, 95th 69.6 $/trade; real 12.0 -> percentile 45.4
  C2 event-shift (the real cross events, circularly shifted): trades mean 1073 (real 1145); random median 14.0, 5th -62.8, 95th 84.1 $/trade; real 12.0 -> percentile 48.0