# Calibration of the London-breakout test — summarised 2026-09-05 18:25 UTC

This is a test record, not advice.

Built by `summarise_log.py` from the per-year lines of `calibrate.py` runs (logs: `calibrate.log`, `calibrate-edge.log`). Null arm: every 1-minute mid return of every real 2026 day multiplied by an independent random sign, spreads and wick sizes kept. Edge arm: the same, plus a drift in the breakout direction after the day's first breakout (see the run's own header for the size). Full pipeline unchanged: rule, random-entry control, halves.

| arm | years | mean trades | excess mean | excess s.d. | ≥2 s.e. | ≥2 s.e. and halves agree | ≥1.5 s.e. | ≥1.5 and agree |
|---|---|---|---|---|---|---|---|---|
| null (sign-flipped returns) | 80 | 105 | +0.05 | 1.07 | 5.0% | 5.0% | 10.0% | 8.8% |
| edge injected | 20 | 105 | +4.06 | 1.03 | 100.0% | 100.0% | 100.0% | 100.0% |

Under a calibrated statistic the null column reads s.d. ≈ 1.00 and ≥2 s.e. ≈ 2.3%.

Excess statistic per null year (sorted):

    -1.95 -1.79 -1.71 -1.70 -1.68 -1.49 -1.46 -1.30 -1.29 -1.28 -1.23 -1.09 -1.08 -0.99 -0.99 -0.94 -0.89 -0.89 -0.88 -0.83 -0.83 -0.60 -0.57 -0.55 -0.48 -0.42 -0.42 -0.42 -0.42 -0.36 -0.36 -0.34 -0.32 -0.26 -0.20 -0.16 -0.06 -0.06 -0.03 -0.01 -0.00 +0.01 +0.03 +0.13 +0.20 +0.21 +0.22 +0.24 +0.27 +0.29 +0.34 +0.39 +0.43 +0.45 +0.51 +0.54 +0.54 +0.60 +0.62 +0.64 +0.67 +0.76 +0.76 +0.96 +1.01 +1.14 +1.18 +1.19 +1.23 +1.28 +1.37 +1.40 +1.53 +1.61 +1.82 +1.99 +2.10 +2.37 +2.51 +2.69

Excess statistic per edge year (sorted):

    +2.71 +2.86 +2.93 +3.16 +3.23 +3.30 +3.34 +3.46 +3.80 +3.99 +4.03 +4.07 +4.18 +4.19 +4.44 +4.72 +4.85 +4.92 +6.40 +6.54