E-13. Rotation sizing measurement + capital retrial (2026-07-13)
- 730 USDT sizing scan: fraction 0.09 → -9.7% (4 trades, killed by minimum-order skips) / 0.125 → +71% (but leg 91<100 blocks K2 + loss windows skipped by accident = the threshold acts as a regime filter) / 0.25 → +46% (MDD -41%). The sizing sensitivity itself is a sign that no edge exists.
- Retrial at capital 2000 (D-015 preregistered, fraction 0.09 — threshold interference removed): OOS Sharpe 0.558 / +19.27% / MDD -17.86% / 179 trades. Excluding the best window (w22 +20.7%), the other 31 sum to -1.14%.
- Verdict: fail (D-016) — only Sharpe passed; MDD and return concentration failed. Strategy 6 rejected.
- Final diagnosis: all performance is altcoin beta from the single month of 2024-11. Edge of 0 under the defined risk. Rejected without using the sealed test — evidence that the D-010 gate is effective.
Source and migration evidence
Source: backtest/docs/experiment-log.md#E-13
Original SHA-256: 069c4aaaf53a20930ca8d94ba6365b4ce2b6bc560171d662a4c179ea55a6857c
Raw data and account information are not part of the public package. The original results do not replace statistical or scientific revalidation. The English text is a translation; numbers, tables and charts are carried over unchanged.