生成时间：2026-03-19 08:58 UTC ｜ 最小 clean replication：固定复用 BTC/ETH/SOL 120d 15m 本地 cache；只测试 shared re-arm gate，不扩写新框架；统一执行 signal 当根及之前数据 + next-bar open + no-overlap + hold 8 bars。
先核对 desk 状态：全 desk 仍无 due-now / overdue；最近 due 点仍是 美股 1d+1wk（SPY/QQQ/AAPL） -> 2026-03-19 20:00 UTC。 manual narrow-paper 最新 refresh @ 2026-03-19T08:35:53Z，new_closed_trades_appended=0。
re-arm 口径冻结：long 侧先要求一次 fresh pullback（close < ema21）把状态置为 armed，再要求 <= 12 根内 close 重新站回 ema21 才允许下一次 continuation/retest；short 侧镜像。hard_invalidation 再把止损钉到 pullback_extreme ± 0.5 ATR。
6bps/side desk 级结果：
baseline：mean total return ≈ -1.97%，mean avg net ret ≈ -0.065%，3-bar fail ≈ 24.67%rearm_gate：mean total return ≈ -0.04%，mean avg net ret ≈ 0.139%，retention ≈ 23.41%，3-bar fail ≈ 25.34%rearm_gate_plus_hard_invalidation：mean total return ≈ 0.60%，mean avg net ret ≈ 0.209%，retention ≈ 23.41%，stopped-early ≈ 22.35%Hard verdict：park / evidence_pool。这轮最小 clean replication 没证明 re-arm 状态机能把三条 archetype 一起修好：要么收益没更好，要么只是靠大幅砍交易数换来的。当前更诚实的结论是 park。
| variant | mean_total_return | positive_cell_ratio | mean_trades | mean_trade_count_retention | mean_avg_net_ret | mean_early_fail_3bars_rate | mean_stopped_early_rate | mean_armed_gap_bars |
|---|---|---|---|---|---|---|---|---|
| baseline | -1.97% | 33.33% | 22.0 | 100.00% | -0.06% | 24.67% | 0.00% | 22.95 |
| rearm_gate | -0.04% | 44.44% | 5.3 | 23.41% | 0.14% | 25.34% | 0.00% | 3.80 |
| rearm_gate_plus_hard_invalidation | 0.60% | 44.44% | 5.3 | 23.41% | 0.21% | 25.34% | 22.35% | 3.80 |
| setup | variant | mean_total_return | positive_asset_ratio | mean_trades | mean_trade_count_retention | mean_avg_net_ret | mean_early_fail_3bars_rate |
|---|---|---|---|---|---|---|---|
| breakout_short | baseline | -3.55% | 0.00% | 20.3 | 100.00% | -0.20% | 10.90% |
| breakout_short | rearm_gate | 0.00% | 0.00% | 0.0 | 0.00% | - | - |
| breakout_short | rearm_gate_plus_hard_invalidation | 0.00% | 0.00% | 0.0 | 0.00% | - | - |
| ema_psar_long | baseline | -3.55% | 33.33% | 34.7 | 100.00% | -0.10% | 29.78% |
| ema_psar_long | rearm_gate | -1.65% | 33.33% | 11.7 | 32.35% | -0.11% | 20.13% |
| ema_psar_long | rearm_gate_plus_hard_invalidation | 0.61% | 33.33% | 11.7 | 32.35% | 0.10% | 20.13% |
| fib_retest_long | baseline | 1.18% | 66.67% | 11.0 | 100.00% | 0.11% | 33.33% |
| fib_retest_long | rearm_gate | 1.54% | 100.00% | 4.3 | 37.88% | 0.39% | 30.56% |
| fib_retest_long | rearm_gate_plus_hard_invalidation | 1.19% | 100.00% | 4.3 | 37.88% | 0.32% | 30.56% |