五色令 & 轻松跟庄 · 信号点迭代报告
2026-09-26 ｜ cache2 全池 5,384 只 ｜ 2014-01-01~2026-09-18 ｜ K=10 槽等权 · 次日开盘 · cost=0.001 ｜ 无门控 / 有门控 双口径分别迭代
✅ 口径校验：五色令-生产复现 21.51% / -25.1% / 0.856（464 笔）｜ 组合/信号重组迭代（§十一、§十二）锚点同口径复现 ✓
21.51%五色令·生产（维持）
22.18%五色令最优变体（不稳健）
10.24%轻松跟庄最优候选（mom+time60）
15.30%基准：全池等权死拿
一、结论速览
1. 五色令（生产）无更优买卖点：全维度迭代（买点执行/卖点/选槽/门槛/K槽/确认/止损）× 双口径后，有门控下生产配置 21.51%/-25.1%/0.856 仍最优；唯一微增益 gapwait5+WZ84=22.18%（+0.67pp）经随机选槽×3（均值≈基准）与年代分层（增益仅集中在 2020+）检验后不成立。无门控下生产仅 -0.09%，最优变体 yang=13.94% 仍低于基准 15.30%。→ 生产配置保持不动。
2. 轻松跟庄迭代有改善但不达标：最优候选 mom+time60@on_qs 10.24%（种子 10.5~17.0%，Calmar 0.34~0.40）；yang@on_ws 11.56%（种子 5.3~8.0%，不稳健）；delay@off 9.10%（种子 -0.3~12.3%，不稳健）。全部低于基准 15.30% 与五色令 21.51%。→ 不进生产；若继续挖掘，方向 = mom+time60 + 市场门控。
3. 数据校正：旧对比脚本 QS 侧 store 覆盖丢失缺陷已修复并重跑（§十），QS_nat 校正为 2.78%/-71.5%/283笔；WS 侧与 QS_t20 不受影响。
4. 组合/信号重组（五色令主 × 轻松跟庄辅）无增益（§十一、§十二）：固定重组（叠加/确认/补位）与全网格重组（买点/卖点/门控/选槽）全部不敌生产 21.51%；结构原因：WS 候选日 ⊂ QS 深水状态（100% zhuli<-25，深度/排序类重组全部 no-op）、QS 高抛 ⊂ WS exitw（仅 1 笔差异），有区分度的金叉/下穿事件重组均为负贡献。→ 维持单一策略。
二、五色令 · 有门控（P95）全量迭代
| 家族 | 变体 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 在场 | 止损 |
|---|
| 买点执行 | open(基准) | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 买点执行 | gapwait g=2% w=3 | 19.69% | -28.75% | 0.685 | 456 | 63.8% | +5.88% | 20.2 | 50.1% | 138 |
| 买点执行 | gapwait g=3% w=3 | 20.04% | -28.28% | 0.709 | 459 | 64.1% | +5.84% | 20.4 | 50.1% | 138 |
| 买点执行 | gapwait g=5% w=3 | 21.69% | -25.14% | 0.863 | 463 | 64.8% | +6.24% | 20.3 | 50.6% | 136 |
| 买点执行 | pullback p=1% w=3 | 13.01% | -35.20% | 0.370 | 406 | 61.1% | +4.59% | 19.8 | 49.7% | 136 |
| 买点执行 | pullback p=2% w=3 | 13.73% | -35.66% | 0.385 | 372 | 60.8% | +5.19% | 19.6 | 47.9% | 126 |
| 买点执行 | pullback p=3% w=3 | 13.45% | -33.80% | 0.398 | 345 | 62.6% | +5.53% | 19.4 | 46.8% | 117 |
| 买点执行 | pullback p=2% w=5 | 13.40% | -35.66% | 0.376 | 382 | 61.0% | +4.97% | 19.9 | 50.1% | 127 |
| 卖点/离场 | 止损 无 | 19.67% | -37.83% | 0.520 | 371 | 70.1% | +6.99% | 26.4 | 47.2% | 0 |
| 卖点/离场 | 止损 15% | 19.20% | -35.29% | 0.544 | 449 | 67.0% | +5.84% | 21.8 | 49.6% | 108 |
| 卖点/离场 | 持30日 | 20.45% | -25.52% | 0.801 | 474 | 63.5% | +5.79% | 19.0 | 45.8% | 131 |
| 卖点/离场 | 持60日 | 18.92% | -28.07% | 0.674 | 456 | 63.6% | +5.66% | 21.3 | 53.9% | 149 |
| 卖点/离场 | 持999日(纯exitw) | 18.64% | -28.07% | 0.664 | 455 | 63.3% | +5.60% | 21.7 | 57.1% | 153 |
| 卖点/离场 | 加exit2(or2) | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 卖点/离场 | exitw WZ80/BD65 | 21.74% | -25.12% | 0.866 | 474 | 66.2% | +6.05% | 18.9 | 47.8% | 137 |
| 卖点/离场 | exitw WZ80/BD70 | 21.73% | -25.12% | 0.865 | 474 | 66.2% | +6.04% | 19.0 | 48.0% | 137 |
| 卖点/离场 | exitw WZ80/BD75 | 21.73% | -25.12% | 0.865 | 474 | 66.2% | +6.04% | 19.0 | 48.0% | 137 |
| 卖点/离场 | exitw WZ85/BD65 | 21.50% | -25.14% | 0.855 | 464 | 64.7% | +6.18% | 20.2 | 50.0% | 137 |
| 卖点/离场 | exitw WZ85/BD70 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 卖点/离场 | exitw WZ85/BD75 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 卖点/离场 | exitw WZ90/BD65 | 20.89% | -26.32% | 0.794 | 453 | 61.4% | +6.21% | 22.1 | 51.8% | 139 |
| 卖点/离场 | exitw WZ90/BD70 | 20.11% | -29.12% | 0.691 | 450 | 61.1% | +6.01% | 22.4 | 52.4% | 138 |
| 卖点/离场 | exitw WZ90/BD75 | 20.29% | -28.77% | 0.705 | 450 | 61.1% | +6.06% | 22.5 | 52.5% | 138 |
| 卖点/离场 | exitw+warn1 | 15.31% | -25.68% | 0.596 | 522 | 64.2% | +4.07% | 13.7 | 41.3% | 139 |
| 选槽 | rank df(基准) | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 选槽 | rank sd_deep | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 选槽 | rank sd_shallow | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 选槽 | rank random s0 | 19.12% | -40.88% | 0.468 | 478 | 60.3% | +5.67% | 19.2 | 49.2% | 166 |
| 选槽 | rank random s1 | 18.07% | -45.14% | 0.400 | 480 | 59.0% | +5.52% | 19.0 | 52.5% | 173 |
| 选槽 | rank random s2 | 16.88% | -38.03% | 0.444 | 468 | 60.9% | +5.14% | 20.4 | 52.4% | 159 |
| 买点门槛 | SD8_DF4 (门控固定SD10口径) | 16.70% | -26.86% | 0.622 | 460 | 63.3% | +5.03% | 20.5 | 49.4% | 141 |
| 买点门槛 | SD9_DF4 (门控固定SD10口径) | 18.24% | -26.37% | 0.692 | 459 | 64.1% | +5.42% | 20.5 | 49.5% | 138 |
| 买点门槛 | SD11_DF4 (门控固定SD10口径) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 买点门槛 | SD12_DF4 (门控固定SD10口径) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 买点门槛 | SD10_DF5 (门控固定SD10口径) | 12.83% | -28.63% | 0.448 | 195 | 66.2% | +9.32% | 19.7 | 36.0% | 54 |
| 组合 | base(基准) | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 组合 | gapwait8 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 组合 | WZ78 | 20.25% | -25.35% | 0.799 | 476 | 66.4% | +5.66% | 18.6 | 47.7% | 137 |
| 组合 | WZ82 | 21.70% | -25.12% | 0.864 | 470 | 66.0% | +6.11% | 19.5 | 49.4% | 136 |
| 组合 | gapwait5+WZ80 | 21.98% | -25.12% | 0.875 | 473 | 66.4% | +6.11% | 19.0 | 48.0% | 136 |
| 组合 | gapwait5+WZ80+hold30 | 19.84% | -26.15% | 0.759 | 480 | 64.6% | +5.52% | 18.1 | 44.5% | 135 |
| 组合 | WZ80+hold30 | 19.54% | -26.15% | 0.748 | 481 | 64.4% | +5.45% | 18.1 | 44.5% | 136 |
| 组合 | gapwait5+WZ84 | 22.18% | -25.11% | 0.883 | 465 | 65.6% | +6.31% | 20.0 | 49.7% | 135 |
| 补测 | SD10_DF4_t6b | 13.14% | -35.08% | 0.375 | 434 | 59.9% | +4.32% | 19.4 | 49.1% | 143 |
| 补测 | SD10_DF6 | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| K槽位 | K5 | 17.03% | -40.67% | 0.419 | 232 | 62.1% | +5.37% | 19.9 | 45.6% | 73 |
| K槽位 | K15 | 17.83% | -33.21% | 0.537 | 700 | 62.1% | +5.20% | 19.9 | 51.8% | 222 |
| K槽位 | K10(基准) | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 50.6% | 137 |
| 买点确认 | yang(收盘>前收) | 11.02% | -33.02% | 0.334 | 240 | 62.9% | +6.71% | 19.3 | 36.3% | 74 |
| 买点确认 | ma5(站上5日线) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 止损变体 | trail10 | 12.52% | -42.85% | 0.292 | 574 | 50.9% | +3.14% | 14.4 | 46.7% | 342 |
| 止损变体 | fixed12+trail10 | 12.52% | -42.85% | 0.292 | 574 | 50.9% | +3.14% | 14.4 | 46.7% | 342 |
| 止损变体 | fixed12+be5 | 17.05% | -31.37% | 0.544 | 521 | 50.3% | +4.46% | 16.6 | 47.2% | 259 |
| 止损变体 | fixed8 | 13.56% | -39.15% | 0.346 | 551 | 48.8% | +3.62% | 15.9 | 48.6% | 269 |
| 止损变体 | fixed12+be5+trail10 | 13.25% | -43.82% | 0.302 | 595 | 46.9% | +3.16% | 13.4 | 45.3% | 381 |
三、五色令 · 无门控 全量迭代
| 家族 | 变体 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 在场 | 止损 |
|---|
| 买点执行 | open(基准) | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 买点执行 | gapwait g=2% w=3 | 0.56% | -69.89% | 0.008 | 1031 | 48.5% | +0.65% | 19.8 | 92.5% | 450 |
| 买点执行 | gapwait g=3% w=3 | -0.51% | -69.36% | -0.007 | 1031 | 48.1% | +0.53% | 19.8 | 92.1% | 450 |
| 买点执行 | gapwait g=5% w=3 | 0.18% | -69.26% | 0.003 | 1042 | 48.2% | +0.62% | 19.6 | 92.2% | 457 |
| 买点执行 | pullback p=1% w=3 | -0.46% | -69.67% | -0.007 | 1006 | 48.9% | +0.50% | 19.9 | 91.5% | 442 |
| 买点执行 | pullback p=2% w=3 | -0.01% | -67.35% | -0.000 | 990 | 47.9% | +0.49% | 19.2 | 92.3% | 449 |
| 买点执行 | pullback p=3% w=3 | 3.86% | -66.94% | 0.058 | 949 | 51.0% | +1.14% | 19.4 | 92.3% | 413 |
| 买点执行 | pullback p=2% w=5 | 1.41% | -67.40% | 0.021 | 995 | 49.2% | +0.65% | 19.3 | 92.4% | 440 |
| 卖点/离场 | 止损 无 | -2.50% | -82.81% | -0.030 | 720 | 53.6% | +0.31% | 29.3 | 91.5% | 0 |
| 卖点/离场 | 止损 15% | 1.89% | -70.72% | 0.027 | 952 | 52.2% | +0.91% | 21.7 | 92.6% | 335 |
| 卖点/离场 | 持30日 | 2.43% | -68.09% | 0.036 | 1097 | 48.5% | +0.81% | 18.3 | 92.2% | 438 |
| 卖点/离场 | 持60日 | 0.32% | -70.17% | 0.004 | 1010 | 49.7% | +0.64% | 20.7 | 92.3% | 463 |
| 卖点/离场 | 持999日(纯exitw) | 1.08% | -70.02% | 0.015 | 1006 | 50.0% | +0.78% | 20.9 | 92.9% | 467 |
| 卖点/离场 | 加exit2(or2) | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 卖点/离场 | exitw WZ80/BD65 | -1.26% | -68.54% | -0.018 | 1066 | 49.2% | +0.39% | 18.8 | 91.4% | 464 |
| 卖点/离场 | exitw WZ80/BD70 | -1.26% | -68.54% | -0.018 | 1066 | 49.2% | +0.39% | 18.8 | 91.4% | 464 |
| 卖点/离场 | exitw WZ80/BD75 | -1.26% | -68.54% | -0.018 | 1066 | 49.2% | +0.39% | 18.8 | 91.4% | 464 |
| 卖点/离场 | exitw WZ85/BD65 | 0.08% | -69.02% | 0.001 | 1042 | 48.0% | +0.60% | 19.7 | 92.2% | 457 |
| 卖点/离场 | exitw WZ85/BD70 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 卖点/离场 | exitw WZ85/BD75 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 卖点/离场 | exitw WZ90/BD65 | 5.88% | -64.77% | 0.091 | 1009 | 48.1% | +1.41% | 20.8 | 93.3% | 439 |
| 卖点/离场 | exitw WZ90/BD70 | 1.98% | -68.41% | 0.029 | 999 | 47.1% | +0.91% | 21.1 | 93.5% | 438 |
| 卖点/离场 | exitw WZ90/BD75 | 0.98% | -68.36% | 0.014 | 1000 | 46.7% | +0.77% | 21.1 | 93.5% | 442 |
| 卖点/离场 | exitw+warn1 | -7.09% | -86.37% | -0.082 | 1301 | 51.9% | -0.31% | 13.7 | 87.0% | 519 |
| 选槽 | rank df(基准) | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 选槽 | rank sd_deep | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 选槽 | rank sd_shallow | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 选槽 | rank random s0 | 2.20% | -72.23% | 0.030 | 1081 | 48.8% | +0.87% | 18.9 | 91.3% | 490 |
| 选槽 | rank random s1 | 5.64% | -64.35% | 0.088 | 1055 | 48.9% | +1.25% | 19.5 | 91.3% | 460 |
| 选槽 | rank random s2 | 1.73% | -71.32% | 0.024 | 1074 | 48.1% | +0.78% | 19.2 | 93.1% | 471 |
| 买点门槛 | SD8_DF4 (门控自口径) | -4.02% | -75.11% | -0.054 | 1117 | 47.8% | +0.11% | 19.6 | 92.4% | 487 |
| 买点门槛 | SD9_DF4 (门控自口径) | -5.47% | -78.20% | -0.070 | 1109 | 47.6% | -0.06% | 19.6 | 92.3% | 489 |
| 买点门槛 | SD11_DF4 (门控自口径) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 买点门槛 | SD12_DF4 (门控自口径) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 买点门槛 | SD10_DF5 (门控自口径) | 4.28% | -60.49% | 0.071 | 604 | 49.0% | +1.66% | 19.4 | 77.6% | 267 |
| 组合 | base(基准) | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 组合 | gapwait8 | 0.14% | -69.26% | 0.002 | 1040 | 48.1% | +0.61% | 19.7 | 92.3% | 456 |
| 组合 | WZ78 | -0.44% | -67.01% | -0.007 | 1085 | 49.7% | +0.49% | 18.3 | 90.9% | 468 |
| 组合 | WZ82 | -0.01% | -66.83% | -0.000 | 1062 | 48.8% | +0.58% | 19.0 | 92.1% | 461 |
| 组合 | gapwait5+WZ80 | -1.19% | -69.09% | -0.017 | 1069 | 49.3% | +0.41% | 18.7 | 91.2% | 465 |
| 组合 | gapwait5+WZ80+hold30 | -1.21% | -74.63% | -0.016 | 1111 | 48.3% | +0.39% | 17.7 | 91.2% | 448 |
| 组合 | WZ80+hold30 | -0.17% | -71.89% | -0.002 | 1111 | 48.7% | +0.49% | 17.7 | 91.4% | 446 |
| 组合 | gapwait5+WZ84 | 0.74% | -70.07% | 0.011 | 1045 | 48.5% | +0.68% | 19.5 | 92.1% | 455 |
| 补测 | SD10_DF4_t6b | 1.19% | -63.66% | 0.019 | 885 | 49.7% | +0.82% | 18.7 | 92.9% | 380 |
| 补测 | SD10_DF6 | -0.16% | -2.01% | -0.079 | 1 | 0.0% | -20.11% | 4.0 | 0.1% | 1 |
| K槽位 | K5 | -7.01% | -89.24% | -0.079 | 577 | 46.4% | -0.27% | 18.8 | 90.9% | 263 |
| K槽位 | K15 | 3.00% | -61.63% | 0.049 | 1481 | 49.4% | +0.97% | 19.6 | 92.8% | 637 |
| K槽位 | K10(基准) | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 92.3% | 457 |
| 买点确认 | yang(收盘>前收) | 13.94% | -49.54% | 0.281 | 632 | 54.0% | +3.39% | 19.8 | 80.9% | 248 |
| 买点确认 | ma5(站上5日线) | 0.00% | 0.00% | — | 0 | — | — | — | 0.0% | 0 |
| 止损变体 | trail10 | -0.89% | -72.03% | -0.012 | 1390 | 41.9% | +0.36% | 14.1 | 91.1% | 933 |
| 止损变体 | fixed12+trail10 | -0.89% | -72.03% | -0.012 | 1390 | 41.9% | +0.36% | 14.1 | 91.1% | 933 |
| 止损变体 | fixed12+be5 | 5.83% | -67.66% | 0.086 | 1247 | 40.7% | +1.00% | 15.7 | 91.0% | 743 |
| 止损变体 | fixed8 | -1.40% | -70.83% | -0.020 | 1272 | 39.9% | +0.33% | 15.7 | 92.2% | 734 |
| 止损变体 | fixed12+be5+trail10 | 2.48% | -71.08% | 0.035 | 1449 | 40.2% | +0.54% | 13.2 | 90.6% | 1017 |
四、五色令 · 稳健性（随机选槽×3 / 年代分层，有门控）
| 配置 | s0 | s1 | s2 | 均值 |
|---|
| base(基准) | 19.12% | 18.07% | 16.88% | 18.02% |
| gapwait8 | 19.51% | 16.96% | 17.00% | 17.82% |
| WZ80 | 19.56% | 16.08% | 17.36% | 17.66% |
| gapwait5+WZ80 | 19.06% | 18.38% | 17.59% | 18.34% |
| 配置 | 2014-2016 | 2017-2019 | 2020-2022 | 2023-2026 |
|---|
| base(基准) | 8.32% | 19.45% | 35.83% | 23.56% |
| gapwait8 | 8.32% | 19.45% | 35.83% | 23.56% |
| WZ80 | 6.56% | 17.56% | 36.97% | 26.79% |
| gapwait5+WZ80 | 6.56% | 17.56% | 36.97% | 27.69% |
基准随机选槽为 19.12%/18.07%/16.88%（均值 18.02%）——即"选槽随机性"本身即 ±1.1pp 量级；各变体均值相对基准无稳定增益。年代分层：WZ80/组合 的增益全部来自 2020-2022（+1.1pp）与 2023-2026（+3.2~4.1pp），2014-2019 为负。
五、轻松跟庄 · 无门控 全量迭代
| 家族 | 变体 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 在场 | 止损 |
|---|
| BK | raw@off | 2.78% | -71.46% | 0.039 | 283 | 58.7% | +2.56% | 105.6 | 99.8% |
| BK | d12@off | 5.32% | -71.03% | 0.075 | 293 | 58.4% | +3.50% | 102.3 | 99.8% |
| BK | d15@off | 5.21% | -77.13% | 0.068 | 278 | 57.9% | +3.48% | 107.1 | 99.8% |
| BK | d20@off | 1.34% | -80.59% | 0.017 | 280 | 56.4% | +1.79% | 105.7 | 99.8% |
| BK | reb@off | 4.80% | -83.09% | 0.058 | 292 | 61.3% | +3.11% | 101.7 | 99.9% |
| BK | yang@off | 4.99% | -81.67% | 0.061 | 302 | 61.9% | +3.23% | 98.4 | 99.9% |
| BK | rsi25@off | 3.64% | -63.70% | 0.057 | 256 | 58.6% | +3.87% | 111.9 | 99.8% |
| BK | sd8@off | -4.15% | -80.17% | -0.052 | 238 | 52.5% | -0.35% | 116.1 | 99.1% |
| BK | ma60@off | 6.53% | -62.91% | 0.104 | 296 | 59.1% | +4.35% | 100.7 | 99.8% |
| BK | delay@off | 9.10% | -62.13% | 0.147 | 287 | 62.4% | +5.09% | 103.1 | 99.8% |
| BK | mom@off | 5.50% | -53.30% | 0.103 | 386 | 58.3% | +2.37% | 77.5 | 99.9% |
| BK | combo@off | -0.01% | -71.27% | -0.000 | 220 | 50.0% | +2.19% | 123.5 | 99.0% |
| BK | bothlow@off | 2.78% | -71.46% | 0.039 | 283 | 58.7% | +2.56% | 105.6 | 99.8% |
| C | raw@off | 2.78% | -71.46% | 0.039 | 283 | 58.7% | +2.56% | 105.6 | 99.8% |
| C | delay@off | 9.10% | -62.13% | 0.147 | 287 | 62.4% | +5.09% | 103.1 | 99.8% |
| C | delay+stop15@off | 6.84% | -63.26% | 0.108 | 515 | 55.5% | +2.61% | 57.7 | 99.8% |
| C | delay+stop8@off | 2.71% | -64.25% | 0.042 | 674 | 41.1% | +1.21% | 41.1 | 99.8% |
| C | d15@off | 5.21% | -77.13% | 0.068 | 278 | 57.9% | +3.48% | 107.1 | 99.8% |
| C | d15+stop15@off | 2.03% | -73.43% | 0.028 | 443 | 51.7% | +1.53% | 61.4 | 99.9% |
六、轻松跟庄 · 有门控 全量迭代
QS 自口径门控（P95, 该买点自身计数）
| 家族 | 变体 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 在场 | 止损 |
|---|
| BK | raw@on_qs | 3.32% | -52.91% | 0.063 | 240 | 60.8% | +2.29% | 107.6 | 95.4% |
| BK | d12@on_qs | 3.39% | -53.29% | 0.064 | 239 | 60.3% | +2.34% | 108.0 | 95.8% |
| BK | d15@on_qs | 6.11% | -53.48% | 0.114 | 239 | 61.9% | +3.81% | 108.1 | 95.8% |
| BK | d20@on_qs | 2.93% | -54.62% | 0.054 | 232 | 58.6% | +2.13% | 108.0 | 94.7% |
| BK | reb@on_qs | 2.88% | -51.92% | 0.055 | 236 | 61.0% | +2.10% | 109.3 | 95.8% |
| BK | yang@on_qs | 5.22% | -53.83% | 0.097 | 238 | 61.8% | +3.45% | 108.8 | 95.8% |
| BK | rsi25@on_qs | 2.03% | -61.32% | 0.033 | 191 | 58.6% | +2.58% | 120.9 | 92.6% |
| BK | sd8@on_qs | 8.74% | -57.51% | 0.152 | 181 | 59.7% | +7.12% | 115.6 | 91.9% |
| BK | ma60@on_qs | 2.98% | -53.29% | 0.056 | 206 | 61.7% | +2.52% | 114.6 | 91.9% |
| BK | delay@on_qs | 3.05% | -49.59% | 0.061 | 243 | 61.3% | +2.04% | 105.7 | 96.6% |
| BK | mom@on_qs | 8.06% | -37.29% | 0.216 | 284 | 66.2% | +3.97% | 82.4 | 96.4% |
| BK | combo@on_qs | 6.98% | -54.67% | 0.128 | 192 | 59.4% | +5.59% | 115.0 | 95.3% |
| BK | bothlow@on_qs | 3.32% | -52.91% | 0.063 | 240 | 60.8% | +2.29% | 107.6 | 95.4% |
| C | mom@on_qs | 8.06% | -37.29% | 0.216 | 284 | 66.2% | +3.97% | 82.4 | 96.4% |
| C | mom+time60@on_qs | 10.24% | -30.05% | 0.341 | 443 | 63.2% | +3.44% | 38.6 | 86.4% |
| C | mom+time40@on_qs | 9.32% | -27.79% | 0.335 | 537 | 58.1% | +2.65% | 28.6 | 81.5% |
| C | mom+trail12@on_qs | 6.79% | -34.09% | 0.199 | 467 | 55.5% | +2.35% | 36.7 | 91.8% |
| C | mom+stop15@on_qs | 4.97% | -44.67% | 0.111 | 372 | 60.5% | +2.12% | 56.6 | 96.4% |
五色令口径门控（SD≥10 计数 P95）
| 家族 | 变体 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 在场 | 止损 |
|---|
| BK | raw@on_ws | 9.97% | -61.77% | 0.161 | 189 | 66.1% | +7.25% | 103.2 | 95.7% |
| BK | d12@on_ws | 10.79% | -61.77% | 0.175 | 187 | 66.8% | +7.90% | 105.0 | 95.7% |
| BK | d15@on_ws | 10.74% | -61.77% | 0.174 | 183 | 67.2% | +8.03% | 107.6 | 95.7% |
| BK | d20@on_ws | 11.03% | -56.16% | 0.196 | 182 | 67.6% | +8.25% | 107.9 | 95.7% |
| BK | reb@on_ws | 9.87% | -46.05% | 0.214 | 183 | 64.5% | +7.33% | 104.8 | 89.5% |
| BK | yang@on_ws | 11.56% | -45.63% | 0.253 | 185 | 65.4% | +8.39% | 104.8 | 92.1% |
| BK | rsi25@on_ws | 4.45% | -55.77% | 0.080 | 174 | 63.2% | +4.18% | 115.0 | 92.8% |
| BK | sd8@on_ws | 6.14% | -57.03% | 0.108 | 170 | 61.8% | +6.08% | 122.3 | 95.5% |
| BK | ma60@on_ws | 8.46% | -53.32% | 0.159 | 180 | 65.0% | +6.84% | 111.2 | 96.1% |
| BK | delay@on_ws | 4.58% | -57.26% | 0.080 | 201 | 62.7% | +3.69% | 102.0 | 92.2% |
| BK | mom@on_ws | 3.57% | -50.40% | 0.071 | 217 | 54.8% | +2.82% | 85.6 | 94.1% |
| BK | combo@on_ws | 7.26% | -54.01% | 0.134 | 159 | 62.9% | +7.19% | 117.1 | 91.9% |
| BK | bothlow@on_ws | 9.97% | -61.77% | 0.161 | 189 | 66.1% | +7.25% | 103.2 | 95.7% |
| C | yang@on_ws | 11.56% | -45.63% | 0.253 | 185 | 65.4% | +8.39% | 104.8 | 92.1% |
| C | d15@on_ws | 10.74% | -61.77% | 0.174 | 183 | 67.2% | +8.03% | 107.6 | 95.7% |
| C | yang+stop15@on_ws | 6.05% | -48.76% | 0.124 | 291 | 54.0% | +3.34% | 62.3 | 89.4% |
| C | d15+stop15@on_ws | 10.39% | -39.17% | 0.265 | 286 | 59.8% | +5.53% | 61.0 | 91.8% |
七、轻松跟庄 · 稳健性
| 配置 | s0 | s1 | s2 | 均值 |
|---|
| delay+stop15@off | 5.06% | 12.29% | -0.32% | 5.68% |
| mom+time60@on_qs | 17.02% | 10.51% | 14.44% | 13.99% |
| yang@on_ws | 7.53% | 5.34% | 7.98% | 6.95% |
| d15@on_ws | 0.55% | 6.14% | 8.65% | 5.11% |
| 配置 | 2014-2016 | 2017-2019 | 2020-2022 | 2023-2026 |
|---|
| delay+stop15@off | 13.58% | -10.59% | 22.49% | 5.35% |
| mom+time60@on_qs | 34.32% | -1.22% | 14.04% | 6.82% |
| yang@on_ws | 9.93% | 0.06% | 21.32% | 12.73% |
| d15@on_ws | 3.56% | 0.68% | 24.75% | 13.37% |
yang@on_ws 确定性口径 11.56%，但随机选槽 7.53/5.34/7.98 → 不稳健；d15@on_ws 0.55/6.14/8.65 波动极大；mom+time60@on_qs 10.51/14.44/17.02 相对稳定但仍有 ±3pp 量级。年代分层：mom+time60 在 2017-19 熊市 -1.22%（其余段为正）。
八、机制洞察（数字）
• 五色令买点本质是接刀：48,602 个候选日中仅 8.3%（4,051）收盘高于前收；站上 5 日线 0 个（0/48,602）。阳线确认在无门控下 -0.09%→13.94%（+14pp），有门控下 21.51%→11.02%（-10.5pp）。
• 门控是收益主体：五色令去门控 -0.09%；无门控全部变体上限 13.94% < 基准 15.30%。
• 轻松跟庄卖点稀缺：卖点 132,329 ≈ 买点 502,059 的 26%；自然版均持 105.6 天；2017-19 熊市无防御 → MDD -71.5%；rev/both 卖点年频 253 笔、年化 -11%。
• QS 买点增益方向：深超跌（d12/d15/d20）与确认类（mom/yang/delay）优于 raw；与五色令门控组合后 yang 11.56% / d20 11.03% / d15 10.74%。
九、持仓层（用户持仓 6 只，2014~2026，小样本）
| 家族 | 配置 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 |
|---|
| WS | base@on(生产) | 2.33% | -19.87% | 0.117 | 27 | 81.5% | +11.18% |
| WS | gapwait5+WZ84@on | 2.39% | -19.87% | 0.120 | 27 | 81.5% | +11.45% |
| WS | base@off | 1.93% | -19.87% | 0.097 | 29 | 72.4% | +8.67% |
| WS | yang@off | 0.54% | -1.05% | 0.512 | 3 | 100.0% | +23.06% |
| QS | raw@off | 6.83% | -30.48% | 0.224 | 101 | 67.3% | +9.20% |
| QS | delay@off | 7.02% | -30.34% | 0.232 | 101 | 67.3% | +9.39% |
| QS | mom+time60@on_qs | 3.47% | -27.26% | 0.127 | 83 | 65.1% | +5.62% |
| QS | yang@on_ws | 1.87% | -18.95% | 0.099 | 22 | 59.1% | +11.44% |
| QS | d15@on_ws | 4.46% | -21.55% | 0.207 | 33 | 66.7% | +18.81% |
持仓池仅 6 只（sz300820 英杰电气 / sz301611 珂玛科技 / sh600584 长电科技 / sz002156 通富微电 / sz002436 兴森科技 / sh603179 新泉股份），K=10 槽位无约束、样本小；ETF（159259/159558）不在 cache2。持仓层上 QS 组合反而优于 WS 生产口径（6.8~7.0% vs 2.3%），主因 WS 的候选在 6 只标的里过少（66 事件）——再次说明 WS 收益依赖全市场选股广度。
十、校正说明（旧对比脚本缺陷与重跑）
缺陷：旧脚本 QSStore 用「类级 _done + 首次 get 覆盖」注入 QS 字段；引擎缓存上限 400（全池 5,384），淘汰后重载新建 dict 时覆盖不再执行 → 引擎读到 cache2 内置的 LXXG exitw（实测 sh600009 卖点与 cache2 exitw 逐日吻合）。即旧 QS 回测的"卖点离场"大部分用的是五色令 exitw。
修复：改 dict 级标记（if '_qs_ok' not in d），淘汰重载必重新覆盖；qs_vs_wuseling.py / qs_robust.py / iter_regress.py 已修复重跑，基线重锚（WS 21.51%/464笔，QS 2.78%/283笔）。
影响：仅 QS 侧 exitw 相关数字变化；WS 与 QS_t20 不变。结论方向不变，败因更正为"卖点稀缺 → 长持 → 熊市深亏"。
| 方案 | 旧年化 | 旧MDD | 旧笔 | 旧均持 | 新年化 | 新MDD | 新笔 | 新均持 |
|---|
| QS_nat 轻松跟庄·自然 | 2.41% | -64.0% | 993 | 29.4 | 2.78% | -71.5% | 283 | 105.6 |
| QS_aligned 同约束(12%止损+40日) | 1.47% | -62.0% | 1531 | 18.8 | 0.68% | -58.6% | 958 | 30.7 |
| QS_t20 持20日 | -1.13% | -65.8% | 1438 | 19.9 | -1.13% | -65.8% | 1438 | 19.9 |
| WS_prod 五色令·生产 | 21.51% | -25.1% | 464 | 20.3 | 21.51% | -25.1% | 464 | 20.3 |
十一、组合迭代（五色令主 × 轻松跟庄辅）
命题：以五色令为主、轻松跟庄为辅，组合能否优化买卖点？口径同前（全池 5,384 只，2014-01~2026-09，K=10，次日开盘，cost=0.001，12%止损+40日上限+exitw WZ85/BD70；门控=P95 win380，WS 口径计数）。锚点校验：ws_base@on 复现 21.51%/-25.1%/464笔 ✓。
CS · 卖点组合（WS exitw + QS 辅助离场）
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| ws_base | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| ws_base | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| ws+qs_rsi85 | 无门控 | -0.08% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| ws+qs_rsi85 | 有门控 | 21.52% | -25.14% | 0.856 | 464 | 64.7% | +6.19% | 20.2 | 137 |
| ws+qs_rev | 无门控 | -6.36% | -84.12% | -0.076 | 1466 | 48.2% | -0.19% | 12.0 | 391 |
| ws+qs_rev | 有门控 | 10.80% | -35.48% | 0.304 | 568 | 56.2% | +2.71% | 11.6 | 116 |
| ws+qs_both | 无门控 | -6.36% | -84.12% | -0.076 | 1466 | 48.2% | -0.19% | 12.0 | 391 |
| ws+qs_both | 有门控 | 10.77% | -35.48% | 0.304 | 568 | 56.2% | +2.71% | 11.6 | 116 |
| qs_rsi85_only | 无门控 | -3.18% | -69.83% | -0.046 | 864 | 42.0% | +0.19% | 26.4 | 397 |
| qs_rsi85_only | 有门控 | 13.76% | -32.59% | 0.422 | 413 | 56.4% | +4.84% | 29.0 | 139 |
| qs_rev_only | 无门控 | -6.09% | -82.02% | -0.074 | 1457 | 48.6% | -0.18% | 12.2 | 388 |
| qs_rev_only | 有门控 | 10.29% | -35.93% | 0.286 | 568 | 56.0% | +2.60% | 11.8 | 117 |
CB · 买点组合（WS 买点 + QS 辅助确认/补位）
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| +qs_raw5 | 无门控 | 4.52% | -68.69% | 0.066 | 911 | 48.7% | +1.21% | 19.2 | 404 |
| +qs_raw5 | 有门控 | 14.31% | -34.01% | 0.421 | 430 | 58.1% | +4.79% | 19.9 | 147 |
| +qs_d15_5 | 无门控 | 5.78% | -69.08% | 0.084 | 896 | 49.0% | +1.31% | 19.3 | 394 |
| +qs_d15_5 | 有门控 | 15.03% | -34.01% | 0.442 | 430 | 58.4% | +5.00% | 20.0 | 146 |
| +qs_bothlow5 | 无门控 | 4.52% | -68.69% | 0.066 | 911 | 48.7% | +1.21% | 19.2 | 404 |
| +qs_bothlow5 | 有门控 | 14.31% | -34.01% | 0.421 | 430 | 58.1% | +4.79% | 19.9 | 147 |
| pool(WS主+QS补位) | 无门控 | 4.08% | -70.48% | 0.058 | 1537 | 53.6% | +0.70% | 18.7 | 507 |
| pool(WS主+QS补位) | 有门控 | 17.15% | -29.36% | 0.584 | 478 | 60.9% | +4.98% | 20.1 | 150 |
结论：组合无增益，维持单一策略。
① QS 高抛（RSI6>85 回落）与 WS exitw 几乎完全重叠：@on 21.51%→21.52%（464 笔中仅 1 笔差异）——QS 高抛从未先于 WS 离场触发。
② QS 下穿散户作辅助离场 = 中途下车：均持 20.3→11.6 天，@on 21.51%→10.80%（-10.7pp），MDD 恶化至 -35.5%。
③ 去掉 WS exitw、仅用 QS 卖点：@on 13.76%（-7.8pp）。
④ QS 金叉确认（信号前 5 日内出现 QS 超卖区金叉）在有门控下是重复过滤：事件 43,166→13,981，@on 21.51%→14.31%（-7.2pp）。
⑤ 无门控下 QS 确认可部分替代门控：-0.09%→+4.52%（d15 版 +5.78%）——但远低于生产 21.51%，即"没门控时有点用、有门控时纯属多余"。
⑥ QS 补位（WS 主 df+10 优先、QS raw 填槽）@on 17.15%（-4.4pp）：QS 信号质量低，稀释槽位。
⑦ QS raw/d15/bothlow 结构性近等价（bothlow 与 raw 在 2,000 只抽样中 100% 全等），不构成独立辅助。
机制：两套系统的离场逻辑互斥——WS 需「持到离场信号」（均持 20.3 天），QS 下穿 = 提前退出（11.6 天）；QS 高抛与 WS exitw 触发点几乎同源（同属冲高回落），叠加无增量信息。
十二、信号重组全网格迭代（组合系统整体）
对象：把两指标信号重组后的单一系统作为迭代对象（WS 为框架，QS 信号重组进 买点/卖点/门控/选槽 四个层面），全池 5,384 只，锚点 ws_base@on 复现 21.51%/-25.1%/464笔 ✓。
RB · 买点重组
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| ws_base(基准) | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| ws_base(基准) | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| 同日共现 WS∧QS金叉 | 无门控 | 8.52% | -49.75% | 0.171 | 726 | 50.4% | +2.10% | 19.1 | 318 |
| 同日共现 WS∧QS金叉 | 有门控 | 8.18% | -48.53% | 0.169 | 294 | 56.8% | +4.24% | 18.3 | 111 |
| QS先确认(前5日) | 无门控 | 4.52% | -68.69% | 0.066 | 911 | 48.7% | +1.21% | 19.2 | 404 |
| QS先确认(前5日) | 有门控 | 14.31% | -34.01% | 0.421 | 430 | 58.1% | +4.79% | 19.9 | 147 |
| WS先→QS触发入场 | 无门控 | 3.75% | -62.73% | 0.060 | 986 | 52.6% | +1.02% | 19.3 | 399 |
| WS先→QS触发入场 | 有门控 | 11.52% | -34.88% | 0.330 | 343 | 64.1% | +4.75% | 18.1 | 107 |
| 深度过滤 zhuli<-15 | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 深度过滤 zhuli<-15 | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| 深度过滤 zhuli<-20 | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 深度过滤 zhuli<-20 | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| 多重融合(深度+RSI6) | 无门控 | 0.15% | -69.02% | 0.002 | 1041 | 47.9% | +0.63% | 19.7 | 457 |
| 多重融合(深度+RSI6) | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| QS补位(WS主+QS填槽) | 无门控 | 4.08% | -70.48% | 0.058 | 1537 | 53.6% | +0.70% | 18.7 | 507 |
| QS补位(WS主+QS填槽) | 有门控 | 17.15% | -29.36% | 0.584 | 478 | 60.9% | +4.98% | 20.1 | 150 |
RS · 卖点重组
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| ws_base(exitw) | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| ws_base(exitw) | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| 或·高抛 | 无门控 | -0.08% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 或·高抛 | 有门控 | 21.52% | -25.14% | 0.856 | 464 | 64.7% | +6.19% | 20.2 | 137 |
| 或·下穿 | 无门控 | -6.36% | -84.12% | -0.076 | 1466 | 48.2% | -0.19% | 12.0 | 391 |
| 或·下穿 | 有门控 | 10.80% | -35.48% | 0.304 | 568 | 56.2% | +2.71% | 11.6 | 116 |
| 且·高抛 | 无门控 | -2.72% | -70.67% | -0.038 | 832 | 40.6% | +0.29% | 27.7 | 382 |
| 且·高抛 | 有门控 | 14.74% | -32.59% | 0.452 | 409 | 55.7% | +5.30% | 30.1 | 137 |
| 且·下穿 | 无门控 | -3.56% | -70.86% | -0.050 | 879 | 42.1% | +0.20% | 25.8 | 403 |
| 且·下穿 | 有门控 | 16.04% | -38.90% | 0.412 | 414 | 56.8% | +5.52% | 27.7 | 139 |
| 下穿仅盈利≥0 | 无门控 | -6.43% | -84.63% | -0.076 | 1321 | 55.8% | -0.20% | 14.0 | 506 |
| 下穿仅盈利≥0 | 有门控 | 13.90% | -35.26% | 0.394 | 522 | 68.8% | +3.69% | 14.0 | 136 |
| 下穿仅盈利≥5% | 无门控 | -6.79% | -83.87% | -0.081 | 1205 | 52.2% | -0.23% | 16.0 | 517 |
| 下穿仅盈利≥5% | 有门控 | 16.43% | -34.59% | 0.475 | 501 | 68.3% | +4.39% | 16.0 | 138 |
| 高抛仅盈利≥0 | 无门控 | -0.08% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 高抛仅盈利≥0 | 有门控 | 21.52% | -25.14% | 0.856 | 464 | 64.7% | +6.19% | 20.2 | 137 |
| 双确认 exitw∧(高抛∨下穿) | 无门控 | -4.15% | -73.22% | -0.057 | 885 | 41.6% | +0.10% | 25.5 | 410 |
| 双确认 exitw∧(高抛∨下穿) | 有门控 | 16.08% | -38.90% | 0.414 | 416 | 56.7% | +5.49% | 27.4 | 139 |
RG · 门控重组（有门控口径）
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| WS门控(基准) | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| QS门控 | 有门控 | 16.78% | -34.89% | 0.481 | 483 | 58.2% | +4.92% | 19.6 | 171 |
| 双门控(WS∧QS) | 有门控 | 3.79% | -42.87% | 0.088 | 136 | 61.0% | +4.32% | 17.6 | 47 |
RR · 选槽重组
| 变体 | 门控 | 年化 | MDD | Calmar | 笔 | 胜率 | 单笔 | 均持 | 止损 |
|---|
| 深度次序(同分内) | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 深度次序(同分内) | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
| 深度优先 | 无门控 | -0.09% | -69.02% | -0.001 | 1041 | 47.9% | +0.58% | 19.7 | 457 |
| 深度优先 | 有门控 | 21.51% | -25.14% | 0.856 | 464 | 64.7% | +6.18% | 20.3 | 137 |
结论：全网格重组仍无增益，原因在结构层（两条包含关系）。
① 五色令候选日 ⊂ QS 深水状态：9,665 个抽样 WS 事件 100% 满足 zhuli<-25（深度桶 9/9 全饱和）、99.93% 满足 RSI6<25 → 深度过滤 / 多重融合 / 选槽排序全部退化为 no-op（与基准逐位相同）。
② QS 高抛触发 ⊂ WS exitw 触发：叠加仅改 1 笔（21.52%），无增量。
有区分度的只剩两类事件重组，均使收益下降：金叉触发入场（同日共现 8.18%、QS 先确认 14.31%、WS 先→QS 触发 11.52%——延迟/重复过滤）与下穿离场（或·下穿 10.80%；加"仅盈利≥5%"门槛可挽回至 16.43%，仍 -5.1pp）。
门控重组：QS 门控 16.78%（-4.7pp）、双门控 3.79%（-17.7pp）；且（双条件）卖点 14.74 / 16.04 全部更差。
最终判定：维持单一五色令策略；两指标信号重组在数据上无增量信息可挖（状态类信号被包含、事件类信号为负贡献）。
附 · 数据与脚本
引擎 /root/lxxg/engine.py（迭代选项 opt-in，回归校验逐位一致）｜ 脚本 iter_ws.py / iter_ws2.py / iter_ws3.py / iter_qs.py / iter_qs2.py / iter_hold.py / iter_combo.py / iter_combo2.py ｜ 数据 data/iter_*.json ｜ 校正版对比报告
qs_vs_wuseling_20260926.html以上为数据分析，不构成投资建议。