| 每輪取前幾名 | 分數上限 | s1觸發%(權益) | s1回落%(權益) | s1賣出% | s2觸發%(權益) | s2回落%(權益) | 單筆等效複利率 | 軌道獲利率 | 截斷率 |
|---|---|---|---|---|---|---|---|---|---|
| 5 | 25.5% | 0.000% | 0.434% | 10.2% | 2.278% | 0.321% | +1.716% | 91.7% | 0.0% |
| 5 | 25.2% | 0.000% | 1.141% | 10.3% | 2.498% | 0.302% | +1.713% | 91.7% | 0.0% |
| 5 | 27.4% | 0.000% | 0.584% | 10.3% | 2.841% | 0.301% | +1.712% | 91.7% | 0.0% |
| 5 | 25.3% | 0.000% | 1.114% | 10.3% | 2.820% | 0.301% | +1.709% | 95.8% | 0.0% |
| 5 | 24.5% | 0.000% | 0.813% | 11.0% | 2.651% | 0.300% | +1.701% | 87.5% | 0.0% |
| 5 | 25.4% | 0.000% | 0.414% | 10.5% | 1.662% | 0.305% | +1.700% | 91.7% | 0.0% |
| 5 | 25.5% | 0.000% | 0.742% | 10.9% | 2.571% | 0.322% | +1.697% | 91.7% | 0.0% |
| 5 | 26.6% | 0.000% | 0.316% | 10.4% | 1.706% | 0.335% | +1.694% | 91.7% | 0.0% |
| 5 | 25.8% | 0.000% | 1.095% | 10.5% | 2.296% | 0.300% | +1.686% | 91.7% | 0.0% |
| 5 | 25.1% | 0.000% | 0.421% | 10.1% | 1.283% | 0.369% | +1.685% | 87.5% | 0.0% |
| 5 | 27.1% | 0.000% | 0.708% | 10.4% | 1.433% | 0.316% | +1.680% | 91.7% | 0.0% |
| 5 | 26.3% | 0.000% | 0.399% | 11.7% | 2.834% | 0.303% | +1.677% | 91.7% | 0.0% |
| 5 | 27.8% | 0.000% | 0.958% | 10.0% | 2.305% | 0.302% | +1.677% | 91.7% | 0.0% |
| 5 | 24.7% | 0.000% | 1.215% | 11.5% | 2.970% | 0.306% | +1.666% | 95.8% | 0.0% |
| 5 | 27.9% | 0.000% | 0.444% | 10.9% | 3.101% | 0.306% | +1.662% | 91.7% | 0.0% |
| 5 | 25.5% | 0.000% | 0.605% | 10.7% | 1.371% | 0.357% | +1.659% | 91.7% | 0.0% |
| 5 | 25.8% | 0.000% | 0.410% | 11.8% | 3.050% | 0.316% | +1.657% | 95.8% | 0.0% |
| 5 | 24.2% | 0.000% | 0.991% | 10.3% | 2.263% | 0.327% | +1.657% | 91.7% | 0.0% |
| 5 | 27.8% | 0.000% | 0.573% | 10.0% | 1.934% | 0.347% | +1.654% | 91.7% | 0.0% |
| 5 | 24.5% | 0.000% | 0.762% | 11.7% | 3.138% | 0.314% | +1.653% | 91.7% | 0.0% |
| 報酬率區間 | 筆數 | 佔比 |
|---|---|---|
| (-inf, -30.0]% | 347 | 5.84% |
| (-30.0, -20.0]% | 873 | 14.68% |
| (-20.0, -10.0]% | 61 | 1.03% |
| (-10.0, -5.0]% | 66 | 1.11% |
| (-5.0, -2.0]% | 189 | 3.18% |
| (-2.0, -1.0]% | 149 | 2.51% |
| (-1.0, -0.5]% | 124 | 2.09% |
| (-0.5, 0.0]% | 157 | 2.64% |
| (0.0, 0.5]% | 159 | 2.67% |
| (0.5, 1.0]% | 172 | 2.89% |
| (1.0, 2.0]% | 346 | 5.82% |
| (2.0, 5.0]% | 773 | 13.00% |
| (5.0, 10.0]% | 920 | 15.48% |
| (10.0, 20.0]% | 842 | 14.16% |
| (20.0, 50.0]% | 608 | 10.23% |
| (50.0, 100.0]% | 141 | 2.37% |
| (100.0, +inf]% | 18 | 0.30% |
用最佳組合實際跑出的每一筆交易報酬率分桶統計，紅色=虧損區間、綠色=獲利區間， 肉眼檢查有沒有不合理的地方(例如某個出場理由的報酬範圍跟預期不符)——這個session就是靠這個 方法抓到過兩個真bug。
x軸為實際時間(涵蓋整個回測區間)，y軸為當下所有複利軌道加總的資金總額—— 用來肉眼檢查複利成長過程是不是平滑合理，有沒有突然的斷崖式下跌等異常。
綠色標示為最佳試驗。兩段式：第一段觸及s1觸發%後開始追蹤最高點，回落s1回落%就賣出s1賣出% 比例的倉位(先落袋一部分)；第二段觸及s2觸發%(通常設得比第一段高很多，確認是真正的大行情)後 改用s2回落%追蹤剩餘倉位，回落就全部賣出。2026-09-08實測發現：s1設得寬鬆、只賣一小部分, 真正的保護留給門檻夠高的s2用緊的回落%鎖利，比單階段版本(optimize_simulation_params) 在同樣修正後的模型下表現好很多。「截斷率」是交易到資料結束都還沒真正觸發出場、只能用最後 一天收盤價估報酬的比例——這是用還沒兌現的帳面數字冒充已實現獲利，搜尋範圍放寬時優化器容易 鑽這個漏洞(截斷率愈高、分數愈好看但愈不可信)，這個版本已經對截斷率設了上限(見上方meta)過濾掉 不可信的組合。僅供參考，非投資建議。