Research · Amount × RRG
金額門檻 × RRG 疊加 Sweep
固定／自適應金額門檻，疊加 RRG 象限與動能閘門；對照純金額 baseline。
窗口 2024-07-01 → 2026-07-17
cells 1600
來源 20260718_amount_rrg_copytrade_sweep.json
假設檢定
H_RRG_IMPROVES_APLD：passed · Some RRG gate beats amount-only (xd fixed) on α/日
H_RRG_IMPROVES_RET：passed · Some RRG gate beats amount-only on period return
H_ADAPT_P90：rejected · Adaptive P90-60 floor beats best fixed xd amount-only
冠軍
主冠軍 · 期間報酬
PRIMARY
L2H12
- mode / thr / gate
- xd · 80M · rs_ratio_gt100
- 期間報酬
- 414.22%
- α / α/日
- 34,761 · 90.52
- 槽 / 捕獲
- 1 · 21.77%
- gate pass / drop
- 154 · 109
次優 · α
ALPHA
L1H12
- mode / thr / gate
- sum · 50M · none
- 期間報酬
- 194.85%
- α / α/日
- 123,036 · 33.84
- 槽 / 捕獲
- 9 · 86.57%
- gate pass / drop
- 363 · 0
次優 · α/日
EFFICIENCY
L2H12
- mode / thr / gate
- and · 150M · leading
- 期間報酬
- 67.92%
- α / α/日
- 4,665 · 129.60
- 槽 / 捕獲
- 1 · 50.00%
- gate pass / drop
- 6 · 11
建議規格
- 金額 ≥ 80M（fixed）
- RRG gate = rs_ratio_gt100
- mode xd · L2H12 · Top-1 · 1 槽
- 對照純金額 baseline：100M L2H12 ret 323.98%
RRG gate vs 純金額（xd fixed）
| gate | 勝 ret | 勝 α/日 | 最佳規格 | ret% | α/日 | n_sig |
| leading | Y | Y | 50M L2H12 | 381.75% | 95.32 | 118 |
| lead_improve | N | Y | 150M L1H12 | 289.06% | 66.32 | 77 |
| not_lagging | N | N | 100M L2H12 | 298.82% | 63.30 | 152 |
| rs_mom_gt100 | N | Y | 150M L1H12 | 289.06% | 66.32 | 77 |
| rs_ratio_gt100 | Y | Y | 80M L2H12 | 414.22% | 90.52 | 147 |
| mono_fresh | N | Y | 50M L1H18 | 70.77% | 109.57 | 5 |
PIT: branch tape + RRG close ffill on signal day. Cost 0 bps. Research only.