Rank 144 · intraday volatility commonality · minimal clean replication

Desk verdict：P0 / park / evidence only

这次不再复读 digest，而是把阈值冻结成 rv_z > 1.0、把 universe 固定成 BTC/ETH/SOL，看 short 侧改善是不是还站得住。

关键读法：short 高共振桶相对低共振桶改善 -2.10 bp；但 best bucket 仍未转正，且只做到 2/3 个币改善，所以不升 P2。

pooled summary

side	bucket	n	mean_net_bp	win_rate
long	c0_1	3630	-21.57	0.34
long	c2_3	555	-18.93	0.39
short	c0_1	3630	-25.75	0.33
short	c2_3	555	-27.85	0.35

asset split summary

asset	side	bucket	n	mean_net_bp	win_rate
BTC-USD	long	c0_1	1210	-22.77	0.32
BTC-USD	long	c2_3	185	-23.69	0.37
BTC-USD	short	c0_1	1210	-24.78	0.31
BTC-USD	short	c2_3	185	-23.64	0.34
ETH-USD	long	c0_1	1210	-20.25	0.37
ETH-USD	long	c2_3	185	-20.58	0.41
ETH-USD	short	c0_1	1210	-26.96	0.33
ETH-USD	short	c2_3	185	-26.11	0.37
SOL-USD	long	c0_1	1210	-21.69	0.35
SOL-USD	long	c2_3	185	-12.52	0.39
SOL-USD	short	c0_1	1210	-25.52	0.35
SOL-USD	short	c2_3	185	-33.80	0.35

scorecard

• usefulness = 1/3
• time_stability = 1/3
• cross_asset_stability = 2/3
• cost_trade_stability = 1/3
• deployability = 2/3
• hard-fail flags = not_post_cost_positive_even_in_best_bucket, cross_asset_asymmetry_persists_in_clean_rep, router_attachment_still_not_done
• recommended_action = park
• why_now = 冻结阈值后做一次真正会改变 verdict 的最小 clean replication。
• main_weakness = 还没接 router，也没做更长窗口时间稳定性。

artifact

• reports/artifacts/scout_rank144_intraday_vol_commonality_15m/asset_split_summary.csv
• reports/artifacts/scout_rank144_intraday_vol_commonality_15m/pooled_summary.csv
• reports/artifacts/scout_rank144_intraday_vol_commonality_15m/threshold_config.csv
• reports/artifacts/scout_rank144_intraday_vol_commonality_15m/promotion_scorecard.json