Split 2016, net 10bps on toggles, sqrt(252). PASS: OOS SR>=0.50, boot-p<0.05, resid>0.
Rule Status Ret Sharpe MaxDD CorrMkt ResidSh boot-p
S1 grinder ok 0.917257 0.723457 -0.230155 0.244176 0.478212 0.0102
S2 earn-gap ok -0.590670 -0.513061 -0.615914 0.033459 -0.548958 0.9474
S3 wk-reversal-reclaim ok 0.192723 0.183143 -0.323684 0.582022 -0.536084 0.2722 S4 spy-alpha-breakout ok 4.488020 1.050377 -0.286631 0.892615 0.223904 0.0004 S5 vol-compress-expansion ok -0.227116 -0.014562 -0.503382 -0.061030 0.050447 0.5162 S6 bear-trap ok 0.584831 0.298690 -0.676273 0.509669 -0.282929 0.1636
ALL DONE
- S1 grinder placebo FAILED (random same-exposure-days basket earns median +460% vs S1 +102%) -> conditioning on 60d-no-spike-up names adds nothing over beta. Both S1 and S4 are beta-plus, not sleeve candidates. S2/S3/S5/S6 REJECT on resid/cost.