"H10 preannouncement-drift -- combined-book diversifier test"

Window 2017-01-24..2026-06-26 (2369 days). Daily net returns, sqrt(252). Book = flagship + PEAD + TSY + crypto. Rule: ADDS if at w<=0.25 in the EW book Sharpe +>=0.05 or maxDD +>=1.0pp, with |corr|<0.5 to every sleeve.

#Sleeves

Sleeve	Sharpe	CAGR	maxDD	corrFlagship
Flagship	0.56	+9.1%	-39.4%	1.00
PEAD	0.93	+11.9%	-18.4%	0.02
TSY	0.37	+1.6%	-12.2%	-0.03
Crypto	1.08	+34.8%	-29.8%	0.02
H10	1.03	+12.8%	-16.1%	0.02

#Pairwise correlation

          Flagship  PEAD   TSY  Crypto   H10
Flagship      1.00  0.02 -0.03    0.02  0.02
PEAD          0.02  1.00  0.00    0.15  0.72
TSY          -0.03  0.00  1.00    0.01  0.01
Crypto        0.02  0.15  0.01    1.00  0.13
H10           0.02  0.72  0.01    0.13  1.00

#Combined book

Book	Sharpe	CAGR	maxDD
4-sleeve equal-weight	1.41	+15.4%	-20.4%
4-sleeve risk-parity	0.56	+2.1%	-11.3%
+H10 equal-weight	1.52	+15.0%	-18.2%
+H10 risk-parity	0.64	+2.6%	-11.7%

#H10 weight sweep into the 4-sleeve EW book

w(H10)	Sharpe	CAGR	maxDD
0.05	1.45	+15.3%	-19.9%
0.10	1.47	+15.2%	-19.3%
0.20	1.52	+15.0%	-18.2%
0.25	1.53	+14.9%	-17.7%

#Verdict: NO-ADD

best H10 weight 0.25: Sharpe 1.53 vs EW4 1.41; maxDD -17.7% vs -20.4%.
H10 corr to sleeves: Flagship 0.02, PEAD 0.72, TSY 0.01, Crypto 0.13.
H10+PEAD 50/50 blend Sharpe 1.06 vs PEAD 0.93 / H10 1.03.
ADDS -> propose a small H10 sleeve in book_allocator; NO-ADD -> H10 stays a research diversifier, not funded.