quantmodel — Improve Vitral Multi-Asset Momentum corr vs AS

Objective

Improve Vitral Multi-Asset Momentum corr vs AS.

Notes: Vitral depth neg_over_b b=N-top_n/2: corr 0.936→0.938 CAGRΔ +1.59%. June/May AS cash+picks exact. Residual eligibility <0.95. Findings: vitral-mam-corr-t112.md

Approach

Reuses existing report research/findings/vitral-mam-corr-t112.md — tables/metrics preserved verbatim via source_report pointer. Body below summarizes grounded evidence from that report and adjacent artifacts.

Files / code / data changed

Associated commits (inferred from log grep t112):

afbdbadf research: t112 Vitral MAM AS live depth (corr 0.938) (#97)
afbdbadf research: t112 Vitral MAM AS live depth (corr 0.938) (#97)
Source report: research/findings/vitral-mam-corr-t112.md
Task-history entry: research/findings/task-history.json#t112

Results

Key table from source report (preserved):

family	best corr	notes
baseline `(n_pos/N)^1`	0.935	paper / prior port
`(n_pos/N)^q` q≈1.25–1.5	~0.940	lifts corr; no live-cash exactness
scores + 252d corr adj	~0.939	BestFolio default; hurts June picks (PDBC vs EWJ)
daily SMA / vol-adj ER	≤0.938	alloc MAE better at dpm=20; corr flat
continuous signal breadth	≤0.937	no edge
`neg_over_b` b=N−top_n/2	0.938	AS live cash exact

Conclusions and metrics are in research/findings/vitral-mam-corr-t112.md; see detail page for preserved tables/metrics.

Conclusions

Vitral depth neg_over_b b=N-top_n/2: corr 0.936→0.938 CAGRΔ +1.59%. June/May AS cash+picks exact. Residual eligibility <0.95. Findings: vitral-mam-corr-t112.md

Problems / follow-ups

See source report or PR discussion for follow-ups. No unsupported follow-ups fabricated.