quantmodel — Investigate and close BAA Balanced CAGR discrepancy vs AllocateSmartly

Objective

Investigate and close BAA Balanced CAGR discrepancy vs AllocateSmartly.

Notes: Root cause: allocatesmartly() same_close was lookahead; AS close-fill = next_close. Also missing 18mo warmup in runners. After fix: BAA corr 0.999 CAGRΔ +0.50% → replicated; HAA corr 0.984 Δ +0.51%. Findings: baa-cagr-discrepancy.md.

Approach

Reuses existing report research/findings/baa-cagr-discrepancy.md — tables/metrics preserved verbatim via source_report pointer. Body below summarizes grounded evidence from that report and adjacent artifacts.

Files / code / data changed

No commit SHA confidently associated with t14 in sampled log [reconstructed — uncertain].
Source report: research/findings/baa-cagr-discrepancy.md
Task-history entry: research/findings/task-history.json#t14

Results

Key table from source report (preserved):

Mode (with 18mo warmup)	Overlap CAGR Δ	Monthly corr
`same_close` + T-bill 10bps	+1.87%	0.961
`next_close` + T-bill 10bps	+0.49%	0.999
`next_close` + zero cash	−0.21%	0.998
`next_close` + T-bill 20bps	+0.05%	0.999

Conclusions and metrics are in research/findings/baa-cagr-discrepancy.md; see detail page for preserved tables/metrics.

Conclusions

Root cause: allocatesmartly() same_close was lookahead; AS close-fill = next_close. Also missing 18mo warmup in runners. After fix: BAA corr 0.999 CAGRΔ +0.50% → replicated; HAA corr 0.984 Δ +0.51%. Findings: baa-cagr-discrepancy.md.

Problems / follow-ups

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