Summarize economic news
Compare each release with its consensus forecast. Scale the surprise using earlier forecast misses, then form Growth and Inflation summaries.
Earlier demonstration · historical context
Where the research started.
The original MVP demonstrated the path from an economic release to a forecast and a portfolio allocation. The current research builds on that demonstration with controlled experiments and explicit reliability checks.
Return to research overview →Original design
One signal process, three portfolio scenarios.
Forecast and moderate
Use Ridge to predict next-month asset returns. Combine those views with a neutral portfolio through Black–Litterman, accounting for uncertainty.
Allocate and compare
Adjust a starting mix of 45% SPY, 35% IEF and 20% LQD within limits. Compare with fixed 60/40 and SPY-only portfolios after trading costs.
This archive describes the earlier prototype.
Its original presentation files remain in the repository. The older interactive replay depends on data bundles absent from this checkout, so it is not included in this research release. Use the current experiment pages for available results and their sample definitions.
The original MVP, the added-signal study, the method comparison and the corrected reference model are separate analyses. Their results should not be combined into one ranking.