Notes from the lab.
Methodology, paper reviews, regime commentary, and multi-strategy framing — written for practitioners and discussed at the level of the underlying signal.
- Methodology · Calendar effects · Regime
Calendar effects as conditioning variables, not signals
Calendar anomalies — turn-of-month, FOMC drift, end-of-quarter rebalancing — fail as stand-alone signals but earn their place as conditioning variables that gate when other signals are allowed to trade.
7 min read - Statistics · Regime · Risk
Regime filtering without overfitting: a practitioner's checklist
Regime models help when they reduce the variance of expected returns conditional on observable state. They hurt when they manufacture states that fit history but not the next month.
9 min read - Portfolio construction · Risk · Volatility
Volatility targeting in a multi-strategy book
Volatility targeting at the strategy level is well-understood. At the portfolio level, in a book with partially correlated strategies, the implementation choices have first-order consequences for drawdown behaviour.
8 min read