| Date | Kind | Title | Tags | Status |
|---|---|---|---|---|
| 2026-08-21 | Paper | Execution Timing and the Cost of ImmediacyUrgency is a cost, not a free variable — timing studies belong inside the backtest rather than beside it. | working note · rev 2026-08-21 | |
| 2026-08-13 | Paper | The Half-Life of Short-Horizon FactorsDecay dominates the signal as turnover rises; the half-life of a factor is estimated, not assumed. | working note · rev 2026-08-13 | |
| 2026-08-05 | Paper | Survivorship Bias in Cross-Sectional BacktestsA universe that excludes delistings silently flatters every long-only factor; membership must be point-in-time. | published · rev 2026-08-05 | |
| 2026-07-28 | Paper | Point-in-Time Data and the Revision ProblemRestatements leak the future into the past; the as-of view is the only view that survives audit. | published · rev 2026-07-28 | |
| 2026-07-16 | Paper | Transaction Costs as a First-Class SignalImplementation shortfall is a signal about capacity and competition, not friction to subtract later. | working note · rev 2026-07-16 |
| Date | Kind | Title | Tags |
|---|---|---|---|
| 2026-07-08 | Note | Drawdown Regimes and Explicit Risk BudgetsA risk budget should respond to the state estimate rather than to the drawdown it is reacting to. | |
| 2026-06-29 | Note | On the Fragility of Overfit Sharpe RatiosIn-sample Sharpe is a selection statistic, not a forecast; multiplicity does the rest of the damage. | |
| 2026-06-18 | Note | Factor Crowding and Capacity DecayCrowding is slow-burn capacity decay; proxies for shared positioning precede the reversion. | |
| 2026-06-08 | Note | Regime-Conditional Volatility TargetingTargeting volatility without conditioning on the state estimate is leverage by another name. |