When Backtests Flatter and Out-of-Sample Refuses
A beautiful backtest is only a hypothesis generator. Live and out-of-sample behaviour is where the hypothesis earns scarce capital—or quietly loses it.
Notes channel
Essays and research notes on markets, risk, and fintech — browse by channel or page through the archive.
Page 4
39 notes
A beautiful backtest is only a hypothesis generator. Live and out-of-sample behaviour is where the hypothesis earns scarce capital—or quietly loses it.
Currency risk rarely headlines investor meetings until it quietly rewrites returns, cash needs, and allocation math in your true base currency.
Scenario decks can become committee performance art. Useful scenarios stay short, decision-linked, and rehearsed before adrenaline arrives at the desk.
Crypto leverage looks new on dashboards. Easy financing, crowded collateral, and liquidation gaps are old credit lessons moving on a faster clock.
Thesis quality matters in digital assets. Access matters first: can you still reach the asset when markets, venues, or keys misbehave under stress?
High nominal yields look persuasive in a screenshot. Without a clear failure-mode story, they are advertising dressed as investment analysis.
Return models love marks and tidy charts. Liquidity decides whether those marks are optional opinions or cash you can actually use when stressed.
Forecasts win the applause in research meetings. Position sizing decides whether a wrong call stays survivable or becomes franchise risk for the book.
Asset allocation looks neat on a slide. Rebalancing is where discipline either holds or quietly dissolves under P&L pressure and ego. Make the cadence explicit.