Sophisticated trading organizations rarely depend on a secret algorithm alone. Their advantage is often the discipline wrapped around prediction. Institutional teams design for uncomfortable realities: forecasts miss, market data degrades, infrastructure breaks, and markets produce regimes absent fr

Early in 2012, a portfolio in JPMorgan Chase’s Chief Investment Office was running over its risk limits. Then a new value-at-risk model was introduced, and the problem appeared to disappear. The new model roughly halved the portfolio’s reported risk. Months later, the trades that beca