Alpha decay

Alpha decay is the erosion of a trading signal's or strategy's predictive power over time, measured as the shrinking correlation between the signal and subsequent returns, or the shrinking profit the signal generates, as time passes since the signal was generated or since it was first discovered. Every signal has some decay profile: some fade within minutes, others over months.

Decay happens for two related reasons: the information content of a signal genuinely goes stale as the market absorbs it, and a strategy's edge can be crowded out as more participants identify and trade the same pattern, competing away the excess return it once captured. This is distinct from ordinary strategy underperformance — alpha decay specifically describes a persistent downward trend in a signal's effectiveness rather than random variation in results from period to period.

Quantitative researchers study alpha decay to decide how quickly a signal must be acted on, how much trading cost it can absorb before decay erases the edge, and when to retire a strategy entirely. It is closely tied to work with alternative data and point-in-time data, since a newly available dataset often carries the richest alpha before it decays as more funds gain access to it. A common pitfall is backtesting a strategy without modeling decay, which overstates how long a live signal will remain profitable.