Here's a question that separates serious quants from hopeful ones: how long does your signal actually stay good? Not how long you hold the position — how long the original insight remains valid. It sounds simple. It isn't. Most traders backtest a strategy, see a decent Sharpe ratio, and never ask whether the edge exists at one hour, one day, or one week. That oversight is expensive.

Signal decay describes how a predictive relationship weakens as the holding period extends. Think of it like a hot tip at the footy. The moment your mate texts you the inside info, it's gold. By the time it's on the radio, it's noise. Every signal has that same arc — peak information value, then a slow (or fast) slide toward irrelevance. The job of a rigorous algo trader is to find exactly where that slide begins.

CONCEPTSignal decay rate tells you the optimal holding period — where edge is maximum before noise overwhelms your predictive advantage.
WARNINGHolding a position beyond your signal's half-life doesn't reduce risk — it replaces your edge with randomness at full cost.
KEY IDEAA signal with a 4-hour half-life run as a daily strategy isn't conservative — it's a different, untested strategy entirely.

The practical measurement approach is straightforward once you know what to look for. Take your signal values at entry and regress forward returns at increasing intervals — 30 minutes, 1 hour, 4 hours, 1 day, 3 days. Plot the information coefficient (IC) at each horizon. You're looking for the curve's peak and, critically, when it crosses zero. That zero-crossing is your signal's expiry stamp.

30m 1h 4h 1d 2d 3d 5d IC 0 Edge gone Information Coefficient vs Holding Period

What the research from quantitative finance literature consistently shows is that many momentum and mean-reversion signals have surprisingly short half-lives — often measured in hours, not days. When you extend the hold beyond that window, transaction costs remain fixed but expected edge shrinks toward zero. The maths turns ugly fast. Measuring this rigorously requires out-of-sample IC analysis, not just in-sample curve-fitting, which is where most DIY backtests quietly fail. Resources like the information ratio framework help contextualise signal strength, while the broader mechanics of algorithmic trading explain why execution timing compounds the decay problem. Understanding how alpha decays relative to benchmarks frames why this isn't just theoretical — it's the difference between a live strategy and an expensive hobby.

The practical takeaway: before you run any strategy live, plot your signal's IC across at least six holding periods and find the peak. Set your exit rules to that window, not to a round number someone on a forum suggested.

Your edge has an expiry date — the only question is whether you read the label before trading.

This content is for educational purposes only and does not constitute financial product advice. Past performance is not indicative of future results. Profit Logic Ltd (ACN 688 669 936) accepts no responsibility for errors or omissions in this content or anywhere on this website. Always seek advice from a licensed financial adviser before making investment decisions.