When Systematic Investing Fails

advanced7 min read

Regime change, crowding and decay — the honest limits of any backtested system.

Systematic investing is powerful, but it’s not magic — and an honest course must cover its limits before teaching the tools. Even a genuinely good, well-tested system can stop working, and knowing why is what lets you respond instead of blindly riding it into the ground.

The deepest truth of systematic investing is that the market is non-stationary — it changes, sometimes permanently, so a system built on the past can quietly become obsolete. Three forces kill systems: (1) Regime change — the conditions the strategy relied on shift (a low-rate era ends, volatility regime flips), and the past stops resembling the future. (2) Crowding — once an edge is discovered by enough people, their trading competes it away; the very success of a strategy attracts the capital that erodes it. (3) Decay — edges naturally fade as markets grow more efficient. This is why a backtest is a hypothesis, not a guarantee, and why later modules obsess over robustness, out-of-sample testing and monitoring for decay. The mature systematic trader holds their system with conviction and humility — trusting it through normal variance, but always watching for the signs it has genuinely stopped working.

Regime change — the market conditions the system depended on shift; past patterns no longer hold.
Crowding — an edge discovered by many is traded away; success attracts the capital that erodes it.
Decay — edges naturally fade as markets become more efficient over time.
The takeaway — a backtest is a hypothesis; combine conviction (ride normal variance) with humility (monitor for real decay).

ExampleA factor that worked beautifully for a decade gets packaged into popular ETFs; billions pour in, bidding up exactly the stocks it buys — and the edge flattens or inverts. Nothing was “wrong” with the backtest; the strategy’s own popularity (crowding) and a shifting regime quietly killed the edge.

Key takeawayEven good systems fail because markets are non-stationary: regime change (conditions shift), crowding (success attracts capital that erodes the edge), and decay (markets grow efficient). A backtest is a hypothesis, not a guarantee — hold systems with conviction and humility, and monitor for genuine decay.

FAQs

If systems inevitably decay, why build them at all?

Because a well-built system still captures a real edge *while it lasts*, with discipline and measurability a discretionary approach can’t match — and good practice (robustness testing, diversification across edges, decay monitoring) extends its life and limits the damage when it fades. The goal isn’t a system that works forever; it’s a rigorous process for finding, exploiting, and retiring edges responsibly.