How to Tell Your Trading System Is Broken

Every system goes through drawdowns; that is normal. But sometimes a drawdown means not an unlucky run but that the edge has stopped working. Telling one from the other is hard and critical: abandoning a working system too early is as harmful as clinging to a broken one too long. Let's go through how to tell that a system has genuinely broken.

Normal drawdown vs. breakage

A normal drawdown is a statistically expected run of losses that fits the system's historical profile. Breakage (edge decay) is when the system consistently performs worse than across its whole history and goes beyond its usual drawdowns. The key question: does the current drawdown resemble what the system went through before, or is it deeper and longer than anything that came before? If it fits the historical range, it is most likely an ordinary run.

Signs of real breakage

Breakage is indicated not by individual losses but by persistent deviations. The drawdown is deeper and longer than any historical one. Key metrics (win rate, average profit, expectancy) have steadily worsened over a large sample, not over a couple of trades. The market conditions the system was built for have changed (for example, a trend system stopped working because of a prolonged flat, or the volatility regime shifted). Several signs coinciding at once carry more weight than any one alone.

Why you can't judge by a few trades

The main mistake is declaring a system broken after a losing streak that is in fact statistically normal. A few losses in a row prove nothing: they happen to any profitable system. A conclusion of breakage requires a sufficient sample, where the deterioration of metrics becomes persistent rather than random. Abandoning a working system during a normal drawdown is a common way to never reach profit, hopping from one strategy to another at the worst moment of each.

What to do when you suspect breakage

On a justified suspicion (several signs over a large sample), it is wise not to abandon the system abruptly but to reduce risk and investigate. Have the market conditions for it changed? Has an error crept into execution? Compare the current statistics with the historical ones on key metrics. Sometimes a system needs not cancellation but adaptation to a new regime. The decision to switch a system off should rest on data (persistent deterioration of metrics), not on emotion after a painful streak.

The practical takeaway

Telling that a system has broken means distinguishing a normal drawdown (fits the historical profile) from edge decay (persistent deterioration of metrics over a large sample, a drawdown deeper and longer than any before, changed market conditions). Don't judge by a few trades: a losing streak is statistically normal and proves nothing. A conclusion of breakage requires a sample and several coinciding signs. On a justified suspicion, reduce risk and investigate (market conditions, execution errors, metric comparison) rather than abandoning the system in panic. Understanding the difference between bad luck and breakage protects you from two extremes, clinging to a dead system and abandoning a working one at the worst moment, and is part of mature trade management.

This material is for educational purposes and is not individual investment advice.