When Stop Losses Support or Undermine a Trading Edge

Summary

The article argues that a stop loss is useful only when losses carry information about likely future returns. For a signal based on a factor such as sentiment, a falling position value does not by itself show that the signal has weakened. Exiting solely because of negative P&L can therefore override the strategy’s predictive signal and require a later re-entry while that signal still holds.

Trend systems are presented as a case where stops can fit the source of the edge: recent declines may predict further declines, making negative performance relevant to future returns. The article says some crypto markets and certain commodities have shown trend characteristics, but advises establishing that evidence rather than assuming it. If stops improve a backtest, the gain may reflect an unmodeled trend or overfitting to historical noise. The discussion offers no comparative test results or detailed stop-setting method; it recommends investigating the edge, considering position sizing alternatives, and treating unusually strong backtests skeptically.

Key ideas

• A loss alone does not establish that a non-trend trading signal has lost its predictive value.
• Stops can align with a strategy when recent negative returns predict further declines.
• Stop improvements in historical tests may indicate either an overlooked trend or overfitting.
• Traders should connect exit rules to the source of the strategy’s edge.
• Alternative position sizing can manage exposure without a binary stop.

Tags

Cited by

• Strategies Perp-Led Selloff Absorption Panel: buy a Binance USD-M alt perp at the next 4H open when its 24h return is <= -2 sigma AND its perp-vs-index premium is simultaneously <= -1.5 sigma (the selling came through the derivative, not spot), exit on a FIXED 12h clock; long-only, 12% per leg, max 5 concurrent, flat ~93% of the time (10 pre-2021 liquid USD-M perps, 4H)
• Hypotheses Perp-Led Selloff Absorption Panel: buy a Binance USD-M alt perp at the next 4H open when its 24h return is <= -2 sigma AND its perp-vs-index premium is simultaneously <= -1.5 sigma (the selling came through the derivative, not spot), exit on a FIXED 12h clock; long-only, 12% per leg, max 5 concurrent, flat ~93% of the time (10 pre-2021 liquid USD-M perps, 4H)
• Hypotheses Delayed Second-Leg After an Alt Upside Shock: Buy Binance SPOT Alt 24h AFTER a >=2.5-Sigma 24h Rally, Hold a Fixed 48h (10 spot majors, 4H, long-only cash, skip the absorption day)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.