Knowing when to retire a strategy

[Amanhasnoname]

All edges decay eventually. The hard question is when to stop. Retire too early, and a working system is lost. Retire too late, and capital is burned.

Define the criteria in advance, before emotions interfere. A statistical rule might be that the rolling expectancy falls below a threshold for a certain number of trades, or the drawdown exceeds a multiple of the backtest maximum.

Also consider structural changes: new broker conditions, changes in market structure, or a strategy's dependence on a pattern that has disappeared. If the reason for the edge no longer exists, the system should be stopped regardless of recent results.

A reduction in size is an intermediate step. Instead of an on-off decision, scale down as evidence of decay accumulates, and scale up if performance recovers.

Keep the research pipeline active. New ideas replace old ones, and the portfolio of strategies evolves.

An honest review at fixed intervals, with predefined criteria, prevents both denial and panic.

What criteria would trigger the retirement of the current main strategy?