ADAPT WITH EVIDENCE
Pause a strategy with rules, not with panic.
A strategy should have keep, reduce, pause and retire conditions before a difficult period begins. Regime change is a research hypothesis to test, not a label applied after every loss.
Before you begin
Prerequisites
Learning objectives
- Explain: Define the monitored assumptions
- Explain: Use multiple stopping signals
- Explain: Research before retuning
Define the monitored assumptions
List the market structure, data, execution, cost and risk assumptions required for the strategy to remain valid. Monitor them separately from P&L so one losing trade is not treated as structural failure.
Use multiple stopping signals
A pause decision is stronger when independent evidence agrees.
- Drawdown or loss sequence breaches a pre-written limit
- Execution cost or fill quality leaves the validated range
- Signal frequency or market structure changes materially
- Release monitoring finds an operational defect
Research before retuning
Freeze the affected version, preserve the evidence and open a separate research branch. Any replacement must pass the same IS, OOS, stress and release gates instead of being repaired directly on the failed period.
Common mistakes
- Reading the result without the stated scope and assumptions.
- Changing the rule after seeing an outcome while still calling the data unseen evidence.
Trader Checklist
- Drawdown or loss sequence breaches a pre-written limit
- Execution cost or fill quality leaves the validated range
- Signal frequency or market structure changes materially
- Release monitoring finds an operational defect
Practice exercise
Choose one of your own trading examples and write one page of rules, evidence and stopping conditions using the principles in "Pause a strategy with rules, not with panic.".
What evidence would overturn the conclusion
The conclusion should be overturned or narrowed if a key assumption cannot be reproduced inside the supported scope, the control cannot be executed, or new out-of-sample evidence repeatedly contradicts it.