Optimization Suggestions

Treat suggestions as experiments

Optimization suggestions should become testable hypotheses, not decisions.

Safe workflow

1. Ask what assumption changed.
2. Run historical testing on a separate version.
3. Check cost and slippage sensitivity.
4. Review drawdown and out-of-sample behavior.
5. Document why the change was accepted or rejected.

Overfitting risk

Repeated parameter changes can fit past data too closely. Keep a written experiment log.