Optimization Stability

Optimization stability refers to the consistency of a strategy's performance parameters over different time periods. A stable strategy will show similar optimal parameters when trained on different subsets of data, indicating that the logic is capturing a fundamental market truth rather than a temporary anomaly.

If the optimal parameters change drastically with every new dataset, the strategy is likely unstable and prone to failure. Evaluating stability is a key step in ensuring that a model is ready for live deployment.

It provides a measure of confidence that the strategy will continue to perform as expected in the future.