Return stability measures how evenly distributed a trading model's profits are across the evaluation period rather than whether those profits are high in absolute terms. A model with high return stability generates its returns consistently across many trades and conditions. A model with low return stability may show similar total returns but achieves them through a small number of outsized outcomes that are unlikely to repeat reliably.
Mutual information is a statistical measure of the degree to which two variables are genuinely dependent — how much knowing one reduces uncertainty about the other. In trading model evaluation, it is used to assess whether a model's entry signals contain real predictive information about trade outcomes, rather than a pattern that appears meaningful but lacks structural dependence.
Market regime strength describes how clearly and decisively a market is expressing a particular character — a strongly trending market is a very different environment from a weakly trending one, even though both are classified as trend-dominant. In darwintIQ, regime strength is reflected in the TrendMatrix and has a direct effect on which trading models are likely to perform well at any given moment.