advanced-math-trading/information-causal

Mutual information quantifies how much one signal reduces uncertainty about another. This Skill computes MI between price and candidate indicators using information theory, helping you identify which signals contain predictive relevance for trading decisions.

Yes. This Skill performs Granger-causality and DAG-based causal checks to confirm that signals causally influence price rather than merely correlating with it, reducing false signal discovery in backtesting and live scenarios.

Correlation only measures co-movement; causal inference determines whether one variable actually drives another. This Skill uses transfer entropy and Granger tests to distinguish causality from spurious correlation, strengthening signal validation.

Bayesian conditioning updates prior probabilities as new evidence arrives. This Skill implements conditional probability updates, letting you refine signal confidence incrementally as fresh data arrives during backtesting or live trading.

This Skill wraps information-theoretic and causal inference methods, but understanding mutual information, transfer entropy, and Granger causality concepts helps interpret results. The dependencies are numpy, pandas, and scipy.

Transfer entropy measures directed information flow between time series, revealing which signals lead others. This Skill estimates TE to identify predictive signal ordering and validate causal chains in multi-signal strategies.