The current implementation of ensemble_compute_statistics_ formulae may contain incorrect or inconsistent statistical logic when aggregating results across ensemble members. This issue proposes a thorough review and correction of the formulas used to compute key statistical metrics such as mean, variance, standard error, and confidence intervals. The goal is to ensure mathematical correctness, alignment with standard statistical definitions, and consistency across different ensemble modes.