Phase 2: statscore-simulation — bootstrap, Monte Carlo, permutation + Python

[armanrasta]

Summary

Simulation methods: bootstrap, permutation tests, Monte Carlo integration. Python bindings ship alongside.

Rust (statscore-simulation)

• Nonparametric and parametric bootstrap
• Percentile and BCa confidence intervals
• Permutation tests
• Monte Carlo and quasi-Monte Carlo (Sobol)
• Simulation-based power analysis

Python bindings (statscore.simulation)

• bootstrap(data, statistic, n_resamples) API
• Bootstrap CI coverage ≥ nominal on 10k simulations

Done when

• Bootstrap CI coverage validated
• Python API matches SciPy/bootstrap conventions