Монте-Карло симуляция на R

data

simulation-study is a Claude Code skill that scaffolds and runs a reproducible Monte Carlo simulation study in R, covering a parameterized data-generating process (DGP), an estimator grid, a seeded replication loop, and a summary table reporting bias, RMSE, empirical SE, confidence interval coverage, size/power, and Monte Carlo standard errors. Before any code is written, the skill produces a Pre-Flight Report that pins down the target estimand and how its true value is derived from DGP parameters — preventing the most common failure modes such as estimand mismatch or coverage-against-the-estimate bugs. The numbered R script is saved to `scripts/R/`, raw per-replication results and summary tables go to `scripts/R/_outputs/`, and the sim-reviewer agent is automatically invoked to catch critical issues before results are presented. Designed for econometricians and statisticians who need to examine finite-sample properties of an estimator, compare competing methods in simulation, or reproduce an experiment from a paper.