MaximumScoreEstimator/MSE-R — reverse-engineered prompt

Build me an R package called maxscoreest that solves the pairwise maximum score estimation problem for matching games, following Jeremy Fox’s maximum score estimator terminology.

I want it to let an R user load matched or unmatched market data, turn it into the inequality member and data array format needed for estimation, create parameter bounds, run the score optimization, and then print the estimated arguments, objective value, and market level stats. Include a working synthetic data example like the README shows, plus vignettes or long form examples for matched data, unmatched data, and cube root bootstrap inference.

Please make it feel like a normal installable R package with documentation for the public functions, tests, sample data in the package, and a clear README showing installation from GitHub and a minimal end to end example. Keep the technical math assumptions in the docs, but make the usage instructions practical for someone who already knows the estimator.