Skip to contents

Applies glmnet coefficients to per-stat Bayesian ratings to produce a composite player rating. This is the career-level, predictive version of player value.

Usage

calculate_psr(ratings, coef_df, center = TRUE)

Arguments

ratings

data.table with {stat}_rating columns from estimate_player_stat_ratings.

coef_df

data.frame with columns stat_name and beta. Optionally sd for standardization.

center

Logical. If TRUE (default), subtract league mean so PSR represents value above average.

Value

data.table with identifier columns plus psr_raw and psr.