Jiaying Gu and Roger Koenker, "Empirical Bayesball Remixed: Empirical Bayes Methods for Longitudinal Data", Journal of Applied Econometrics, Vol. 32, No. 3, 2017, pp. 575-599. There are four files (in addition to this README) for this paper: bball.csv -- a comma-separated, ascii file with the data used in the paper. It is zipped in gk-files.zip. DataConstruction.tar.gz -- a gzipped tar archive with R files to construct the data from the original source: ESPN Website: bayesballR3.R.tar.gz -- a gzipped tar archive with R files for the analysis AnalysisFiles.txt -- a roadmap of what R files are used for the tables and figures of the paper. This file also mentions several files saved in the R .Rda format that are generated as intermediate products of the analysis and can be regenerated with one or more of the .R files included in the bayesballR3.R.tar.gz file. All of the analysis relies heavily on the REBayes package for R, which in turn relies on the Mosek optimization engine and the Rmosek interface package. This file is also zipped in gk-files.zip. Data in bball.csv have been aggregated into half seasons. Halfseason indicates whether the observation is in the first or second half of the season of a given year. Only players who have more than 10 at bats in any half season are included, and only players who have more than three half seasons are represented. The transformed batting average is arcsin(sqrt((H + 1/4)/(AB + 1/2))). • Name • IdNum • Year • Halfseason • Pitcher • HA transformed batting average; • AB at bats • H hits • BB walks • YOB Year of Birth; • age age of the player • agesq age squared For Rgonauts, the data in bball.csv are available from the R package REBayes with the instruction data(bball). Note that the two ASCII files in gk-files.zip are in DOS format, so Unix/Linux users should use "unzip -a". The ASCII files in DataConstruction.tar.gz and bayesballR3.R.tar.gz are in Unix format.