In the setup.R file are all the code that estimate parameters and cross-validates their quality. It is mostly just a bunch of functions that call better packages in a uniform way.
All the estimation examples are in separate RMarkdown files. Whenever called they produce html files that are relatively easy to look at and store all their estimations in the errors,contained and intervals folders.
The paper itself is the nofreelunch.Rmd file, and it reads from those folders to write the tables. That's it.