This is a silly example of ex post evaluation of an ex post model. You need to have a holdout sample. It's made more silly by the fact that they use the unive

This is a silly example of ex post evaluation of an ex post model. You need to have a holdout sample. It's made more silly by the fact that they use the universe of Waldo books, so their "strategy" is essentially just "here's approximately the answer".