If you play fantasy football and also know nothing about football, all you can rely on are ESPN's projections. Obviously, the world needs to know how accurate these projections are. Otherwise, the bragging rights of millions of people are in jeopardy.
So here is a fully Bayesian hierarchical linear model which tests their efficacy. It uses the projections to predict the point spread. Also, a logistic regression predicts winning the fantasy football game. Both of these analyses indicate ESPN's projections are terrible.
A write-up is attached, in case you are unfamiliar with fantasy football, hierarchical linear modeling, or both.
The raw data is also available, along with the code for the analysis. If you wish to replicate my analysis, you will need a copy of WinBUGS, which you can find here: https://www.mrc-bsu.cam.ac.uk/software/bugs/