Do you not think he would test to see if keeping previous years data would improve the model overall? That he just did it for kicks?
BYU is an outlier. He or any other of those guys will tell you they're a good candidate to outperform the projection, but they're trying to be objective, they don't mess with specific teams.