Yep, which is why so many of these "projection models" are laughable. The good players regress. The bad players progress. Everyone ends up in the middle. Used in the aggregate, projection models might have some use. But on an individual basis, they're laughably bad. Mike Trout has two full seasons in MLB. He has posted an OPS of .963 and .988. He is going into his age 22 (!) season. Davenport's projection? An .871 OPS. Miguel Cabrera is the best hitter in the league. In the past four seasons, he has posted OPSes of 1.042, 1.033, .999, 1.078. He, like Joe Mauer, is entering his age 31 season. Davenport's projection? An .895 OPS. Bryce Harper is a phenom, hitting MLB at the age of 19. He has posted an OPS of .817 and .854 in his two seasons. He is entering his age 21 (!!) season. Davenport's projection? .816. Another regression candidate. Used individually, these projection models are more than inaccurate. They're broken and nonsensical.