Thought some would be interested in an academic paper I found about the usefulness of MILB stats in predicting MLB success.
Here is the abstract; the rest can be found at the link.
We consider the problem of projecting future success of Minor League baseball players at each level of the farm system. Using tree based methods, in particular random forests, we consider which statistics are most correlated with Major League success, how Major League teams use these statistics dif- ferently in handling prospects, and how prior belief in a players ability, mea- sured through draft position, is used throughout a players Minor League ca- reer. We find that roughly the 18th round prospect corresponds to being draft neutral for a team, at which point teams essentially make decisions based strictly on performance. We use for our data all position players drafted be- tween 1999 and 2002.