I just want to add soemthing to Bill said, and that is that high school pitchers are always going to have greater "upside" than college pitchers because less is known about them. Say the true ability of a pitcher is some random variable X. Now know one knows what X really is, but using different measures, we can estimate X . Call this Xe. Now since Xe is an estimate, there is a band around this value that we can similarily expect the player to fall into. The size of the band depends upon the amount of data available. Since college players have more data available, the size of this band will be smaller than that for high school players. So if we take a college pitcher and high school pitcher who have the same Xe, the high school player will be said to have more "upside" because his band is looser. So all the upside really is is a reflection of our lack f knowledge.