It has been clear for years that the science of player projection is something of a scam. There is a finite amount of stuff we can know about a player’s performance the next year and a certain amount that is stochastic, random, unknowable. I’ve put the unknowable part at about 25 percent, based on various ways of measuring the accuracy of my expert projections.
This big random component means that the lens of a single season tells us only a little about about a player’s actual abilities. And while we use these small slices to tell us more about the player’s game, as a player ages his game changes. The measures that matter for a 25 year old are different for a 30 year old and different still for a 35 year old. The very smart Tom Tango set out to see how much of the potentially knowable 75 percent he could project using a very raw set of weighted averages building in regressive factors, and writes about it here.