Without getting too in-depth on my method, I use a combination of shooting rates and per-minute floor stats. When two players are being compared, I normalize (z-score) the pertinent stats and then add them up in what can be called a z-sum. The lower the z-sum, the more alike two players are. By limiting the pool of players to those of the same age, I can compare players at the same part of their career curves. That is, there is little use comparing a veteran at the tail end of his career to a neophyte still learning the ropes.
Using this process, I’m able to come up with a list of players who have similar statistical profiles at the same age. Hence by looking at the top 10 most compatible players, one could get a good idea at the ceiling and cellar for a given player at a given age.