A while back I posted my Batting Average projections for players. In the comments section, somebody noticed that Neifi Perez was down for a batting average of around .290. Neifi Perez, however, is a terrible hitter and will get nowhere near this number. I said I would look into what was going on. I have recently begun to do so.
When Ron Shandler projects batting averages, he considers two factors very important: contact percentage and walk rate. Contact percentage deals with the concept that the more balls a player puts in play, the more balls will fall for hits. Walk rate is a way of measuring selectivity. From what I've heard, Shandler thinks that the more walks a player takes, the better the pitches are that he ultimately decides to swing at and the better percentage of their balls in play will actually fall for hits.
My old formula did not take selectivity into account, but the new one will. I even believe that I've come up with a better way to determine selectivity than Shandler. Using data from retrosheet.org, I have put together a list for the past few years of the types of pitches each batter sees: swinging strikes, called strikes, fouls, balls, and a few more. I've only recently compiled this list and have not yet tried to analyze it, but I believe that by looking deeper into an at-bat (instead of simply looking at the outcome of it), we will be better able to determine selectivity of a hitter.
I still have to work more with these numbers, but off the top of my head I think this makes sense. The more pitches a batter doesn't swing at (called strikes and balls), the pickier he is about the ones he will swing at. I believe that this will prove to be key in gaining a better understanding of hitting selectivity.
I will let you know when I find some conclusive results.