This will (hopefully) be the least-analytical thing I post here for quite a while. I could, of course, build a projection system and make my picks that way, except for the fact that I’ve sworn never to build a baseball projection system, for the following reasons:
1. Assembling a projection system would take a significant amount of time and effort. And since there are already systems out there that are significantly more sophisticated than anything I would build, that time and effort would be a prodigious waste, even moreso than the rest of my unpaid work already is.
2. I would enjoy the sport less as a result. Whenever I make predictions, I find myself rooting for them to come true instead of hoping for entertaining or surprising things to happen. The primary rule of baseball projection is “this guy will keep doing roughly what he has done before.” But one of the best parts of baseball is the fact that players don't always follow that rule. When the next Jose Bautista or Chris Davis hits 50 homers or something, I don’t want to be the guy who insists on predicting that this year’s breakout star is bound to regress to the mean; I'd rather hang on and enjoy the experience while it lasts.
So my picks are generally made with as little study as I can manage, which in turn allows me to slide by with minimal emotional investment in them. It also ensures that they will turn out to be thoroughly incorrect – but then, everyone else’s usually are as well.