Power Ranking Methodologies
A couple weeks ago, my friend John Gibson asked me to do a monthly NPB power rankings segment for his Japan Baseball Weekly podcast. I’ve been on the ‘cast a couple times and it’s always a lot of fun, so it was an easy decision to accept, despite my lack of recent writing activity.
The other thing that made this an easy decision is the opportunity to do something different with a concept like power rankings. I have to admit I’m not much of a fan of power rankings; I tend to think of them as subjective and lacking real analysis. So if I’m going to do this, I’m going to try to put my own spin on it. Even if turns out to be analytically shallow, it should at least be fun.
The question is how to rank the teams. For me, there is a spectrum of possible answers, ranging from “objective” to “subjective”. On the “objective” side of the spectrum, there’s data. Not just wins and losses and runs scored and allowed, but ideally atomic, granular data on each play, that should be more predictive of how a team is actually performing. On the other side, there’s eyeballs and intuition. “Sure, Hanshin is winning, but they don’t feel like a first place team.” Power is in the eye of the beholder.
In my first attempt I’ve come down somewhere in the middle. At the moment I don’t have the type of play data I want for an entirely data-driven approach, so I used data points that I thought were most indicative of each team’s ability to compete along with my own instincts. The top and bottom teams were pretty obvious; the middle tier not so much. The rankings will be in this week’s JPW podcast, so we’ll see how I did then.