World Series time! Enjoy Premium-level access to most features through the end of the Series!
October 3, 2012
The Platoon Advantage
What We Mean By Worthless
I’ve spent a lot of time recently, much more than I care to admit, discussing the end-of-season awards with casual fans. I don’t know why, but it’s something I do. I think that baseball fans becoming generally more informed is a good thing for all concerned. I have no illusions that my ranting and raving actually accomplishes that, but I keep chasing that carrot (often in blatantly unproductive ways). Anyway, something good came out of it, because I’ve had an idea about how to improve this sort of discourse, from our sabermetrically-inclined end. It’s not an entirely new idea—Bill James, for one, has talked about things like this quite a bit over the years—but it’s one I think it’s worth calling back to mind now and then, and I hope I might have a slightly different way of framing it.
You know what the big debate going on right now is—and, no worries, this post won’t actually touch on that whole thing at all—and you can imagine that certain issues are coming up a lot. Things like the value of RBI and batting average. Elsewhere—in discussing David Price vs. Justin Verlander for the Cy Young Award, for instance—pitcher wins come up quite a bit.
The most common words statheads use when those types of statistics come up all end in “less”: meaningless, useless, worthless. I’ve used them. It’s easy. And in a way, I think it’s totally right.
It’s also terribly confusing to the uninitiated, though, and seems to create an extra barrier that really doesn’t need to be there. The natural response to that sort of language, of course, is defensive and incredulous: “You’re telling me a guy who drove in 120 runs didn’t have a great season? Scoring runs is the whole point!” “When’s the last time a bad pitcher won 20 games?” And that’s understandable. It’s not as though these stats don’t tell you anything; if you see that a hitter has 100 RBI, odds are he’s probably pretty good. The same is true for a pitcher with 17 or 18 wins. Besides, no one wants to hear that the things they’ve grown up paying attention to and caring about are “meaningless.” The average fan, hearing that language, shuts down. Wants to fight back. I would, too, probably.
Yet, in a very real sense, those numbers (just to take the two most obvious examples) are totally meaningless. A VCR in good condition will still allow you to watch a tape; a floppy disk will presumably still keep copies of your Word documents or other small files (if you can find a device to read it). These things technically have a use, but are more or less worthless now. We’ve got other things that do their jobs better, and do more besides.
It’s the same way with RBI and wins (and the like); these are “meaningless” and/or “worthless” only because we have other tools that do their jobs better, and more. A hundred-plus RBI might tell you that the hitter was probably good that year (not always), but also gives you a bunch of other information—opportunities with runners on, most obviously—and it’s impossible to separate the two. It’s like static or tracking problems on a VHS tape (to stretch that analogy to its very limits). We have many other ways of tracking things hitters do that lead to producing runs—ISO, TAv, and so forth—that also keep track of the good things, but that weed most of that other junk out. The same is true with wins; a lot of wins probably means a good pitcher, but also probably means a good team (or at least a good offense). Other stats do that better. None of them is perfect, of course, but better.
So it’s not that RBI and wins are literally “useless,” in the sense of having no actual utility. It’s just that, in a world with so many tools that have the same use but are much better at providing it, the old stats are of no practical use to us. Citing those stats in the context of a value discussion has no meaning, as long as we have more advanced and nuanced information available.
So while it can be true in some ways to call them “meaningless,” nobody wants to hear that, and the better and less confrontational point is that they’re imprecise; 120 RBI, without more, tell you a hitter likely had a very strong season, but it won’t tell you how strong, and it certainly won’t tell you whether it was better or worse than another player with 100, or 140, or 75 RBI. A pitcher who has gone 21-5 has probably had an excellent season, but there’s no way, without much more, to reliably compare him to another pitcher who went 16-10 or 14-13 (or in one case, 8-16). Our newer surrogates can tell us not only that a player was good or bad, but can go a much longer way toward telling us that Great Player A was even greater, or less great, than Great Player B. Wins and RBI can be interesting, but really can’t do that at all.
It’s not discarding useless stats for useful ones, it’s putting aside the somewhat telling but imprecise old stats for the much more reliable and accurate new ones. It’s about a certain (imperfect, but greater) degree of precision, not “usefulness” or “meaning.” Sweeping away the noise, getting right down to the good stuff. It’s that simple!
Is phrasing things this way likely to convert the masses? Of course not. There are legions out there who will immediately discard anything they didn’t see on the back of a baseball card when they were 10 years old, no matter how it’s presented. But it’s a more palatable, more easily relatable (and probably a bit more accurate) way of framing the way we tend to look at things, and might get more people listening eventually, especially as they’re increasingly inundated with slightly more advanced statistics in mainstream articles and TV broadcasts.
I believe that a world with more knowledgeable baseball fans is a good thing, and that thoughtful discussions are better than screaming past each other. It might help quite a bit on both fronts if we started talking about sabermetrics as an attempt not merely to stamp out but to improve upon the precision of those old measuring sticks—which it is—and stopped treating the old stats as entirely useless or meaningless.
Even if they kind of are that, too.