It might be helpful, I think, to draw a distinction. A large part of sabermetrics is what might be called "the language of decision making." These are the numbers and acronyms that help GMs and coaches evaluate players and execute strategies. It doesn't much matter how cumbersome it is, because the talk is deliberate and behind-the-scenes.
Another and quite different way to use advanced stats is to employ them as a "language of description." When we make fun of old fashioned broadcasters, we generally malign their descriptive powers. For instance, we think that they should use FIP instead of Wins and ERA to describe a pitcher's goodness, or wOBA instead of RBIs and Home Runs to describe a hitter's.
It's clear enough to me that the critics of sabermetrics confuse these two types of language. They are so irate about perceived arrogance regarding the latter that they totally impugn the former. In actuality, I doubt that anyone would mind the "language of decision making" if it was relegated to front office boardrooms. It would just be another arcane industry vocabulary.
What more interests me is the question of whether we confuse the two as well. I may overstate the case a bit, but in so doing I hope that there might be some valuable self-reflection. This is the postulate I think we might debate:
The sabermetric community is so convinced of their stats' analytic superiority that they incorrectly equate this with descriptive superiority. As this represents a misunderstanding of language, the sabermetric community is in part responsible for the hostility it receives.
I'll begin with the glaring example. It's often been asked how anyone in their right mind could care about a pitcher's wins. It's so obviously a flawed analytical tool that it must be apparent to everyone involved with the game. I think that's true; it is apparent. Here's Jim Bouton in Ball Four:
About a week ago Segui won two games. He pitched about two innings and gave up two runs and then about four innings and gave up two more runs. He was pitching lousy, but he was in there when our team was scoring runs, so he got credit for two wins. And so the reporters started coming around. "Gee, Diego, you're starting to win ballgames. Tell us what you're doing different.
He wasn't doing anything different, except maybe pitching worse.
It's obviously impossible for any pitcher to not understand this concept. After all, they are the guys getting judged unfairly. But elsewhere in this great book, Bouton uses wins constantly as a measure of success. Not only does he talk of the thrill of getting a "big W" -- even as a reliever -- he speaks of "20-game winners" and guys going "0-8" before getting cut, etc. etc., to beat the band, and he means those as judgments of quality. In fact, in the Historical Baseball Abstract Bill James does the same thing. Are they idiots? No. Did they have access to at least ERA? Yes. Then what's going on?
Before I spill my beans, one more example to chew on. Here is a quote from Roger Angell's classic, The Summer Game, and pay attention to the descriptive tools he uses:
The Mets are short on heroes. Their gamecock, their small nova, is Ron Hunt, the second baseman. Now in his second season, Hunt appears to be one of those rare ballplayers who improve from year to year. Four years ago, he batted .191 in the low minors. Last year, as a major-league rookie, he hit a solid .272, and so far in this young season a stream of modest singles and doubles has kept him constantly above the .300 mark. Originally a third baseman, he now has mastered the pivot at second and he fields with assurance, if not brilliance. His new partnership with Rou McMillan, the veteran shortstop just purchased from the Braves, cements the Mets' infield. McMillan has slowed down a step or two and is having trouble at the plate, but he is a tough, tobacco-chewing, old-time pro and a boon to this young team.
I don't think that many of us would deny that Angell is a wonderful writer. Even the most ultra-saber purist isn't going to deny him flashes like "their small nova." Of course not. The question is, if it were written today, would it be improved by scratching out things like batting average, small sample eye-ball assessments, and the implied moral weight of the "tough, tobacco-chewing" characterization? Would it get better by becoming more precise and accurate, and by using our new analytical language of accuracy? Maybe, but maybe not.
Here, finally, is what I'm getting at. There is nothing that says that descriptive language is made better by precision and accuracy. Every writer knows this. When one describes a room, he does not describe it; he creates a new language-room that gives the feeling of the room he has in his head. To do anything else is to get bogged down in details and create something in fact more distant from the actual room. A writer also knows that he should use concrete, instantly communicable language. One should call them "curtains," not "window treatments," even if in this case they're not technically curtains. The simpler word is infinitely better.
When we watch baseball, only one part of our mind is thinking like a GM. The other part is doing, in effect, what Angell is doing: re-processing the game into language that strikes a chord with our psyches and our buddies. This is why Bouton and James and are interested in "Wins." Though it's a rotten analytical tool, "Wins" has a concrete reality for these men; they are an almost tanglible presence and thus the logical choice for description. RBI, of course, is the same way: rotten in one way, but far more instantly communicable and "real feeling" that wOBA. At least to 90-plus percent of fans. Same with batting average and ERA. These things have a literal correspondence that many metrics -- with their weights and park corrections -- lack. One could argue they're more concrete.
This gets me to my final point. My argument is not that the language of the game should or will remain static. As "Wins" become less tangible, they'll drift out the language, and I agree that this is for the better. But it is no sure thing that the sabermetric "language of decision making" is the right tool for describing the game of baseball, over the radio or in the newspapers or even at the ballpark to each other. Maybe it is, but we have to prove it. You can't just shout at a Victorian novelist and tell him that "naturalism" is the way to go. And we can't just shout at beat reporters.
Namely, we should think about a sabermetric "language of description" in real, journalistic terms. It should poach the best, most communicable stuff -- the stuff that can get into the blood -- and leave the rest to the experts. Some things are just too abstract to play, other things could use better names.
Most importantly, we should keep writing and writing. We should write game recaps that are narratively familiar to the average reader but employ our new language. We should write Angell-like essays that tell stories as they analyze, and analyze as a means to tell stories. I'm not saying that there isn't also a place for esoteric blogging and razor-sharp recaps. There is. But we should do more to integrate our language into the extant, storied baseball vocabulary.
If we insist on sabermetric language's de facto descriptive superiority, we misunderstand language. Then we're the ones who are wrong.