almost immediately had a response for me back at his own site. It's excellent and thoughtful, again, and there's an interesting discussion that goes on in the comments, under which I put my own reaction to it. Go there and read it yourself, but in a nutshell, Mark wonders whether there are things that we still need to consider that are more difficult or perhaps impossible to quantify -- leadership, intimidation, distracting a pitcher with the threat to steal, and so on. I think those things, or many of them, absolutely do exist and have an effect on the game, but that while sabermetricians don't have any way of measuring those things, it's important to note that batting average, RBI and fielding percentage don't measure these things either. So I don't see how these things bolster the anti-saber crowd's arguments...except that Joe Morgan and Steve Phillips keep telling us that they do.
-----------------------------------------------------
In the comments to yesterday's post, Ron from Baseball Over Here pointed me toward something he wrote about six months ago in defense of the RBI, and again, I think this is excellent and definitely worth a read.
Ron starts by showing that the list of career leaders in RBI is populated almost exclusively by great players. The conclusion seems to be (correct me if I'm wrong, Ron) that RBI must be measuring something useful, if only great players are getting a lot of them.
A potential problem with that is that you need to pretty much be a great player to be among the all-time leaders in any stat. The top 14 in At-Bats, which doesn't measure anything but one's propensity for being penciled into the lineup and not walking or sacrificing, are all in or surely headed to the Hall [edit: or are Pete Rose]; you can draw your own conclusions about Raffy at #15, but then the next ten after him are already in too (though some you could argue about -- Maranville, Aparicio). The top three batters in career strikeouts are Reggie, Slammin' Sammy and Thome; the top six pitchers in career losses are all in the Hall. So by itself, I'm not sure that that line of thinking gets us very far.
I'll definitely accept the general premise, though. You can even just limit it to a single season. There are definitely a few exceptions (Jose Guillen and Emil Brown were mentioned in the comments to Ron's piece), but really, if you've got 100 or even 80 or 90 RBI, the odds are very good that you were an excellent hitter (for power, at least) within that season. The problem is that that extremely high-level thing is pretty much all it does; if dude A has 100 RBI and dude B has 120, we have no better idea who was the better player with that info than we had without it.
Ron is all over that. He recognizes that RBI totals are a poor way to decide the MVP race, for instance. But he concludes that RBI is a good stat anyway, because they tell us something important--how many times the guy did something that brought a run home. He has a lot of interesting analysis about how many different ways a run scores, and basically shows that, you know, RBI are usually pretty necessary to scoring runs most of the time.
But I'm not sure I understand why that makes the RBI stat itself important. We already know what a player can do that tends to lead to a run scoring (in rapidly descending order: (1) get on base; (2) hit with power; (3) run the bases well). We can track how well he does these things and get a pretty good sense of how good he is, all else being equal, at producing runs. If we already know these things, what does it add to consider RBI themselves, knowing as we do that so much of what we're really measuring is the opportunities that that player's teammates created for him?
I'm really asking. From where I'm sitting, it looks like RBI themselves are superfluous when you already have all the other, non-context-dependent stats that make good RBI guys good RBI guys. I'm certainly open to discussion and new ideas on this....I'm just not seeing it right now.
Very well put.
ReplyDeleteIt might be a while before we see stats such as RBI go to the wayside, and for the sake of the kids watching the game, I hope they don't (RBI's are much easier to understand than, let's say wOBA; you might be able to tell them that it's good, but they likely won't understand). Anyway, the authority in baseball still maintains that those stats are good. Journalists in the BBWAA are still favoring the older stats. I'd probably say that Selig does, too. Arbiters in arbitration cases still focus on the older stats. And for the most part, fans are still use only the older stats. We talk amongst ourselves about these stats, but most regular fans still like batting average and RBI.
For these to be truly accepted, there probably needs to be a new commissioner and about 15 years for the "traditionalists" to literally die out and relinquish their positions. And then, the population at large is going to have to stop hating math so much.
Yeah, I'm with you, Mark. I don't think stats like RBI should go away. I like seeing them on the scoreboard and whatever for the reason I gave in my initial post on this -- they give put the player's performance in some context. And the point about the kids is a good one, too.
ReplyDeleteThat's why discussions like this are good, not the fighting and name calling.
ReplyDeleteI won't sit here and say the RBI is a great stat. It's not, and it's use is overblown. No offense to Morneau.
But I think it provides an additional look at what is happening on the field, and shouldn't be discounted.
As far the great player/lots of RBIs comment, I might have overstated that. I drew a conclusion that wasn't necessarily there.
But if you have Morneau batting 4th for 162 games, he's likely to have 162 RBI's.
If you put Cuddyer 4th for 162 games, with the rest of the lineup exactly the same, he's not going to have 162 RBI's. Or anything close.
That tells me something, even it doesn't mean anything to anyone else.
And I guess that's why I like RBI's.
Good points, Ron. But look at it the other way--if the better hitter (or better "RBI guy," or whatever) is hitting in the position with fewer RBI chances, then looking at their RBI can really distort things.
ReplyDeleteConsider the 2008 Yankees. Johnny Damon batted leadoff almost exclusively, but had an excellent year with the bat and did even better (1.013 OPS) with runners in scoring position, so he ended up with 71 RBI. Jason Giambi, meanwhile, batted fourth, fifth, or sixth all year, and hit about as well as Damon overall, but fell flat with RISP (.692), yet still managed 96 RBI.
Giambi did drive himself in 15 more times than Damon did, but his RBI total benefitted hugely from battong order position: he had about 160 more runners on base during his plate appearances than Damon did.
If Giambi and Damon were flipped in the order, so that Damon batted with as many runners on base as real-life Giambi did and vice-versa, and they drove in those runners at the same percentage (a nearly league-leading 19.355% for Damon, and a mediocre 14.612% for Giambi), we'd expect Damon to have 101 RBI and Giambi to have just 73.
That's kind of an extreme case, but I think it illustrates the point. Yes, a great hitter hitting 4th is going to have more RBI than an average hitter hitting 4th. But a great hitter hitting 1st or 8th will very likely not have more RBI than an average hitter hitting 4th...