RBI?
The following is a lengthy discussion/argument that Bryan and I had on April 29 about RBI. I believe that Bryan has somewhat changed his views since, which I'm sure had very little to do with this discussion.

Marc:
Tue, 29 Apr 2003 08:42:32:
>> why would peter gammons suggest that the sox are dying to trade jillenbrand
>> to the cubs when the guy has 8 million rbi? other than the fact that he's a
>> dopey no-information-having buffoon?


Ryan, 12:03:

> Because RBI are a pretty meaningless stat, and we're only a month into the
> season, so the stats don't mean that much anyway. Mueller is at least as
> good a player as Hillenbrand, so they don't really need him. The Cubs are
> apparently unhappy with Bellhorn, so maybe they will finally be willing to
> give up Juan Cruz.
>


Bryan:

Mueller cannot hit like Hillenbrand, never has. Remember also Hillenbrand is
still developing, Mueller's pretty much set. Gammons is doing the Gammons
thing, which is spouting all sorts of rumors that will not happen. Having
watched probably a third of the Sox games this year, I can say Hillenbrand's
RBI are not meaningless; he has been a clutch hitter. Granted they could be
meaningless, but in this case they're not. And in most cases, they're not.
Whether I'm going to start some sort of holy war in which I won't
participate I don't know, but why everyone thinks after 100 years there are
better simple baseball stats than average, homers, and RBI is beyond me. If
the key word there is simple, so be it.


Ryan, 12:26:

> I don't want to get into a lengthy Hillenbrand-Mueller debate. I will say
> that Hillenbrand is probably a slightly better hitter, but it's very close.
> Mueller is supposed to be a better fielder. I think they're fairly
> comparable.
>
> I will however say that RBI are indeed a pretty worthless stat. I guess it's
> "simple", but what does that really mean? (Wouldn't it be simpler to have
> one stat to tell how good a player is instead of 3?)
>
> RBIs largely reflect the opportunities a player has. If there are frequently
> players in scoring position when you come up, you'll get a lot of RBI.
> Hillenbrand may have a few clutch hits this year, but anyone who has ever
> done a study on it has found that "clutch" hitters don't exist. It's just a
> matter of luck. Players aren't able to consistently hit much better or worse
> with runners in scoring position.


Bryan, 12:36:
"matter of luck. Players aren't able to consistently hit much better or worse
with runners in scoring position."

I'm not so sure there is anyone who plays baseball who would corroborate
this. Sticking with the Red Sox because that's what I do best, if you're
talking about manufacturing runs, here are two innings from Sunday's game
that stick out: early in the game, Bill Mueller gets a hustle double. Nomar
then flies out to right, easily advancing him. Against a drawn-in infield,
Manny gets a soft-hit ground ball through the left side, scoring the run.
That's two situational hits, with only one RBI to show for it. In the 10th
inning, there's one out and a man on third after Hillenbrand gets a double
and the pinch runner is bunted over. Jason Varitek strikes out swinging.
Now, it's one case, but Manny Ramirez and Nomar Garciaparra are considered
two of the best players in baseball, and Varitek's not. Will different
things happen under the same circumstances? Of course they will. But will
this happen again? Probably. And that's why you can't call RBIs meaningless,
because better hitters do the right thing more often when it counts, which
is precisely what makes them better hitters.


Ryan, 1:47 :
> One, I think people who play baseball aren't always the best people to make
> these judgments. Sometimes it helps to be able to look at something from
> some distance. Obivously, a player has to believe that he can turn it up a
> notch in a pressure situation. It doesn't mean he can.
>
> And those good things that good players do show up in other stats. Manny
> Ramirez and Nomar have great OPSs, too.
> Sure, in general, a guy who gets 120 RBI is better than a guy who gets 60.
> But I think it's much better to use stats that better reflect a players
> overall skills like OPS than to use RBI.


Bryan, 1:57:
Well, I can say that I wasn't intentionally contrasting it to OPS, but I
guess that's the natural comparison. I personally feel like you're right,
big RBI numbers show up in other categories like OPS, but I think the
breakdown is pretty much the same. That is, there are high guys like Barry
Bonds and Sammy Sosa, and low guys like Rey Ordonez and Doug Mirabelli in
both categories, but most people are in between. Though it's an admittedly
tenuous argument, I'd also say that OPS is affected by team performance in
the sense that if your team sucks, it's harder to hit/walk and therefore
your OPS numbers are skewed, as RBIs clearly are. But I don't think OPS says
everything, because baseball is too multi-faceted, and that's why I like the
old three better than just one. And let me clarify something: I don't mean
there are necessarily clutch hitters, but if batters averages' don't
"increase" in clutch situations, they're still higher than each other's, and
Manny Ramirez' is still better than Jason Varitek's. Over the course of a
long season that translates to RBIs in these small ball situations, while
Ramirez' homers take care of the rest.


Ryan, 5:09:
"I don't mean there are necessarily clutch hitters, but if batters averages' don't "increase" in clutch situations, they're still higher than each other's, and Manny Ramirez' is still better than Jason Varitek's. Over the course of a long season that translates to RBIs in these small ball situations, while Ramirez' homers take care of the rest."

If you're willing to accept that there aren't clutch hitters and that RBIs just reflect that people get more hits and homeruns, then why not cut "the big 3" stats down to 2, batting average and HRs.

And if you do that, you'd have to recognize that there are somethings that aren't recognized (like walks, doubles, and triples). So, if you wanted to add those things, you'd get something like OPS.

If your critcism of OPS is that "baseball is too multi-faceted", I will agree that it doesn't take every possible element of the game into account. I don't think we will ever come up with a stat that will evaluate everyone with 100% accuracy. But I think it's much better than the "big 3" because as I said, it essentially takes into account batting average and homeruns, as well as other important facets of the game.



Bryan, 5:27:
Well, that's if they don't increase in clutch situations. Though I've read
plenty about the myth of the clutch hitter, good hitters' averages are often
higher with runners in scoring position, which would mean they do increase
in clutch situations, whether it's because the pitcher gets nervous, he
knows a fastball is coming, etc. So I'm not really willing to accept there
aren't clutch hitters, and one example I'll cite is Gary Sheffield, who has
something like four home runs in 1-0 games over the past two years. It's
pretty absurd, and it's not like I'm the world's biggest Gary Sheffield fan.
It usually comes down to being able to hit the fastball. And why not just
average and homers? Well, Mike Piazza has three home runs and four RBIs this
year, which means he's not hitting with runners on. And OPS can't tell you
that, and the combination of the three can. My argument is since you can't
come up with something that is 100% accurate, and since there is no real
complement to OPS that makes it easy to break down a player, the big three
stats are the easiest way to find out what you typically want to know about
a player, whereas OPS is just a general gauge, and therefore is not likely
to supplant the big three. Here's one situation where I think OPS could be
useful, but I would still stick with RBIs: playoffs, late innings, let's say
Varitek or a pinch hitter. Varitek's got low average and homer numbers and a
decent OPS, David Ortiz, on the bench has a higher OPS but is a part-time
player. Now, instinct would be to go with the guy with the higher OPS. But
what if Varitek has a slew of RBIs? That would tell me he's hitting well
with runners in scoring position, basically hitting the fastball. The irony
of the whole thing is, I think OPS' natural complement is RBIs, because on
base plus slugging basically takes care of the first two of the big three,
but not the third. However, everyone's going to want to know how many homers
some guy's hit - and that's why I don't see OPS taking off in the capacity
in which it could.



Ryan, 5:46:
I have a very simple solution for what to use to complement OPS:

OPS with runners in scoring position.

I think it reflects clutch performance much better than RBI because like I've said RBI are very much based on the situations you get. When Jeff Kent won the MVP, a big part of that was that he had 125 RBI to Bonds' 106, even though Bonds had a better OPS with runners in scoring position (it was fairly close). Obviously, a big reason that Kent had all those RBI was because Bonds was batting in front of him and always on base.

Also, I'm not arguing that we should never mention RBI or HR ever again. My problem is when people throw out lines like "we should trade for him, he hit 88 RBI last year" or "he sucks, he's hitting .250". That's all well and good, but without knowing the rest of his numbers, that doesn't tell me much. A player might be very good and still only hit .250 (Adam Dunn). A player might hit 88 RBI and not be (Raul Mondesi last year).
If you tell me a player has a .900 OPS, I feel fully confident saying he is a very good offensive player.



Bryan, 6:07:
> OPS with runners in scoring position.
With a man on second, a walk would net you a 1.000 OPS. But you didn't
score. I mean, I watched the Angels and I know the value of getting on base,
but that's an instance where an RBI is more important than getting on base
(however marginally, but the 1.000 OPS makes whoever look like a
world-beater).

I thought of Mondesi while thinking of how RBIs alone are flawed. But on the
opposite tack, Ichiro's OPS two years ago was .838, not exactly
earth-shattering, but he was one of the best five players in the league
(and, of course, the MVP). Look at it the other way, too: if a player has a
.900 OPS, those stats are represented, for the most part, in the three
others, unless they have a ton of walks. Walks are a natural sticking point,
but here's the thing: only good hitters usually walk a lot, and an example
is that OPS shows Bonds' ridiculous value from two years ago, when he hit 73
homers but wasn't among the league leaders in RBI. However, I would argue
that the 73 homers from two years ago and the .370 average from last year
illustrate as effectively as OPS how good he was, and RBI is a measure of
what they allowed him to do. Basically, I see the three stats as average =
consistency, homers = power, and RBI = clutch, and I think it's dangerous to
label any one of these "meaningless." I have more to say but I'll say it
later.



Ryan, 7:48:
As for OPS with runners in scoring position, I'm sort of inclined to agree with you on the walks. THat can be easily solved by looking at Avg. or Slg. with runners in scoring position. My basic argument is that using RBI to measure clutch hitting is inferior to actually looking at what a player did in clutch situations.

Also, since most people who have studied the matter agree that there is not thing as "clutch" hitting, I don't think it's important to measure at all when predicting future performance.



Bryan, 4/30, 5:06 pm:
I forgot this was waiting for me. If you're going to defer to average with
runners in scoring position, then what's the case for OPS? That some guy
walks a lot? Great, but what happens when he gets to second? The next guy is
up, and you don't look at his OPS, you look at his average. I don't think -
and this is where we disagree - you can evaluate players completely
independently of their teams, because it's always related. While I like OPS
as a rate, as I've said before, the one stat it doesn't account for is
cumulative clutch hitting. Note that I've made a distinction between clutch
hitting and cumulative clutch hitting. "People" who have dismissed the myth
of the clutch hitter can talk all they want, but that doesn't account for
players having high averages with runners in scoring position, nor does it
account for the fact that hitting at any point is clutch. That is, OPS is
good for measuring the ability to get somewhere on the basepaths, while the
big three stats measure how often someone is able achieve the best possible
result while at the plate, i.e., a hit, home run, or "clutch" hit. And if
there are no "clutch" hitters, why would anyone's OPS with RISP be any
higher than their normal OPS? Because the situations are different. When the
situations are different the variables change, be it for the cleanup hitter
who gets a lot of chances or the eighth hitter who doesn't. But my argument
is that person is batting fourth for a reason, the reason being to drive in
runs. And that's why RBIs are important, especially for comparing across
teams.

OPS is a valuable gauge stat. And there is a compelling case for it to
replace average and homers, even though it never will in the public eye. But
OPS with RISP is no RBIs, and I would argue has little value at all.




Ryan, 5:18:
You still haven't addressed my biggest problem with RBI, that they are too influenced by how often you're lucky enough to come up with runners in scoring position.



Bryan, 5:26:
Argument: it's not luck. You're hitting fourth for a reason, that is, your
manager wants you to drive in runs. If you do it, you stay there, you get
more RBIs. If you don't, you get dropped, your RBIs go down. That's a very
simplistic answer, but it's also very simplistic to say someone is "lucky"
to get a chance to hit with RISP when you can easily bat 500 times a year.
The current baseball economy skews all stats, not just RBI, so I'm not quite
ready to throw them out.



Ryan, 5:56:
That's bullshit. Maybe the manager is just an idiot for batting you fourth. Regardless, your high RBI total still is reflecting the fact that you bat fourth (presumably because you're a power hitter which is reflected in your HR), not that you are inherently a more clutch hitter than the guy hitting first. And if you bat fourth on a team with 3 guys in front of you with OBPs over .400, you're going to get a lot more RBI opportunities than if you're batting fourth on a team that has 3 guys with OBPs around .300 in front of you.
I don't think any study has shown that other stats, like OPS, avg., and HRs, are affected much, by who is batting behind you, so I don't think it matters much who else is in your lineup for those stats, but it matters tremendously for RBI.

My other main argument against RBI which I'm probably not going to convince you of, is that clutch hitting doesn't exist. Yes, I know, you can isolate a few players and say they have higher averages with RISP, but you can also isolate a few players who hit much better on Wednesdays, that's just random fluctuations. If I flip enough coins 10 times each, one of them will get heads 8 or 9 times. That doesn't mean it's a clutch coin. There hasn't been any study that has shown that "clutch hitting" is a consistent ability.



Bryan, 6:17
Sure, but OPS and average are rates. You can determine them after one at-bat
or after 500. It's not cumulative. It's not something you get. Whether it's
fair or not to someone on the Tigers that Raul Mondesi had 88 RBI last year
is irrelevant. No matter how you slice it, that's a lot of RBI. Though a
walk sometimes ends up being as good as a hit, it never starts that way,
because on a hit there's a possibility for an error, possibility to stretch
it, whatever. That's why hits are relevant. Homers are relevant because you
eliminate all other probabilities. And RBIs are relevant because they are
the goal of every at-bat. If you're upset that some players get shortchanged
because they play for bad teams, I would argue strongly that affects every
stat and has for as long as baseball or any sport has existed. You take the
top 10 RBI guys every year and they will be great players. Same thing for
OPS. If it was meaningless, as in literally had no meaning, that wouldn't be
the case.

And you won't convince me that a clutch hitter doesn't exist, and the
Wednesdays argument is ludicrous. Within the context of a game, there are
variables that change the circumstances. And most players' averages' go up
with RISP, some by a lot, and that's because the hitter has an advantage
when the pitcher needs to make certain pitches in certain situations and the
hitter knows what those pitches are.



Ryan, 6:52:
I think I'm finally willing to give up on this debate because I don't think either one of us is goint to convince the other of anything, but I have 2 final points in response to that email.

1. If you want a cumulative stat, you can use something like Runs Created.

2. The problem with RBI isn't just that players on bad teams have low #s, its also that certain lineup spots have low #s. When Ichiro won the MVP, a big part of the reason was that he hit well in the clutch that year, hitting around .450. But he only had 69 RBI. If RBI were a good measure of "clutch" hitting, then how could one of the most clutch hitters in the league have such a low total?



Bryan, 10:26
My only answers to these questions:

1) We never got into runs created, but to me, it sounds similar enough to
RBI and complicated enough that RBI still be more useful.

2) I'd say his batting average speaks for itself.