Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
"Joe Carter was the RBI machine of his generation". So said coliver in MED's Hall of Famers thread. Now I don't want to pick on coliver; this type of comment is frequently made about Joe. But is it right? Was Joe Carter a uniquely good RBI man? If so, despite the fashion among analysts to deride RBI at the expense of other metrics, it might be a good reason to revisit our opinions of Touch 'Em All Joe.


I think there are four ways to look at whether Carter was an "RBI machine", any of which might be considered valid, but some more important than others. First, is whether he amassed a lot of raw RBI. Second, and more importantly, is whether he amassed a lot of RBI, given his opportunities at bat. Third, is whether he amassed a lot of RBI, considering his overall offensive production (this is less important, but still an interesting question). Fourth, is whether he was a good hitter with men on base, or whether his RBI total was higher than expected given his other offensive numbers. This I call a "run element" analysis.

1. Raw RBI totals.

First, let me point out that Joe stands 48th on the career RBI list. That's a very good accomplishment. Joe led his league in RBI once, came in the top three five times, and the top ten eight times. So he has some solid opening RBI credentials.

Joe's "era", I would estimate, is roughly 1981-2000, a nice two-decade period, allowing two years on either side. Joe should certainly have more RBI than most of his contemporaries during that time, since Joe played 16 seasons during that time while the average RBI man who played in Joe's era could only be expected to play 10 seasons.

The top ten RBI men from 1981 to 2000 :

Cal Ripken        1627
Harold Baines 1573
Eddie Murray 1519
Joe Carter 1445
Barry Bonds 1405
Chili Davis 1372
Jose Canseco 1358
Mark McGwire 1350
Rafael Palmeiro 1347
Gary Gaetti 1341


While Joe doesn't show up as the RBI machine here, fourth overall behind two first-ballot Hall of Famers and Harold Baines isn't too bad.

Overall, there are 42 men (plus Dante Bichette) who had 1000 RBI in that timespan, and those are the guys I will study here. 1000 is a nice cutoff because Rickey Henderson and Tim Raines are just below, at 973 and 964 (along with Mo Vaughn, Julio Franco, B.J. Surhoff, Jay Buhner, and Greg Vaughn) and it's nice to cut them out of the study. It's still a diverse group; it runs from Wade Boggs and Paul Molitor to Kent Hrbek and Cecil Fielder.

2. RBI per opportunity

This is what truly separates the RBI men from the boys... or the Paul Molitors and Tony Gwynns, anyway. More than raw totals, the true RBI machines are the guys who drive in runs often, as opposed to guys who just get a lot of trips to the plate.

There are two important measures here. The first is RBI per plate appearance, which measures how good a guy was in driving in runs based on the chances he got at the plate. The second is RBI per out, which measures how many runs a guy drove in, based on the number of opportunities he took away from the team.

In RBI per 100 plate appearances, Joe Carter ranks 14th among the 43 leading RBI men of his generation, with 15.4. The top ten:

Juan Gonzalez   19.8
Albert Belle 18.6
Mark McGwire 18.5
Frank Thomas 17.4
Jose Canseco 17.4
Ken Griffey Jr. 17.4
Cecil Fielder 17.0
Dante Bichette 17.0
Jeff Bagwell 16.8
Sammy Sosa 16.6


Now personally, I would submit that these guys are the RBI men of Joe Carter's generation. If you want to rank only the guys who are within four or five years of age to Carter, then he still ranks behind Canseco and Fielder and Galarraga and Darryl Strawberry.

In RBI per 27 outs, Joe Carter ranks 24th among the 43... below-average. The top ten:

Mark McGwire    8.03
Frank Thomas 8.00
Juan Gonzalez 7.79
Albert Belle 7.53
Jeff Bagwell 7.36
Ken Griffey Jr. 7.27
Jose Canseco 6.88
Barry Bonds 6.73
Cecil Fielder 6.70
Dante Bichette 6.54


This is the true measure of how a player's RBI impact his team. Guys who can drive in runs without taking opportunities away from their teammates are the RBI guys you want.

3. RBI as a percentage of offensive value

I used Runs Created as a measure of offensive value, then looked at how many RBI a guy had relative to that Runs Created number. A guy who had absolutely average offensive skills (in an absolutely average context) would have a 100% figure here... he would drive in exactly as many runs as he had RC).

Joe Carter is the champ here. Well, Joe and Cecil Fielder (another player on this year's Hall of Fame ballot). Joe's RBI were 20% larger than his overall offensive value, coming 2nd out of the 43 players. The top ten:

Cecil Fielder   121
Joe Carter 120
Matt Williams 118
Ruben Sierra 116
Gary Gaetti 116
Juan Gonzalez 116
George Bell 115
Dante Bichette 113
Jose Canseco 111
Andre Dawson 110


These are the guys for whom the RBI represents what they do best as offensive players. It doesn't say much about their actual value, since value is used as a benchmark here. But it's an interesting list. Another way of looking at this list, if you like, is that these are the guys most overrated by looking at their RBI totals. (Most underrated? Boggs, Gwynn, Molitor, Barry Bonds, and Ryne Sandberg).

4. Run Elements

I will always remember Bill James's shorthand for calculating RBI in his Brock2 projection system, because it gives a great succinct (approximate) formula for calculating how many RBI a player will amass...

RBI = .235*TB + HR

This works very well even today, by the way. For the majors as a whole, in 2003 this estimate is off by just 1%.

For Joe Carter, this estimate for his career gives 1315 RBI, which is 130 less than his real total of 1445. Carter's total of +130 ranks very well, 5th overall out of the 43, but not well enough to capture the title "RBI machine of his generation". The top ten:

Paul O'Neill    +185
Harold Baines +170
Wally Joyner +167
Eddie Murray +137
Joe Carter +130
Ruben Sierra +130
Dante Bichette +124
Mark Grace +107
Cecil Fielder +105
Dave Winfield +105


Incidentally, some of Joe's positive total is explained by added opportunities, not by hitting well with runners in scoring position. Over the 10 years for which Retrosheet has data, Joe hit .270/.335/.453 with RISP, as opposed to .263/.308/.467 overall. His lower SLG with RISP wouldn't result in an abnormally high number of RBI.

So there you have it. I think, overall, that guys like Juan Gonzalez, and even Harold Baines, are more deserving of the title "RBI Machine of his generation" than Joe Carter.

For Jays fans, he'll still be Our Hero.
Joe Carter, Ribbie Machine | 20 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
Pistol - Friday, January 09 2004 @ 11:27 AM EST (#81570) #
I believe it was ESPN’s magazine which had something on ‘expected RBIs’ some time during last season. Basically, they gave an expected RBI value given for each possible situation which I believe was based on historical results (one for example would be, with runners on 2nd and 3rd with 1 out the expected RBIs in an AB is 1.1 RBIs).

From there they took each player and all the scenarios that the batter came up in, and compared the actual RBIs to the expected RBIs to see which batters were driving in more runs than expected. I don’t believe it was done, but you could go further and measure the runs above average per AB to make good comparisons between players.

What expected RBI did was largely limit the affect that teammates had RBI totals.

I can’t remember seeing anything on this anywhere else, but I think it does have value and am surprised that no one has explored this further (that I know of).
Mike D - Friday, January 09 2004 @ 11:28 AM EST (#81571) #
First of all, Craig, this is tremendous work! Your third and fourth points, in particular, are really interesting. It's thought-provoking and well-written.

I'll try and sneak one pseudo-critical thought in before the choir to whom you're preaching echoes its agreement. After all, Joe Carter's name has become a punchline in the analytical community.

Guys who can drive in runs without taking opportunities away from their teammates are the RBI guys you want.

This is the much-documented and difficult-to-argue point about the gaping hole in Joe's offensive game -- his penchant for making outs. It's also the classic argument against overvaluing RBI. It doesn't do what your other data does, in my opinion, which is to take the "ribbie machine" argument at face value, and to meet it head-on.

RBI per 27 outs, to me, is a back-door use of Joe's poor OBP, and nothing more.

Craig, does your data include the number of plate appearances with runners in scoring position? A comparative analysis of RBI per RISP PA would be interesting -- it would get us closer to that still-unaccessible Elias "RBI Efficiency" matrix (which showed this year that Carlos and Barry have a decided skill for knockin' em in, even after controlling for number of opportunities).

The RBI per 100 plate appearances is valuable, and when combined with Joe's career numbers with RISP, it supports your conclusion that Joe was a good, but by no means pre-eminent, RBI man.
Mike D - Friday, January 09 2004 @ 11:30 AM EST (#81572) #
The "RBI Efficiency" matrix is what Pistol's referring to. I also read about it in ESPN Magazine, and Pistol's right: It basically evaluated RBI against "expected RBI," and thereby took the role of teammates (for the most part) out of the equation in a hitter's RBI totals.
Pistol - Friday, January 09 2004 @ 11:35 AM EST (#81573) #
http://www.baseballstuff.com/btf/scholars/ruane/articles/rbi_production.htm
I googled expected RBI and found an article on it at Baseball Stuff, which coincidentally used Joe Carter as an example. COMN.
_R Billie - Friday, January 09 2004 @ 11:36 AM EST (#81574) #
The value Joe brought to the Jays was more character and durability than anything else. He was a useful player but probably the weakest offensive performer among the stars of the 1993 WAMCO lineup. He'd always managed a lot of extra base hits and total bases while hitting cleanup behind good OBP players which is the real source of his RBI as opposed to some mystical ability to drive in 100 runs. If Joe had a mystical ability it was staying healthy over a full schedule of games year after year and hitting some pitches which were out of the strikezone outside. I never really thought of Joe as a special offensive player but he was an easy guy to get behind and it was fun to watch his RBIs roll over 100 at the end of every year.
robertdudek - Friday, January 09 2004 @ 11:40 AM EST (#81575) #
Mike D,

Making outs takes RBIs away from teammates, so it has to be factored in in an assessment of the value of a player's RBI abilities.

RBI per plate appearances penalizes the guys with high walk rates. Maybe one ought to use RBI per AB, which is a happy medium between PA and Outs.
Craig B - Friday, January 09 2004 @ 11:46 AM EST (#81576) #
Quick hits...

Unfortunately, the data doesn't have RISP opps. Retrosheet has this data up to 1992, and if someone wanted to study that, I can give them a hand. E-mail at tybalt4@yahoo.com!

RBI per 27 outs is intended to measure the team impact of RBI. Yes, it is a bit like RBI * OBP. But the outs are the unit in which runs always have to be measured. I like Robert's suggestion of RBI/AB as a "happy medium", and thought of including it, but it would be repetitious. The long and short of it that among high-RBI guys, Carter ain't noticeably good.

If Joe had a mystical ability it was staying healthy over a full schedule of games year after year

Yes, this is something I didn't address. Joe was durable as hell.

Pistol, good article.
_coliver - Friday, January 09 2004 @ 12:32 PM EST (#81577) #
Despite the fact that most of my comments are slammed, I am honored that I was mentioned in the introductory paragraph of "Joe Carter, Ribbie Machine".

I now feel like a full-fledged member of the Battersbox.ca community.

It is the best Jays site on the web.
_S.K. - Friday, January 09 2004 @ 12:40 PM EST (#81578) #
Ahhhh WAMCO... which later became WHACOM, or HWMCOA, but which really should have been HAMCOW all along.
_Mark - Friday, January 09 2004 @ 01:14 PM EST (#81579) #
Nice article! Point #3 is a really interesting way of looking at the data. That list certainly matches up with who I expected might be on it (except Canseco maybe). Minor nitpick, I'm not sure the title for #3 is exactly accurate, it's more of an index rather than a percentage since you're taking a ratio of two completely different metrics.

As that simple Bill James formula suggests, rbi is a largely a function of slugging average and opportunities. Inherent in the opportunities is the player's position in the batting order. Getting 100-110 rbi hitting fourth in a good batting order isn't evidence of some special skill. While the public seems to understand that leadoff hitters have low rbi totals because they have fewer opportunities, somehow they don't see that cleanup hitters are expected to have higher totals, due to their position. Instead, it's seen as some sort of skill.

Just for kicks I ran a list of all players with a .500 SLG in 400 AB (1753 players up to 2002). Lowest RBI total: 44 (Sixto Lezcano, 1979, .502 SLG, 400 AB, right at the cutoff). The median # of RBIs is around 103 (higher than I might have guessed). Our man JC is only on the list 3 times (1986, 1991, 1994).
_steve - Friday, January 09 2004 @ 02:01 PM EST (#81580) #
although all these statistics are mostly against joe carter. i would like to say that joe carter usually does come up with the timely RBIs. anytime there is a guy on third base, he would usually drive him in. i think carter was a fairly good situational hitter.
_Mick - Friday, January 09 2004 @ 02:58 PM EST (#81581) #
I knew Paul O'Neill was a Hall of Famer!
_SportsmanTO - Friday, January 09 2004 @ 03:37 PM EST (#81582) #
Interesting stuff and quite thought provoking. I heard Joe the other day talking about how he was left out in the cold by the baseball writers on the radio. (FAN 590) He sounded pretty upset but he knew deep in his heart that he may not be HOF material but he was deeply hurt and upset that he got less than 5% of the vote. He compared his career to that of Tony Perez which I think may be a good comparison because both men played for big offensive machines. I wonder, with this data how the 2 would fare against each other. Since I know next to nothing about finding/calculating that data I can't do it :(

Anyways I think he provided more of the intangibles than being a great RBI guy and another thing that might've hurt him was that at the end of his career, his performance really dropped off the table.
_Mark - Friday, January 09 2004 @ 05:59 PM EST (#81583) #
uh, correction on my tangential earlier post, what i meant to say was sixto lezcano, 1977, 400 AB, .502 SLG, 49 rbi. and an honorable mention to lou brock in 1964 for 419 AB with the cards, 44 rbi, .527 SLG.
Craig B - Saturday, January 10 2004 @ 11:37 AM EST (#81584) #
Yeah, but Lou Brock actually only slugged .464 in '64, since he played 52 games for the Cubbies as well and slugged .340 there (before the famous trade for Ernie Broglio).
Joe - Saturday, January 10 2004 @ 12:35 PM EST (#81585) #
http://me.woot.net
Hasn't it been established that 'timely hitting' doesn't exist?
Coach - Saturday, January 10 2004 @ 01:05 PM EST (#81586) #
RBI per 27 outs, to me, is a back-door use of Joe's poor OBP, and nothing more.

In the article that Pistol linked, Tom Ruane looked at OPS, not just OBP, and concluded that Carter might have had the worst season ever for someone with 100+ RBI.

The guy remains a Canadian hero for that one great moment, something like Paul Henderson. Though Joe was a lot better player than he was a commentator, he wasn’t nearly as good as Dawson, Parker, Rice and several other borderline HoF candidates. That said, I’m surprised he didn’t get more votes from the writers, and I understand his disappointment. I thought he'd stick around on the ballot for a few years, at least.

Thanks, Craig. This was fun, except for Mick raising my blood pressure with his O'Neill remark.
robertdudek - Saturday, January 10 2004 @ 01:24 PM EST (#81587) #
Hasn't it been established that 'timely hitting' doesn't exist?

Timely hitting surely exists. Joe Carter's season ending homerun in 1993 was timely.

A hitter who can produce timely hits with significantly greater frequency in key situations than the average major league hitter has not been shown to exist. Well, there are a few guys who did do very well in RISP or late inning/close situations over their careers, but we would expect such outliers to occur by chance given the thousands of players who've played regularly in the majors (similar idea to a coin having a long run of 'heads', though it is not weighted - see Rosencrantz and Guildenstern are Dead). Thus we can't be sure whether those outlier were or were not clutch hitters.

Clutch hitters haven't been proven NOT to exist. The ability (if it exists) may be so small that it is not detectable given the sample sizes of a player's career.
_steve - Saturday, January 10 2004 @ 02:53 PM EST (#81588) #
example of timely hitting

most of tony batista's home runs when he had a big year were in the 8th and 9th innings.
_Mark - Saturday, January 10 2004 @ 06:12 PM EST (#81589) #
Craig B: Yep, that was why it was an honorable mention :) First time with my query on the Lahman database I forgot to group players by year and player so it counted each of the stints with each team separately and so that came to the top.
Joe Carter, Ribbie Machine | 20 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.