In 1980, the Elias Sports Bureau – baseball's statistical keepers – quietly introduced a new statistic into the sport's vernacular: the game-winning RBI. The introduction was so quiet, in fact, that I can't find a single article mentioning the new statistic in 1980. Instead, it just suddenly appeared in box scores that spring, innocently tracking players' ability to hit in the clutch. Or so Elias hoped.
The phrase "game-winning rbi" invokes images of big hits, players coming through in the clutch to put their team ahead for good. A two-run double in the bottom of the eighth to make it 3-2, a tenth-inning leadoff home run, even a bases-loaded sacrifice fly… The problem with statistics, though, is that they need a rigid definition to be useful; a vague "I know it when I see it" just won't do. The game-winning RBI was defined in Rule 1004-a as "the RBI that gives a club the lead it never relinquishes."
It sounds simple enough, but therein lies the problem. As any of a thousand different articles complaining about the statistic would tell you, that simple definition means that an early-inning bloop single can be just as much of a game-winning RBI as a two-out, ninth-inning grand slam. As one 1989 article put it, "the GWRBI went to the player who drove in the run that put his team ahead for good. That might be a grand slam in the bottom of the ninth inning for a 10-9 victory or a weak grounder in the top of the first in a 13-0 blowout."
The weaknesses in the statistic are best highlighted in this New York Times article, published in the wake of its demise (Elias officially retired the statistic after the 1988 season):
The game-winning r.b.i. was, in truth, a misleading measure of clutch hitting, the much-admired and ceaselessly debated-over ability to produce runs under pressure. … The main problem with the game-winning r.b.i. was that it did not adequately distinguish the achievement it was meant to reward. Even a casual fan understands that hitting a routine single in an early inning to drive in the first run of a 10-0 victory is a lesser accomplishment than clobbering the decisive home run in the bottom of the ninth. Yet each qualified as a game-winning r.b.i.
Further, the statistic did not account for conditions or occurrences that are beyond a player's control. Any player, no matter what the quality of his team, has himself to blame if his batting average is low. But as John Thorn points out, ''teams without g.w.'s don't get many g.w.r.b.i.'s.'' A corollary is that the statistic failed to recognize genuinely notable feats. ''Suppose you hit a grand slam in the top of the ninth to put your team ahead,'' said Steve Hirdt, the executive vice president of the Elias Sports Bureau. ''You've done a great job. If your team loses the lead in the bottom of the ninth, does that diminish what you've done?''
Players didn't like the statistic either.
"Good. It was an absolutely ridiculous statistic,'' Mike Schmidt said. “I believe in the potential of the stat, but not the way it was."
"It could have been a better stat, maybe for something from the seventh inning on," Keith Hernandez said.
Hernandez holds the all-time record, and always will, with 129 GWRBI. He set the single-season mark with 24 in 1985 with the New York Mets. "If you win 10-0, you shouldn't have one. You hit a sacrifice fly in the first inning, there's not that kind of pressure early in the game," he said. "Still, it's nice to be the leader. I'm sure I have a lot of legitimate ones mixed in with the other ones."
"I just think it's a stupid stat," [George Brett] said. "I could understand having it for games that are won from the sixth inning on. But who's going to decide what the game-winner is before that?"
"There should be one. Maybe it could be treated like a save," [Mark McGwire] said. "It could start in the seventh inning and be used in games decided by three runs or less."
Clearly, the drawbacks of the young statistic – the obvious inclusions of non-clutch situations in what was meant to be a clutch statistic – were too much to keep it alive.
"It'd been on its last legs for a couple of years," said Bill Murray, chairman of the baseball committee and also its subcommittee on scoring rules. "There had been requests for its elimination, from sportswriters and other interested parties. It sort of lost its appeal."
The death of the game-winning RBI was almost as quiet as its birth. Other than the two articles quoted above, most mentions of the statistics's demise early in the 1989 season had more to do with the fact that Rafael Palmeiro ended the 1988 season with the second-highest batting average in the National League but zero game-winning RBIs, a quirky stat Chicago and Texas writers found funny thanks to the winter trade that moved the first-baseman to the Rangers.
It's now been over twenty years since the last official game-winning RBI was tabulated by Elias. This is not a bad thing. As Mike Schmidt, George Brett and the rest of the Hall of Fame-level talent said that spring, it just wasn't a good statistic. It's purpose was questionable to begin with and it's implementation was terrible. If a ballplayer who does nothing but benefit from the statistic (like Keith Hernandez) can see its flaws, then that's a good sign that something is wrong with it.
It makes you wonder, though. If everyone could so easily see the weaknesses in the language of the game-winning RBI, why can't they see the same weaknesses in the language of Rules 10.17(a) and 10.17(d)? In case you need a refresher, that language reads:
10.17 (a) The official scorer shall credit as the winning pitcher that pitcher whose team assumes a lead while such pitcher is in the game, or during the inning on offense in which such pitcher is removed from the game, and does not relinquish such lead…
10.17 (d) A losing pitcher is a pitcher who is responsible for the run that gives the winning team a lead that the winning team does not relinquish
At least with the game-winning RBI, the credit was given to the player who actually knocked in the run, not the pitcher who just so happened to be on the mound that same inning. It's a question for which I don't ever expect to hear an answer.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.Subscribe now
While we're at it, maybe someone could shed light on when MLB started listed "2-out RBI" as a separate entry in the box score, in addition to the plain old "RBI." Is this another attempt to measure "clutch-ness" or what? And why do we list "Runners left in scoring position, 2 out"? I would rather publicly humiliate the guy who left a RISP with 0 outs or 1 out. It's tougher with two outs.
Finally, why list RBI in the batting stats when it is already in the line score? Here's the listing from WS Game 7 last year with the RBI and 2-out RBI separate listings. Why are we emphasizing what happened with two outs?
2B: Freese (3, Harrison).
HR: Craig (3, 3rd inning off Harrison, 0 on, 1 out).
TB: Craig 4; Berkman; Freese 2; Molina, Y 2; Furcal 2.
RBI: Freese 2 (7), Craig (5), Molina, Y 2 (9), Furcal (1).
2-out RBI: Freese 2; Molina, Y; Furcal.
Runners left in scoring position, 2 out: Molina, Y; Carpenter 2; Schumaker 4.
Team RISP: 2-for-9.
Team LOB: 8.
Good stuff Larry - you are a must-read at BP.
and wish that someone at BP would run these numbers.