In assessing a league’s MVP, no statistic is looked at more closely than the RBI: what wins games is runs, and he produces – i.e., drives in – the most runs must be the most valuable. But looked at more closely, the rbi, like many of baseball’s treasured stats, is somewhat suspect.
One, not all rbi’s are equal. The run batted in late in a close game surely should be valued more than an rbi in a blowout. Specific rbi situations could be assigned specific values – like the elements of a gymnastics floor exercise – but they aren’t.
Two, rbi opportunities are not spread around equally. Except for rbi’s via the homerun, they depend on having teammates already on the bases, preferably “in scoring position” – i.e., standing on second or third. The cleanup hitter usually leads his team in rbi’s, not only because he is inherently the best rbi-man, but because he typically bats after the teammates with the best on-base percentage. Statistically, he will have the greatest rbi opportunities on the team. A leadoff hitter, by contrast, will have the fewest, because he leads off the game and thereafter bats after the weakest-hitting teammates. A more accurate assessment of rbi value should require that the number of rbi’s be divided into the number of rbi opportunities presented. What that last figure itself should be is debatable, but let’s say it’s men-in-scoring-position for a start.
The above, of course, gets us into a bit of a chicken-and-egg debate. Is the cleanup hitter the team’s best rbi-man because he is the cleanup hitter? Or is he the team’s cleanup hitter because he is the best rbi-man? This will never be answered unless and until the stat is refined as suggested above.
P.S. The mention of “men-in-scoring-position” brings to mind a relatively new stat that has been prominently mentioned in connection with the Twins this year (2008), because they lead the majors in batting average with runners in scoring position (“risp”). The underlying basis for keeping such a figure is that, of course, runs win ballgames and hits with risp produce runs. Why this is not necessarily a valid assumption was illustrated, not for the first time, in yesterday’s Twins 6-4 loss to the Tigers. The Twins were 4-for-12 in risp situations, a healthy .333 average, well above their league-leading .311 mark. Of those 4 hits, though, one was Joe Mauer’s single in the first with two men on; the lead runner, Denard Span, was held at third. In the 9th, Alexi Casilla bunted safely, again with two men on. No run scored in that inning. The Twins, because of all their banjo hitters, undoubtedly also lead the league in hits with runners in scoring position that do not produce runs. Again, for this statistic to have real meaning, it should be tweaked, to reduce, if not eliminate, the value of hits that don’t produce runs.
The flip side of this statistic, of course, was shown in how the Tigers won the game: two homeruns with, in each case, a man on first. No hits with a runner in scoring postion, but four runs.
September 2008

The “Save”

Among official baseball statistics there are a few pure ones, like batting average. Unlike rbi, batting average doesn’t depend on what the hitters ahead of you have done, nor, unlike runs scored, on what the hitters who follow do. It is also valid as a measure of worth: someone who hits .300 will be an asset to the team; someone who hits .225 will not. By contrast, a player can hit three home runs and still be valuable, or field at a .995 clip and still be a defensive liability. But among all the impure stats, one stands out because not only is it often invalid as a measure of worth, it alone affects how the game itself is played.

The “save,” first of all, didn’t exist until 1969. Baseball got along fine for all those years without it. It was introduced to acknowledge the increasing importance of the relief pitcher; but it, in turn, has itself increased that importance. Counting the number of saves has contributed to the mystique of “the closer.” As this mystique has grown, it has become legend that “the ninth inning belongs to the closer.” As a result, there are fewer complete games than in the old days. Once the starter reaches the ninth and a save situation develops, most managers will turn to their closer now, regardless of how well the starter performed in the eighth.

The next problem involves the definition of a “save.” If you enter the 9th with a three-run lead and finish the game with your team still ahead, you get a save, regardless of how well or poorly you pitched. If you give up three hits, two walks and two runs while recording only three outs, that would give you an ERA of 18.00 but you still get a save. Take another example: your team is leading by three runs in the 9th, there are two outs, there is a runner on first. You come in, throw one pitch, the batter drives it 400 feet to deep centerfield, where the ball is caught. You get a save, because the potential tying run was on deck when you entered the game. Or worse: the ball is not caught, but the runner is thrown out at home. You have thrown one pitch, given up a double and recorded a save.

On the other hand are the Mariano Riveras and Goose Gossages who enter the 9th with a one-run lead and strike out the side for their save. The wide disparity in the degree of difficulty for different save situations and the fact that a bad performance can reap the same statistical reward as the good performance are major flaws in the save category. Another obvious problem: you can only get a save when your team is ahead near game’s end. The closer for the New York Yankees has a major head start over the closer for the Washington Nationals (see, e.g., Rivera and Gossage, cited above, not to mention Sparky Lyle).

What bothers me most, however, is the way this statistic doesn’t just reflect the game, it affects the way the game is managed. To wit, Ron Gardenhire (and he’s just my local example) generally won’t put Joe Nathan in a game unless and until it’s a save situation. If the Twins are winning 5-2 at home with two out in the bottom of the 8th, Nathan will be warming in the bullpen, ready to come in. If a Twin then hits a home run, building the lead to 6-2, Gardy will sit Nathan down and bring in a lesser reliever. (Once two batters get on base, however, Nathan will be summoned.) The only magical difference in baseball terms between a three-run and a four-run lead is that one will produce the statistical oddity of a save, while the other will not, yet it has changed the manager’s choice of pitcher.

This is not to say that Rivera, Nathan and Jonathan Papelbon are not valuable to their respective teams. It is to say, though, that the difference between Rivera, who records his saves with cold-blooded efficiency, and Nathan, who produces heartburn as often as relief, is much greater than their number of saves will reflect. I am sure that the baseball stat-heads have come up with a closer efficiency rating that takes care of all the issues I have raised (just as SI informs me of a metric called Equivalent Baserunning Runs that measures baserunning value!). For starters, I would like to see a one-run save valued more highly than a cheap three-run save, with a two-run save in between. A closer is supposed to handle pressure, and there simply is much less pressure when you are ahead by three runs in the ninth. There also should be a penalty for allowing a run, let alone two runs, in a save situation – even to the extent of eliminating the save itself. Finally, the field should be leveled between the good and bad teams by ranking closers on their save percentage, not the save total.

Maybe the manager then would pay more attention to the quality of his closer’s performance and not be so automatic in how he deploys his forces. Even if it’s too late for that, at least the fans will be given an official statistic that, if not perfect, will be less impure.
August 2009