The Ten Most Questionable MVP Awards in Baseball History

You Think Ichiro! Was Bad? Check Out These Choices…

I know that everyone’s favorite sub-species of homo sapiens, the stathead, is still reeling at the recent announcement that Ichiro Suzuki was the American League Most Valuable Player of 2001.

The hue and cry (possibly the most redundant phrase extant: look it up sometime…) over this occurrence is not likely to diminish for some time. While many sportswriters (including my old sparring partner Rob Neyer at ESPN.com) have tried to put a band-aid over this, for most dyed-in-the-wool (yet another highly redundant phrase—where ELSE would you put the dye, unless you were casting it, that is?) stat-mongers, this is the worst voting outrage since…

Since…

Now that’s the question that should be on the minds of those venting their frustrations at the failures of baseball’s meritocracy. (The failure of meritocracy in general is a more interesting topic, but we’re in the wrong line of work for something like that.)

Just what is the anatomy of error in Most Valuable Player voting over the 71-year history of the award? Just how often have the baseball writers picked the same player that “advanced statistical methods” designate as the best player of the year in each league?

And which have been the most questionable MVP picks? Where does the selection of Ichiro Suzuki rank in that anti-pantheon of decision-making?

To do such a study properly, mind you, would take a book. That book, however, would run the risk of tedium, because it would have to evaluate a whole set of statistical measures, attempt to discern particular patterns of bias, and generally go into more detail than even the average stathead is willing to wade through. The general fan population would stay away from such analysis in droves.

So what can we do to summarize this? What I decided to do was lean on the only current method that is readily available for purposes of historical comparison—Total Baseball’s rating system, known to most of you as Total Player Rating (TPR). (As you know, Bill James has a new method out—sort of, that is; while we’ve been assured by his legion of admirers that it’s wonderful, we don’t have all the data, so we can neither evaluate it nor use it.)

So it’s TPR to the rescue. Here’s how it works: we look at the TPR of the player winning the MVP award in each league for every season since 1931 (when the BBWAA was authorized to conduct the vote) and compare it to the TPR leader for that season. The difference (if any) is logged, and then we see which MVP winners are furthest away from the “sabermetric MVP” as determined by Total Player Rating.

Once we do that, we can assemble a “top ten” most questionable MVP picks list.

There are some caveats that those of you who wish to check the data here must keep in mind, however. As many of you know, TPR has some quirks—mostly in how it handles defense. In more than a few cases, its component for this aspect of player performance—measured by a stat called “fielding runs”—is clearly out of reasonable bounds, and colors the overall rating.

What I’ve done is to adjust certain seasons where a player’s rating has been unduly affected by this quirk in the system. When “fielding runs” represent more than 25% of a player’s rated performance, they’ve been removed from the ranking.

This was necessary in just about two dozen seasons (out of 142 MVP awards). You’re free to disagree with that decision, but remember this is meant to be a quick-and-dirty (ever wonder why there’s no contrasting catchphrase ... “slow-and-clean”?) look at the matter in question.

Before we look at the top ten list, here are a few related facts:

—The writers and TPR have agreed on the identity of the MVP 30% of the time (43 times). The NL writers have matched TPR 20 times, while the often-malinged AL writers have done so 23 times.

—The average “distance” from the media choice and the sabermetric choice is 1.16 TPR. It’s a little higher in the AL (1.21) than in the NL (1.12), which means that the writers in the AL—remember, they’ve matched the sabermetric choice a few times more than the NL writers have—get further out of whack than their NL counterparts.

—It’s clear that the introduction of the Cy Young Award in 1957 has affected the way writers think about the MVP award. Prior to existence of the Cy (from 1931-56), pitchers won 11 MVP awards out of 52 total, or 21%. From 1957 on, pitchers have won only 8 MVP awards out of 90, or 9%.

Another related stat has to do with the number of pitchers who are ranked as the “sabermetric MVP” but who were not voted the award from the writers. From 1931-56, this happened six times (about 11.5% of the time). Since 1956, it’s happened 14 times (out of 90 possible awards, or 15.6% of the time).

—A big factor in winning the MVP award is playing on a team that wins a pennant or a division. Out of 142 MVP awards, 99 of them (70%) have been given to players on such teams. The AL voters correlate more highly (73%, of 52 of 71, as opposed to the NL’s 66%, or 47 of 71).

Non-pennant winner MVP recipients are a bit more likely to match up between the writers’ assessment and the sabermetric choice. Fifteen of the writer-TPR matches are from the forty-three players on non-pennant winners, or 34%. The writer-TPR matches on pennant/division winners are a bit lower (28%); the closeness of these numbers indicates that the two variables probably have little relationship to one another.

—The AL writers have been in a serious slump over the past decade. They were ahead of the NL going into 1990, but their mean distance from the sabermetric winner has been the worst of any decade since the award was created (1.94). This fact is reflected both in the noisy disapproval of statheads, and in the Top Ten list of most questionable MVP selections that you’ll find below.

Another way to look at MVP voting patterns would be to catalogue how the players’ component stats rank in the league. Baseball writers haven’t had the benefit of sabermetric measures for more than seventy percent of the time frame in question, and they clearly have demonstrated more than a little resistance to those tools in the past twenty years, preferring to rely on a hard-to-pin-down combination of traditional stats. (It’s often charged that only the “Triple Crown” stats—batting average, home runs, and RBI—get utilized, but the evidence shows that this isn’t really the case.) Such a statistical anatomization, however, would be part of that book that we’re NOT writing here…

10. Dick Groat, 1960 NL (-3.1)

Groat, shortstop for the pennant-winning Pittsburgh Pirates, led the league in hitting (.325) and finished third in hits (186). TPR suggests that Willie Mays was the best player in the NL that year (first in hits, fourth in runs, stolen bases and slugging average, third in triples, total bases and batting average), though he had strong competition from Ernie Banks, Hank Aaron and Eddie Mathews), but the Giants finished fifth (in an eight-team league), 16 games out.

9. Ken Boyer, 1964 NL (-3.4)

Boyer, the long-time star at third base for the Cardinals, was apparently rewarded for his fine career as a result of St. Louis’ September pennant charge. He was the RBI leader (119), and finished fifth in walks (but with only 70; this was the sixties). TPR shows another close race for top player between Mays (who led in HRs, doubles and slugging average), and Ron Santo (who led in triples, on-base percentage and walks).

8. Jackie Jensen, 1958 AL (-3.4)

Jensen, playing right field for the Boston Red Sox, batted behind Ted Williams for most of the year and wound up leading the league in RBI (122). He was fifth in total bases, doubles and homers, second in bases on balls. Boston finished third, but was only four games over .500. It’s likely that the aversion to consecutive awards that the BBWAA has shown is at work here; Mickey Mantle had won the MVP in 1956 and 1957, and the writers apparently were loath to hand it to him a third time, even though TPR shows that he was the top player in the AL.

7. Maury Wills, 1962 NL (-3.4)

Seduced by all those stolen bases, the NL writers voted for Mr. Excitement, the ever-aggrandizing Mr. Wills, who, in addition to those 104 steals (a new record at the time), was second in runs, second in hits, and led the league in one other offensive category—triples. (Actually, he was tied for first with three other players). TPR, as was often the case from 1954-65, deemed Willie Mays to be the top player in the league. Wills and the Dodgers probably would have won the pennant in a breeze if Sandy Koufax’ finger hadn’t impersonated a grape midway through the season, but they didn’t, dropping a playoff series to Mays’ Giants that was—well, a lot stranger than its predecessor in 1951.

So far we’re 2-2 in terms of pennant/division winners vs. also-rans for our “questionable” MVP winners.

6. Mo Vaughn, 1995 AL (-4.0)

Now here’s one that will get the foam oozing from the mouths of statheads everywhere. The AL BBWAA probably wasn’t as bad at MVP voting as it looks, if you consider that they would not, could not, and did not bring themselves to vote for Albert Belle, who was clearly deserving of the award in 1995 and 1998.

With that ambient condition firmly in place, Vaughn was the lucky recipient of the booby prize, thanks to his tie for first (with Belle) in RBI. Vaughn cracked the top five in only one other offensive category (total bases), while Belle led the league in homers, doubles (tied with Edgar Martinez), and slugging average.

The Red Sox did win their division in 1995, but the Indians went 100-44, which isn’t too far off the pace of the 2001 Mariners and 1998 Yankees, who followed in their footsteps of AL domination.

5. Yogi Berra, 1955 AL (-4.3)

The baseball writers of the fifties were in love with catchers, and two catchers in particular: Berra and Roy Campanella, who each won three MVP awards in the first half of the decade. Berra’s awards in 1951 and 1954 weren’t nearly as far-fetched as this one was. Yogi didn’t even the lead the league in RBI in ‘55, finishing third, and didn’t crack the top five in any other offensive category. A bushel of other players were ahead of him according to TPR, led by Mickey Mantle .

4. Frank McCormick, 1940 NL (-4.4)

McCormick, a big, tall first baseman for the Cincinnati Reds, had his two best years for the pennant winning teams of 1939-40. The writers picked the worse of his two seasons, however, to anoint as MVP-worthy. In 1939, Frank at least led the league in RBI. It appears that, as with Boyer, the writers bestowed a kind of “cumulative” award here, based on the fact that McCormick had finished 1-2 in RBI in these two years, 1-1 in hits, 4-1 in doubles, 2-2 in total bases, and somewhere near the bottom each year in walk percentage. TPR suggests that either Claude Passeau (the Cubs’ crafty righty who won 20 games for a sub-.500 team) or Arky Vaughan were the best players, but Johnny Mize would have been my #1 vote had I been there.

3. Marty Marion, 1944 NL (-4.5)

This is probably an award cooked up in the back room somewhere, as I’ve always suspected was the case when Willie Stargell somehow managed to “tie” Keith Hernandez in 1979. (I don’t know what the procedures are for safeguarding the MVP vote, but they’re probably even worse than the conditions in Florida this time last year.)

Marion was a great shortstop, and the Cards had just won their third pennant in a cakewalk, but Marty is the only player to win an MVP award without finishing in the TOP TEN of any offensive category. He beat out the Cubs’ Bill Nicholson by one vote, which is seven less than Ichiro’s margin over Jason Giambi. TPR says Stan Musial, but it’s taken 71 years (and 73 homers from Barry Bonds) to get past the three-MVPs-to-any-one-player unwritten rule.

2. Mickey Cochrane, 1934 AL (-5.1)

Lou Gehrig won the triple crown in 1934, but the Cochrane-managed Detroit Tigers won the pennant. Cochrane finished fifth in OBP, but he was not even on the same lap as Lou. Manager of the year, yes; MVP, no.

1. Ivan Rodriguez, 1999 AL (-5.3)

Here it is, the most questionable MVP pick to date. Pudge had a nice little season in 1999, and the Rangers did win their division, but he finished just fifth in hits, seventh in runs, and tenth in slugging average. Pedro Martinez, one of those pitchers victimized by the increasing role specialization in award voting, was the top player in the league according to TPR (and a whole bunch of the rest of us).

Those of you who plunked down your bucks for the most recent Total Baseball (edition #7) may find that the 2000 AL MVP, Jason Giambi, looks to be about 4.0 TPR behind Pedro Martinez. That’s right—so why isn’t he on the list? The main reason is that fielding runs not only pad some player’s TPR totals, they take away too much for others. Jason now has the reputation of being a sub-par first baseman, but the new fielding run computation system that Pete Palmer introduced penalizes first basemen a bit more than it used to, and drops Giambi’s ranking with a thud. The distance between Giambi and Martinez is certainly not as great as the system represents it as being, so I’ve compensated by using Giambi’s offensive ranking (batting runs, in the Total Baseball system) as the measure being compared with Martinez’ performance.

What’s the final tally for players on pennant/division winners vs. also-rans? Eight-to-two for winners.

So where does this leave Ichiro Suzuki? How does his TPR compare to the best player in the AL in 2001?  Exactly how big is that gap? I’ll leave that question for the active number crunchers among the stathead brigade. My hunch is that Ichiro probably escapes the Top Ten, however. While his OPS is about 300 points lower than Giambi’s, he’ll gain ground in fielding runs and even more in stolen base runs (his 56 steals in 70 attempts are worth about seven such runs in Palmer’s system).

Somebody out there is going to post the answer, so I’ll just sit back and let the fur fly (another one of those ridiculous sayings: the only way this happens is when rich dames in mink coats get on airplanes).

That said, it’s still a seriously questionable choice. The AL writers seem to have been groping (and they get a lot of complaints about this from the three feminists who attend baseball games…) for some way to pay tribute to the Mariners’ amazing season; they looked at Ichiro’s hit total—242, the ninth highest total in baseball history—and enough of them became mesmerized by it to tilt the contest in favor of the first man in baseball history whose surname is the same as a brand of foreign car (no, Richard Dotson doesn’t count; as Johnnie Cochran, now a native New Yorker, once said: if the spelling don’t fit, you must acquit).

What’s my hope out of all this? That statheads will not escalate their derision over this vote into full-fledged disdain of Suzuki—who had nothing to do with the actual vote. The phenomenon is known as “transference”: statheads have a deep-seated ideological problem with singles hitters, and an event such as this one is just what the doctor ordered (as opposed to the apple a day…) to push them into a merciless air strike of negativism. This little history lesson was meant to re-affirm that the meritocratic process is inherently flawed, and that no amount of wailing will tear down the wall. (And frankly, anyone who thinks that the Pink Floyd was worth a pitcher of warm spit after Syd—not Marty—Barrett left needs to drink that pitcher of warm spit.)

Ichiro Suzuki—innocent victim of baseball’s bizarre lip service to multi-culturalism. And you thought that those Luis Bunuel films were surreal…