I think WAR is a fantastic cumulative stat. I don’t love it, and it certainly isn’t perfect, but it is better than most single-stat measures of a player, and it serves as a really good starting point while comparing players.
Of course, it is relatively new to the average baseball fan, and it was completely unknown to the awards voters in the 1950’s. Mostly out of curiosity, I decided to take a look at the National League MVP’s of the 1950’s and compare their WAR scores to the league.
I am not implying that just because a WAR score doesn’t match the choice for MVP the choice is wrong. There are plenty of “x factors” involved (as I will address below). But sometimes it’s interesting when an MVP and his accompanying WAR simply don’t match at all. It makes me wonder: what was the story behind that?
In the chart below, you will see the actual NL MVP for each year, their fWAR score, their fWAR rank (in the NL only) and, in the final column, the player with the highest fWAR in the NL that year. Sometimes they match, sometimes they don’t. For the two years in which a Pitcher won the MVP, I have their fWAR rank amongst pitchers (P) and amongst hitters (H), just to give a better sense of where they stood.
If it was left completely up to fWAR, Willie Mays would have had four MVP awards in the 1950’s, and Duke Snider and Jackie Robinson would have had two a piece. As it is, Mays grabbed just one, and the other two were shut out.
But let’s look at what did happen (rather than what might have happened). Roy Campanella won three MVPs that decade, and in each season, he was top 10 in fWAR, finishing as high as 3rd and 4th. Given that he was a catcher, and so his Plate Appearances were relatively limited, the ability to accumulate WAR “points” was also relatively limited, leaving these choices seeming pretty well in-line with MVP. Not to mention that catching statistics were even more limited than they are today, and so his handling of the pitching staff, his pitch calling and pitch-framing are all elements that might have been witnessed, but not quantified.
There were five other non-catcher, non-pitcher NL MVPs that decade. In four of those, the actual MVP was also either 1st or 2nd in fWAR. Not bad, right? In the 5th case (Sauer), he was 4th. Again, not bad at all.
Much like pitch-framing for catchers, we only have so much data on the base-running ability of position players, and we have to use the mildly-inferior Total Zone rating for defense, leaving the WAR scores of the 50’s not quite as reflective of what happened on the field as some of the scores are now (not they they’re perfect yet). Even so, the voters must have been in unwitting agreement with WAR, as seven of their choices for NL MVP were top four in fWAR, and the other three were non-position players (thus clouding the WAR comparison).
Now on to the pitchers. The NL MVP was a pitcher in both 1950 and 1956. In ’56, Don Newcombe was 3rd in WAR amongst pitchers and was 10th when compared only to hitters (leaving him as 12th overall). This is by no means a “crazy” pick, especially when we consider that strikeouts – a major contributor to fWAR for pitchers – were not in abundance in the 1950’s (Sam Jones lead the NL that year with just 176 K’s).
If we were to switch from fWAR (depending solely on K, BB and HR) and look at RA9WAR (more of a reflection of what happened on the field), then Newcombe jumps to 2nd amongst pitchers and 3rd among all NL players. Suddenly things fall into place.
And then we come to Jim Konstanty and his minuscule 1.0 fWAR in 1950. How did this discrepancy occur? He was 65th in WAR that year. Even when we switch to RA9WAR, he ranks 35th overall in the NL. There must have been something else going on… a huge September? Extraordinary leadership? He told great jokes?
Turns out ol’ Jim was a relief pitcher. An old-school reliever, pitching 152 innings, leading the league with 22 saves, “winning” 16 games, and setting a then-record of 74 appearances. He was apparently so good that in Game 1 of the World Series that year, they switched him to start the game (he lost, 1-0). He also had a 2.76 ERA, lowest among all pitchers with 150 IP or more.
So there you have it! This is a really good example of context; I’m not agreeing unequivocally that Konstanty deserved the MVP, but after discovering what his role actually was on that team, it makes a whole lot more sense than simply looking at WAR. [And it still beats one of the weirdest award choices involving a reliever ever: Steve Bedrosian for Cy Young in ’87.]
Overall I’m a bit impressed by the relatively close correlation between fWAR and the awards. I wonder if something like this might help the saber metric cause? I mean, guys who didn’t use much more than their eyes and a few stats (BA, HR, RBI) to judge were actually relatively close to what the objective numbers say.