The brilliant Tom Tango sent along an email that discusses a fascinating point about analyzing baseball (and perhaps everything), something that has to do with consistency. Maybe the best way to explain it is through Bill James' classic piece on Ken Keltner and the Hall of Fame. You might remember that Bill had received several letters from an advocacy group (a group that included Bud Selig) making an impassioned case for former Indians third baseman Ken Keltner as a Hall of Famer.
A central part of that case was that Keltner:
1. Had a higher lifetime average than Hall of Famer Eddie Mathews.
2. Had more RBIs than Hall of Famer Jackie Robinson.
3. Had more hits than Hall of Famer Ralph Kiner.
If you look now -- with the benefit of years and years of such pleasant-sounding Hall of Fame arguments -- the absurdity smacks you right in the face. Mathews, of course, made it into the Hall of Fame DESPITE his .271 batting average, not because of it. You could argue that low batting average kept him out of the Hall for his first four years of eligibility, a blight on the writers’ Hall of Fame voting records. Robinson was elected DESPITE his mere 734 RBIs (and it's still striking to see that he receives 77.5 percent of the vote). Ralph Kiner made it DESPITE a rather paltry looking 1,451 hits -- and it took him the full 15 years to get elected.
It's clear that Kevin McReynolds should not be in the Hall of Fame because he was better defensively than Hall of Famer Ted Williams or, to take it up another notch, that Rick Rhoden should not be in the Hall of Fame because he was a much better hitter than Sandy Koufax … well, that's basically the argument being made with Keltner.
But the larger point is that nobody in the Keltner group really believes the Keltner Standard should be applied for the Hall of Fame. There are 234 players currently not in the Hall of Fame who meet the Keltner Standard, and 90 percent at least are not even viewed as viable Hall of Fame candidates. They are perfectly fine players, but that's about it. The list includes Willie Montanez (whose lifetime WAR is 0.0). The list includes B.J. Surhoff and Richie Hebner and Wally Joyner and Rico Carty and so on.
But I'm sure the Committee to Elect Ken Keltner would tell you they would not want that standard to be used in other cases. They would tell you that Keltner is different because he was a great defensive player (he was), because he was a hero for the 1948 World Champion Indians (he was) and because he was a fine gentleman who represented the game well (he was and did).
The thing is … a fan of any player can come up with such exceptions if they choose. The point is not the exceptions. The point is that because the Ken Keltner group began with their conclusion -- that Keltner belongs in the Hall of Fame -- the rest was amplifying numbers that backed that argument, conveniently passing on the ones that didn't, playing hide and seek with a fair argument.
And while the Ken Keltner example seems kind of obvious and transparent, Tango believes -- and I agree -- that this is the sort of inconsistent analysis we often get in. I’ll quote him:
“A fan will sometimes want to consider clutch and other times not. He will give bonus points for making the playoffs, but not always, and maybe not to the same degree each year. Maybe he thinks a low BABIP should be rewarded for Verlander, and maybe a high BABIP should be dismissed in some cases.”
I think that’s right. The MVP races are a great example. My editor often gripes about how, when it comes to the MVP races, many people will choose from the enormous menu of options to get to the player they wanted in the first place. If the player of choice plays for a winning team, then that becomes important -- how can you be “valuable” for a losing team? If he doesn’t play for a winning team, however, then the argument becomes that a player should not be penalized for having lousy teammates. If he has a high on-base percentage, then that’s important. But if he doesn’t then perhaps his job is not to get on base but to drive in runs. Leadership is wielded often and usually without anything more concrete than vague stories about helping out a teammate or giving a speech. Same goes for clutch performance and fielding performance (sometimes with numbers, sometimes without). Obviously not EVERYBODY works this sort of conclusion first analysis, but many do … at least to some extent.
So if I’m understanding Tango, he’s not so concerned about the FRAMEWORK people use. It can be WAR, VORP, Win Shares, whatever. But he thinks once you choose that framework you have to be consistent about implementation … no matter the results. That -- and that alone -- tells you the value of the framework. If you believe RBIs trump all for instance, that’s fine, but then you have to live with the fact that in 1957, the year Mickey Mantle won the MVP, five players were better than him because they drove in more runs -- this would include Vic Wertz and Frank Malzone. Pick a framework … any framework you want. But live with the results.
Which leads us to the Tango challenge … and Jack Morris … and Rick Reuschel.
You probably know -- it’s been mentioned here a time or two -- that Baseball Reference’s WAR says Rick Reuschel was a better pitcher than Jack Morris. WAR is not so precise that you can feel supremely confident that a player with a 5.4 WAR season was definitely better than a player with a 5.1 season.
But with Reuschel and Morris, um, it’s not that close.
Rick Reuschel: 64.6
Jack Morris: 39.3
No, that’s not close. Of course, you might then argue that Morris had better individual seasons than Reuschel. WAR says no soup for you.*
*WAR prefers outdated "Seinfeld" references. WAR is like that.
Best seasons by WAR
1. Reuschel, 1977, 9.2
2. Reuschel, 1985, 6.0
3. Morris, 1979, 5.6
(tie) Reuschel, 1973, 5.6
5. Reuschel, 1979, 5.5
6. Reuschel, 1980, 5.4
7. Reuschel, 1978, 5.2
8. Morris, 1986, 4.8
(tie) Morris, 1987, 4.8
10. Morris, 1986, 4.6
So six of the top seven WAR seasons belong to Reuschel, including the top two.
Of course, not many people outside the statistical world think Reuschel was even close to as good a pitcher as Jack Morris (though it is noted that at this moment, he leads the poll on this blog). In 1997, Reuschel got exactly two Hall of Fame votes. Morris, meanwhile, is on the brink of being elected. People talk about Morris quite a lot; nobody seems to talk about Reuschel (except, perhaps, in reference to Morris). Some of this is fairly easy to explain -- these are those famous “exceptions” that the Ken Keltner crowd used. Morris pitched for winning teams. He was Opening Day starter a lot. He pitched one of the most famous games in baseball history. He had a certain bulldog presence that Reuschel lacked.
But none of these are particularly convincing arguments that Morris WAS better than Reuschel, either at his peak or over a whole career.
So the Tango Challenge -- come up with a framework that shows that Jack Morris was better than Rick Reuschel. On the bright side, you don’t have to defend the framework itself -- that is to say, if you want to say Morris is better because he won more games (254-214) then that can be your framework.
On the downside, however, you would have then concede that every pitcher who won more games than Jack Morris was a better pitcher than him and is more deserving than Morris for the Hall of Fame. This would include Jim Kaat and Tommy John and Jamie Moyer. Maybe you could live with this. However, if you are really going to use career wins as your framework, you would also have to live with Bob Forsch (168) being better than Sandy Koufax (165) and Jerry Reuss (220) being better than Pedro Martinez (219).
You could go with winning percentage -- Morris’ .563 was quite a bit better than Reuschel’s .528. But again, you would be opening the door. Dwight Gooden had a much higher winning percentage than Morris, and so did Bob Welch, David Wells, Dave McNally and Wes Ferrell, who also hit a lot better than Morris.
No, these are obviously flawed methods. OK, how about this? Morris won 20 three times, Reuschel only once. But again, there’s Wes Ferrell with six 20-win seasons. Wilbur Wood, Mike Cuellar, Luis Tiant, Dave McNally, Johnny Sain all won 20 more often than Morris … no, if you’re trying to make Morris’ case you don’t want to measure by 20-win seasons. Strikeouts will lead down the same path.
You could go with Morris’ amazing Game 7 as the critical factor. But then you would have to think hard about Don Larsen as a Hall of Famer, and Jim Lonborg, and Mickey Lolich and Allie Reynolds and so on.
So you look elsewhere. The advanced stats seem to be conspiring. Win Shares? No. Morris had 225 Win Shares; Reuschel had 240. FanGraphs? Definitely not. Reuschel had 73.4 FanGraphs Win Shares, Morris only 56.9. How about Baseball Prospectus’ VORP? It’s close but … no. Reuschel 35.1 WARP (Wins Above Replacement Player) and Morris 33.3.
So now what? That’s your challenge. If you believe Jack Morris was better than Rick Reuschel -- ESPECIALLY if you believe this should not even be a discussion -- then throw out a framework that proves it. And then, let’s test it.