If yer haven't caved yet, to spend the $3 to get Aristotle's baseball stylings, this article "Ain't Gonna Study WAR No More" should do it.
James groans that the argument "WAR = Ultimate Truth" feels to him like the 1970's, when he argued that pitcher W-L records were flawed. People would nod towards his argument and then come up with 9,000 reasons that couldn't possibly be true. This phenomenon re-occurs, with grim reconstructive precision, in the firestorm of 119 comments to his article.
Ain't Gonna Argue WAR No More, dept.
What was so controversial about his article? He proved, tonelessly and with the ruthless efficacy of a T-1000 termination unit, that one type of 2.0 WAR hitter is helping his team 3x as much as another kind. James posited two types of hitters:
|WAR "color type"||Player||Stats||WAR|
|Green||Teddy, 1953||(1.410 OPS in 37 games)||2.0|
|Red||Eddie Bressoud, 1956||(.225 and 23 errors in 133 games)||2.0|
Bill remarked, casually ... obviously Teddy's 2.0 WAR are much more desirable.
To Bill's utter amazement, the WAR crowd argued in 9,000 ways that Teddy's 2.0 WAR are NOT more desirable. WAR is WAR. What the replacements to, in the extra 96 games, have nothing to do with Teddy. Yada yada yada.
So James, blinking his eyes in confusion, dealt with the situation calmly. He refrained from blustering, swearing, sputtering, or calling names, preferring to publish a very simple and airtight study:
|WAR "color type"||PAIRED Player||Stats||WAR|
|Green||225 of 'em||(Great in short seasons)||2.0|
|Red||225 of 'em||(Kinda cruddy in long seasons)||2.0|
And he asked, did the "GREEN" type of 2.0 WAR player appear on teams that won more games? The GREEN 2.0 WAR players appeared on teams that won +4.0 more games on average.
Think about that for a second. How did one roster slot, in these situations, account for four extra wins on average? If Nelson Cruz slugs .575 for fifty games and then does a back dive off the Safeco roof in jubilation ..., would you expect that this would get the Mariners FOUR wins BEYOND what some other player did with the same WAR?
And that's on average ---- > So. Many, many times the 2.0 WAR player netted his team 5-10 wins.
Consistency is the Hobgoblin Of ... dept.
Like we said, 119 posts follow, including posts from three or four of sabermetrics' biggest names -- names I used to respect a lot more before seeing that comments thread.
How many of these 119 posts attend to the fact that the GREEN 2.0 WAR players appeared on winning teams? Exactly 0 of them. These 119 posts get about the business of explaining away the +4 wins as to why they should not be credited to the GREEN players, and why WAR is perfectly intact, meaning this:
- WAR dogmatists: One 2.0 WAR player is exactly like another; they both help their teams win exactly the same.
- James (and Clarke): WAR is not "value." One 2.0 WAR player might help his team win much more than another.
Like, one comment goes, "These 2.0 WAR players are biased toward platoon players, and those +4 win teams tended to have good platoon players to pair with them."
First of all: Okay, let's say that's true. So a 2.0 WAR platoon player helps his team win much more than other 2.0 players do! Where is this reflected on FanGraphs?
And this logic is true for every "excuse" that Studes and Tango make when they argue that WAR is an absolute (Studebaker more so than 'Tango', to be fair). If the GREEN 2.0 WAR players are playing on better teams, they are helping their teams win more than the RED 2.0 WAR players.
Another reply: "If you have a better stat than WAR, show us." This objection has been written on the 'net 1,000,000 times if it's been written twice.
The analogy is: "I think I can take apart an airplane with one tool. My choice is a crescent wrench. Okay, you've shown that the fuselage is intact. What's your tool?"
My tool is this: you shouldn't be trying to use one tool for this job. Period. Why are you trying to limit us to one tool?
WAR is useful; RLP is very useful; UZR is useful. The sleight-of-hand, Three Card Monte slip occurs when you ask, "What BETTER absolute stat do you have?" Yours were never absolute to start with. When did you stop beating your wife?
Replacement Level Players (RLP) Detonation, dept.
There is a hidden light bulb here. Did you catch it? It's in the realm of Replacement Level Players.
Wins Above Replacement (WAR) dictates that if the GREEN players (Teddy 1953) got injured, then they would be replaced by 0.0 WAR players, right? So Teddy's teams should never show an advantage.
You get it? I'm packing 3 WAR in my bat and glove, but if I get injured after 81 games, my 1.5 WAR will be subbed out for Willie Bloomquist and the position will finish at 1.5 WAR.
This didn't happen. The 225 Teddy teams piled an average of 4.0 WAR into the empty spaces, not 0.0 WAR. If you studied another 225 teams, the same thing would happen. Teams would get, on average, more than 0.0 WAR out of a blank slot on their roster. This calls into great question the 0.0 WAR paradigm.
And, in fact, the average baseball team does not win .320 of its games (25 x RLP). It wins .500 of its games, right? It sounds obvious, but this RLP = 0.0 assumption is a key basis to the WAR absolutism. They stand or fall together.
They argued about this, too, for dozens and dozens of posts -- in a state-of-the-art saber thread -- and they argued about everything EXCEPT the real existence of the +4 wins per ballclub.
The 119 comments to James' simple argument were rather melancholy to Dr. D. They were a reminder to him, just how badly a person wants to believe what he believes.
James himself is relatively free from this, which is precisely the reason that reading him is like breathing fresh oxygen. You know what other site is like that? Seattle Sports Insider. I flat-out enjoy exchanging ideas with you mooks. Even the Konspiracy Korners are light, airy, and refreshing.
What's that worth?
Gimme the 100 Ribbies, dept.
Dr. D didn't post this as a Nelson Cruz apology. If a scrub 4th outfielder can win me as many games as the MLB home run leader, give me the benchie. But Dr. D is in line with 30 major league GM's, in that he will pay more for RBI than WAR says he should. Maybe, just maybe, we sabes sitting behind our monitors in mom's basement* have missed something here that the players and managers have understood. Why should that enrage us so? The players must know nothing that we don't?
It's important to know whether a player had 1.0 WAR, or 3.0 WAR, or 40 homers, or 97 RBI, or 6.2 runs per 27 outs. But none of those stats take us to the end of the line. Nelson Cruz could finish with 2.1 WAR this season, but his relative effect be 6-7 wins.
*just kidding of course. I'm one of 'em.
Image: Keith Allison