The WAR over Brandon McCarthy

There are people who have issues with advanced stats and at times it can be understandable. Rickey Henderson stole 1,406 bases. He actually stole them, it wasn't derived from some formula. It can't be reassessed and determined that it really should be 1382.6, it is 1,406 and should he enter another major league game and swipe another it will go up accordingly to 1,407. While these benchmarks have their role and are certainly very important (maybe the most important) statistics, there are also the advanced stats, the stats that seek to go beyond what we just see with our eyes and tally up, and seek to find true value, true worth. Stats that give us the opportunity to say what would it be like if Babe Ruth played in 2011's American League? Or that answer questions like, everyone is pitching better this year, is that one pitcher really that much better?

In the end the object of a baseball game is still the same - had Coco Crisp stolen four bases yesterday in a losing effort, who really cares right? Hideki Matsui's 14-game hitting streak quickly becomes irrelevant if coupled with a 14-game losing streak doesn't it? What matters most is winning. We are all familiar with a statistic that answers that simple question, how much did x-player contribute to winning the game - wins above replacement player (WAR). All of us for the most part are also familiar with how there are two calculations for it, one care of Baseball-Reference.com and one from Fangraphs.com. Therefore, unlike looking up Rickey Henderson's stolen bases on either site and seeing that he stole 1,406 according to both databases, with WAR you will get a differentiation, because like all fans they value certain contributions to the game differently. Seldom is there a big difference in the two systems, but occasionally due to their construct you can have big questions over the value of one player versus the other. For the A's this year, Brandon McCarthy is that player.

I have been a huge McCarthy fan all year, I think he is doing great things and has pitched well throughout the year - Fangraphs agrees, they say he has been worth 3.1 wins more than whomever we have starting in Sacramento. Baseball-Reference however says, not so fast, McCarthy has been alright but he has only been worth 2.3 wins better than some scrub for the River Cats. When I first noticed this split, it was July 26th when McCarthy was taking on the Rays at the Coliseum. Then, Fangraphs as has been the case for most of the season and still is today, had McCarthy as the most valuable player on the team, yet Baseball-Reference had him as not only not the team's best pitcher, but their fifth best at that. Today that discrepancy is no longer as wide, but still exists, as Baseball-Reference still views Gio Gonzalez and Trevor Cahill as more valuable than McCarthy. So what digs? Who is right?

The problem lies because the big question in calculating advanced pitching stats, is where does the pitching end and the defense begin? Is it McCarthy's fault if someone scorches a groundball between Cliff Pennington and Scott Sizemore for a single? Or is it Pennington and Sizemore's fault for not getting to it? Fangraphs simply says, I am not going to answer that question, there are merely three things pitchers control, strikeouts, walks and home runs (now of course nothing could be clear cut right, what if Coco or David DeJesus scales the wall and steals back a home run, surely that isn't the pitcher "keeping it in the yard" well, Fangraphs says it is). Baseball-Reference chooses to instead look at the end results, cold-hard runs. That means pitchers who have strong peripherals like McCarthy (1.3 BB/9 and 0.4 HR/9) find their FIP (Fielding Independent Pitching) lower and in some cases much lower than their ERA (for McCarthy he sports a 3.31 ERA but a 2.69 FIP), which Fangraphs values much more highly than the runs that actually crossed the plate which is what Baseball-Reference is monitoring.

Let's look at the three pitchers in question: 

One thing to note is that WAR does favor those who play more, just like a counting stat in that respect, it is easier to accumulate WAR by playing more, therefore Cahill despite worse peripherals and more runs than McCarthy does benefit from having pitched almost 40 more innings (this is true of both calculations).

Where does this leave us? It is interesting to see how different pitchers are helped or penalized. Cahill is helped because his 69 runs, while higher than the others, and resulting in a higher ERA, are 26 fewer than Cahill's AAA replacement would provide.Fangraphs however knocks him for not helping himself out with a high walk rate and higher home run allowed rate which erodes his value as their WAR number is based off the defense independent FIP.

Which is right? Unlike Henderson's stolen bases, there is room for argument. I personally, like Fangraphs' method of taking out what I see as things beyond the pitcher's control. I think there is validity to the argument that once a ball leaves the batter's bat if it is somewhere in the field of play, it is no longer up to the pitcher what happens at that point (you could make the argument that on bunt hits, it is very much so up to the pitcher - a defender - as to what comes of it, but even traditional stats don't fault the pitchers on errors, giving him an unearned run on runners who get on base care of his mistakes, despite the fact that he is wholly at fault). In support of defense-independent pitching statistics was an interesting post the other day which said, if pitchers really have no control over the batted non-home-run ball, there should be little variance regardless of the situation. The post, by Jeff Sullivan at Baseball Nation, looked into those rare times when position players take the mound. He found this when looking at the 126 position players since 1970 who stepped up that hill to stare down MLB hitters,