Is Sonny Gray Really the Best Pitcher in the AL?

Fact: Sonny Gray is currently the AL leader in ERA. He's got a league-leading 2.06 ERA.

He's just edging out former-teammate Scott Kazmir at 2.12 and Kazmir's current teammate Dallas Keuchel at 2.36. Sonny has a legitimate shot to be the A's first Cy Young winner since Barry Zito, and a serious claim as the best pitcher in the AL.

Fact: In his last start, he gave up 4 runs in 5 2/3 innings. He gave up a 3-run homer, but because those runs came after a Sogard error that would have been the final out of the inning, none of them were earned. Despite giving up 4 runs in the game, his ERA actually went down.

While ERA as traditionally been the statistic for evaluating pitchers, it's especially prone to hiccups like this. As someone who tends to rely on advanced statistics, I usually look stats other than ERA to get a sense of who has provided the most value to their team. The most famous of these stats is, of course, WAR.

AL Pitching WAR Leaders 2015 (via Baseball Reference)

Not only does WAR seem to bear out our hypothesis that Sonny is the best pitcher in the AL, it turns it into a blowout. Sonny is leading 2nd-place Keuchel by nearly a full win! ERA tells us Sonny's the best, WAR tells us Sonny's the best, our hearts tell us Sonny's the best, case closed, SONNY'S THE BEST!

... except, not so fast.

Here are the AL WAR leaders according to Fangraphs:

AL Pitching WAR Leaders 2015 (via Fangraphs)

... wait, WHAT?! No typo there. Sonny doesn't appear on the AL fWAR pitching leaderboard until #6, over half a win behind Chris Archer with 3.9 WAR. How are Fangraphs and Baseball-Reference reporting such different numbers?!

It would be easy to look at the WAR leaderboards from Fangraphs and Baseball-Reference and say they're reporting completely different stats, and you wouldn't be entirely wrong. Baseball-Reference and Fangraphs evaluate their pitchers with similar formulas, but differ in one fundamental way: B-Ref uses Runs Allowed/9 innings (RA/9) as baseline for evaluating pitcher performance, while Fangraphs uses FIP.

FIP stands for Fielding Independent Pitching. The theory behind it is that defense plays a huge role in whether a pitcher succeeds or fails: a grounder along the third baseline can just as easily be a double into the left field corner as a simple groundout depending on who's playing 3rd base. FIP attempts to quantify a pitcher's true talent by removing pitching from the equation, counting only those stats that don't involve the defense: walks, strikeouts, and home runs. While FIP obviously doesn't account for things like pitch-framing or "pitching to contact", it actually does a really good job of evaluating pitchers. In fact, FIP is correlated with a pitcher's ERA in the future even better than their current ERA! FIP is an invaluable tool in modern pitcher evaluation because it attempts to determine how many runs a pitcher should have given up with an average defense behind them.

The reason Sonny lags behind the leaders in fWAR is precisely because he sits behind those pitchers in FIP as well. Sonny is #7 in the American League in FIP in 2015, with his 2.93 FIP lagging his 2.06 ERA significantly.

So, who's right here?

That depends. The difference between fWAR and bWAR reflects a fundamental difference in the understanding of how baseball players' worth is evaluated. In fWAR, which is FIP-based, the number of runs a pitcher deserved to give up based on their personal performance is the determinant of how much they're worth. In bWAR, which relies on RA/9, at it's core the most valuable players allowed the fewest number of runs to score (notice, by the way, that whether runs are earned or not doesn't factor into this equation at all). 

At the end of the day, I personally use bWAR for pitchers. It just doesn't make sense to me to give pitchers wins for their FIP rather than their runs allowed. Think about hitters, for example: even if a hitter has had an incredibly luck-driven season based on a sky-high BABIP, we still credit them for the doubles and triples that fell in, even when they were lucky. So too, even when a pitcher gets lucky or unlucky based on their peripheral stats, at the end of the day we still evaluate them on how many runs they ended up giving up, even if it's then used to calculate something more fancy like WAR.

None of this is to say that I think people who use fWAR for pitchers are wrong or dumb. It's just looking at something totally different. FIP looks at the results a player deserved rather than the results a player actually saw on the field. That's an incredibly useful thing to know, and I use FIP all the time. But in my eyes, it's just not the best way to tell you which player was most valuable on the field.

So is Sonny Gray really the best pitcher in the AL? It all depends on how you understand what it means to be the best.