Using Contact Quality to Sort Out the AL Cy Young Mess

The American League Cy Young race is pretty messed up this season. The current WAR leader, while apparently healthy, might throw so few innings in September that he fails to qualify for the ERA title as a result. The pitcher currently ranked second by WAR in the league hasn’t pitched in a month. A third pitcher who, as of July 1, had authored a sub-2.00 ERA and fantastic peripherals — and was probably the favorite for the award — is now an afterthought.

Overall, there are probably eight candidates who deserve to appear on a ballot — and that’s without even considering the credentials of dominant relievers like Edwin Diaz and Blake Treinen. Voters, however, can only choose five names — and, as a result, it is possible that totally defensible ballots will omit the eventual winner (or that a pitcher who would have otherwise won will be omitted from a totally defensible ballot).

As I noted yesterday with regard to the NL’s Cy Young field, this award invites multiple questions about how best to evaluate pitching performance. Unavoidably, one’s choice for Cy Young will depend on how one weighs what a pitcher can and cannot control — and how best to quantify those effects. In this post, I’ll look at various metrics and consider the implications of each regarding luck, defense, and pitcher skill.

Before we get to how contact and defense might be playing a role in voters’ minds, though, let’s look at some fairly standard statistics at FanGraphs.

Jay Jaffe made the case for Chris Sale’s candidacy last week, and that case certainly looks quite strong — or would, if the season ended today. Problem is, Sale might not get too many more opportunities to build said case. The left-hander is scheduled to throw two innings for Boston today and then another three innings on the 16th. If he records those five innings and then, say, another 10 over his final two starters, he won’t qualify for the ERA title and will potentially allow other pitchers the opportunity to catch up in value.

One pitcher unlikely to do any catching up is the one who currently resides just behind Sale on the WAR leaderboard — namely, Trevor Bauer. Bauer hasn’t pitched in a month due to a leg injury and how much time he will pitch over the remaining weeks is very much in the air. Luis Severino was going great through his start on July 1, but has a 4.66 FIP and 6.83 ERA since then without pitching deep into games. Those games in April, May, and June certainly count, but Severino’s poor second half has allowed other pitchers to surpass him.

To get a sense of how other WAR-type metrics characterize the race, see below.

Blake Snell does very well by those measures based largely on runs allowed, such as RA9/WAR hosted here at FanGraphs and Baseball-Reference’s version of WAR, as well, which starts with RA9 and then makes adjustments based on park, opponents, and team defense. That said, his strikeout and walk marks lag behind those produced by other candidates and his home-run rate is merely good but not great. His low BABIP could be an indicator that he has allowed weak contact, but it might also mean that he’s been lucky or benefited from good defense.

Conversely, Justin Verlander fares better outside of the run-based metrics, leading the Baseball Prospectus version of WAR (based on DRA) — though the error bars at BPro further indicate what a tossup this race has become — and coming in third in WAR here at FanGraphs. That’s a subject worth exploring further.

To get a sense of the type of contact pitchers are giving up, we will look at a pitcher’s xwOBA, available at Baseball Savant, and compare it to their actual wOBA. These numbers do incorporate strikeouts and walks, so we wouldn’t necessarily expect a big gap between the two numbers overall.

For the pitchers whose names appear above the league-average line in the table here, the cause is likely some combination of bad luck or bad defense. It appears that Verlander, for example, has suffered the most among Cy Young contenders, while Corey Kluber and Snell have received the biggest benefit. Curiously, the combination of Gerrit Cole and Verlander plus also Carlos Carrasco and Kluber are on the same teams — and yet reside on the opposite sides of league average. This could mean that luck, more than team defense, has caused the differences between wOBA and xwOBA — or it could mean that the two pitchers simply received different defensive performances over their roughly two-dozen starts this season.

Before isolating the differences between wOBA and xwOBA, the table below shows generally how the team defense has performed this season for the contending starters.

It would appear that Carrasco, Kluber, Severino, and Snell have likely received some decent defensive play, while the jury is still out regarding the Astros pitchers and Sale. To further test the impact, here is xwOBA and wOBA on batted balls, including homers.

This is pretty close to the same table we observed above, but this allows us to find a bit more separation by removing strikeouts and walks. There’s a roughly 40-point gap between Verlander’s wOBA-xwOBA differential and the league-average figure, further suggesting that he’s been the least lucky pitcher. Snell, who receives a similar benefit in the other direction, has been luckiest by this measure.

Now, let’s eliminate homers and see how defense and luck have impacted the pitcher’s expected results based on balls in play.

Look what happens to Verlander above. That big gap we saw before basically disappears. The issue for Verlander isn’t defense, it’s that he has given up a decent amount of cheap homers. Verlander is tied for the MLB in homers travelling under 360 feet, with five, while 20 of his 27 homers have failed to travel past 400 feet, well above the MLB average of around 50%. Meanwhile, it looks like Chris Sale hasn’t had great luck on batted balls that have stayed in the yard, with Cole, Kluber, and Snell all reaping the benefits of some extra outs on harder contact.

What a voter should do with this information is a tough question. Should Verlander be given the benefit of the doubt given that most of his homers weren’t no-doubters, or should those homers count against him because they left the park in the first place? Kluber and Snell look really good in the run-prevention department, but they’ve clearly been aided by some combination of good luck and defense. Further still, the innings gaps that are likely to exist between these players complicate matters, as well.

*****

Just as I did yesterday, I’ve included here a rough WAR estimate based on xwOBA, should you choose to desire such a thing.