Weak Contact and the American League Cy Young Race

Over in the National League, differing philosophical differences could shape the voting for the Cy Young award. Unless voters choose to embrace a closer like Zach Britton or look at only wins, however, we don’t have the same type of arguments over which to rage in the American League. In the AL, for example, there’s no pitcher with a massive, Kyle Hendricks-like difference in ERA and FIP. There’s no Clayton Kershaw-size innings gap between most of the contenders. Rather, the AL offers a large group of deserving candidates. To decipher which candidate is the most deserving, we’re going to have to split hairs. Let’s start splitting by discussing weak contact and its role in the candidates success.

To determine potential candidates for the Cy Young, just as I did for the National League, I looked at those in the top 10 of both RA/9-WAR as well as the WAR used on this site. If the pitcher appears among both groups, he’s included below. I also included J.A. Happ because he has a lot of pitching wins, and whether you agree or disagree with the value of a pitching win (I honestly had no idea Happ had 20 wins before beginning to write this, if you want to know the value this author places on them), some voters will consider them, so he’s on the list. A few relevant stats, sorted by WAR:

Those top four candidates seem to have the most compelling cases. Of those candidates, only Sale doesn’t appear among the top five of both ERA and FIP, but he also leads the AL in innings pitched this season. Rick Porcello has presented a strong argument for his candidacy in recent weeks, Tanaka leads the league in ERA, and Kluber looks to have best combination between FIP and ERA. There probably isn’t one right way to separate these candidates, but one aspect of the season at which we can choose to take a look is the impact that weak and strong contact has made in turning batted balls into outs.

The first thing we can do is look at the average exit velocity for these pitchers, per Baseball Savant. Below is the average exit velocity and where those players rank among MLB pitchers with at least 190 batted balls (173 pitchers).

While we can’t say exactly how much influence a pitcher has over exit velocity, the answer is at least some. We see here that the only real standout in the group is Kluber, with most of the group residing in the middle and Happ and Sanchez hardly distinguishing themselves. Looking at the average exit velocity by batted-ball type, we see much of the same.

Jose Quintana distinguishes himself with weaker contact through the air, while Porcello has done the same on the ground. While this information is generally illustrative in demonstrating the general contact tendencies of these pitchers, we can drill down further, using the Statcast search function available at Baseball Savant. At that site, it’s possible to separate batted balls by their expected batting average (xBA), which incorporates exit velocity and launch angle, both of which are very important in determining whether a ball becomes a hit and what kind of hit it becomes.

Here, we will separate batted balls into buckets based on their xBA. The easiest balls to turn into outs will show up where xBA is under .200. On those batted balls, and there have been more than 20,000 of them this season, the total batting average has been .064. By looking at how the Cy Young candidates fared on those batted balls, we can see how many extra outs those pitchers gained and lost compared to what we would expect from their contact profile. Here is how the candidates have done on those sure outs.

On those easy outs, Happ and Sale have gained the most from either luck or defense. As I mentioned in the piece on the NL:

Because we have the expected batting average based on the contact profile (exit velocity and launch angle), it seems nearly impossible that these respective gains and losses are the product of some sort of pitcher skill. What we have done is neutralized a portion of the weak contact argument by showing how all pitchers with that same contact profile should perform.

We can keep going by creating more buckets depending on xBA. Here are the buckets I created with their respective league-average BABIP.

Taking these buckets, we can calculate the number of outs a pitcher has gained or lost relative to their contact profiles, as shown below.

Two of the bigger beneficiaries reside in Toronto. How much credit should go to Kevin Pillar is not entirely clear, but given the average exit velocities those two have produced on fly balls, he certainly isn’t hurting matters. Tanaka has received a considerable benefit — as have Porcello and Sale — while Kluber, Verlander, and Quintana have been roughly neutral. Translating those extra outs into an adjusted BABIP, we can see how much effect those outs have had.

J.A. Happ has a lot of wins, but there’s some evidence to suggest that he isn’t completely responsible for them. He’s had a good season, but there is a healthy difference between his ERA and his FIP, and there’s reason to believe he’s not doing anything special to create that difference. In somewhat the same way, Tanaka is definitely deserving of his top-four Cy Young status, but his league-leading ERA has had some help. Adding 19 points to Porcello still gives him a below-average expected BABIP, indicating he has induced some weak contact. How much credit he deserves for that contact is up for interpretation. Perhaps the biggest surprise on this list is Verlander, who has a very low .251 BABIP, but his batted-ball profile this season nearly matches that number. If you are inclined to believe that type of batted-ball contact matters and is the responsibility of the pitcher, you should look more to Verlander’s ERA than his FIP. As for Corey Kluber, he still looks good and there doesn’t appear to be a compelling reason to knock him off favored status.

Author’s Note: In the comments of the post on the Cy Young candidates, several commenters requested that the data be run to include home runs. After all, Scherzer had allowed a lot of home runs, meaning he was allowing hard contact. For the AL, Verlander’s case might be similar. I did not use home runs when looking at the BABIP for this analysis, as I consider FIP my baseline pitching stat. When I look to a pitcher’s quality, and not everyone needs to do this, I generally look at FIP. As home runs are accounted for FIP, I omitted home runs from the BABIP portion of the analysis because I was looking for potential reasons to move away from FIP when considering these candidates. Adding home runs back in muddles that process and does not help in providing any context to the baseline analysis. In any event, your comments are appreciated and considered for future posts.