Is The Cy Young Debate Over?

Over the last few weeks, as CC Sabathia struggled a bit while Justin Verlander remained dominant, Verlander gained plenty of momentum in the race for the A.L. Cy Young award. In fact, Rob Neyer made the following statement this morning:

Unless he really sucks in his next few starts, Justin Verlander’s going to be the unanimous choice in the American League. You know a pitcher’s in good shape when usually reasonable people start talking about the MVP Award.

Neyer is undoubtedly right, as Verlander has the sort of buzz that makes his Cy Young victory all but certain and has pushed him into the MVP conversation as well. The question that I would like to address is whether the race between Verlander and CC Sabathia should be considered over. While there are other viable candidates, I am going to stick to Sabathia and Verlander to try and illustrate how the use of different metrics can impact our view of the race.

The Basics

[table id=50 /]

Looking at the basic metrics makes a strong case for Verlander. He has more wins, fewer losses, a lower ERA, a better strikeout rate, and allows fewer overall baserunners. That said, there are reasons to dig deeper than these numbers, such as relative quality of opposition, ballpark, and the impact of defense.

Isolating and Contextualizing Pitcher Performance

The simplest way to dig deeper is to utilize a metric like FIP, which tries to isolate a pitcher’s performance from that of his teammates by looking primarily at events that are solely within the pitcher’s control (HR, BB, HBP, and K). One common misconception about FIP is that it measures “what should have happened.” This is not accurate, as FIP measure what actually happened in the statistical categories that can be exclusively attributed to the pitcher (Measures such as xFIP and SIERA are the ones that control certain variables to try and estimate what “should have happened” if the pitcher had not had anomalous results in areas such as HR/FB%, which are generally out of their control). Recently, Fangraphs added FIP-, which adjusts FIP for park and league.

[table id=51 /]

These numbers paint a different picture, as they suggest that when considering the things within the pitcher’s control, CC Sabathia has been better than Justin Verlander. Additionally, we might want to look at the relative quality of the opponents each pitcher has faced. As friend of the blog and former TYA writer Stephen Rhoads noted in an email:

Quality of opposition faced – BP has CC facing .754 OPS hitters, good for 34th highest of 136 pitchers with at least 100 IP. Verlander is at .739, good for 123 of 136. That should matter, I would think.

If you stopped here, you might conclude that although the basic numbers favor Verlander, the advanced metrics and some contextual data point to CC Sabathia as the best pitcher in the AL in 2011.

Digging Into What “Should Have” Happened

As I mentioned above, there are a slew of pitching estimators that try and dig deeper than FIP, controlling for elements within a pitcher’s statistical profile that are anomalous and are unconnected to any sort of skill. For an example, see the THT definition for xFIP:

This is an experimental stat that adjusts FIP and “normalizes” the home run component. Research has shown that home runs allowed are pretty much a function of flyballs allowed and home park, so xFIP is based on the average number of home runs allowed per outfield fly. Theoretically, this should be a better predicter of a pitcher’s future ERA.

The logic behind these metrics is that by controlling for these variables, we eliminate the vagaries of luck and distill a pitcher’s performance down to the products of his talent. There is much debate about using such metrics when it comes to All-Star and award voting, as some contend that we should only consider what actually happened when rewarding past performance, and that such metrics should be used solely in a predictive capacity. While I think there is merit to that argument, looking at these metrics can be useful when trying to determine whether one candidate has been exceedingly lucky.

[table id=52 /]

These metrics paint a murky picture but lean towards Verlander, as xFIP slightly prefers Sabathia while SIERA and tERA favor the Tigers’ ace. I also included BABIP (Batting Average on Balls In Play) and xBABIP (expected BABIP based on batted ball types), which suggest that some of the difference between the two players exhibited by the basic metrics such as ERA and WHIP (the advanced metrics do not include balls in play and therefore are immune to this issue) can be attributed to luck.

So, Who Wins?

This is a tough call. Pointing at FIP or WAR and declaring Sabathia the winner is just as foolish as pointing to wins and ERA and crowning Verlander. I am wary about waving away Verlander’s edge in the basic metrics by using BABIP and xBABIP, being that if we are going to delve into “expected performance” we need to look at xFIP, SIERA, and tERA as well, which point to Verlander as the AL’s best pitcher. If I had a vote, I would likely focus primarily on the basic metrics and the first level of advanced metrics I provided above, namely FIP. Seeing how Sabathia only has a slight edge in FIP and the remainder of the metrics suggest that Verlander has been more difficult to hit and score upon, my vote would go to the Detroit Tigers’ ace. However, the key point I am trying to highlight is that there are various angles that an analyst could take here, each providing the opportunity to reach a different conclusion. This is not the slam dunk that some have painted it to be, and CC should have a chance to steal the award by outpitching Verlander in September.

For more on this debate, see Joe Pawlikowski’s excellent post from this afternoon.

7 thoughts on “Is The Cy Young Debate Over?

  1. Great post, Mo. This should be required reading for anyone looking to enhance/expand their understanding of utilizing advanced stats to evaluate pitcher performance.

  2. I agree, for award consideration I think I would lean towards stats that measure what has actually happened instead of what should have. I don’t get at all how the “debate is over” but I guess that’s just Neyer being Neyer.

  3. I am not a fan of using stats that measure what should have happened to compare what did happen. Even if Verlander was more lucky, I think he has the edge.

    Also, I’d be interested to see performance based on opponent OPS, not just an aggregate. It would be great if performance could be weighted on that basis.