Tuesday, October 04, 2011

The continuing story of Justin Verlander, magical pitcher

Justin Verlander and narratives -- in that link, Joe Posnanski discusses how sportswriters and commentators are clinging to the story that Justin Verlander was the reason the Tigers won last night, despite all evidence being markedly to the contrary.  It's a great read, it's spot on, and it has one problem and one problem alone. Alas, that problem is that Posnanski himself is guilty of perpetuating a narrative that numbers didn't necessarily back up with Verlander this season. See, for instance, two ways to look at MVP voting and the obvious MVP choice (he rightly concludes there isn't one)

While Posnanski is willing to acknowledge that Verlander is not a guarantee for MVP, he does seem to buy into the gospel that Verlander is a uniquely dominant pitching force in the American League. In fact, numbers indicate that he's a superb pitcher whose difference from other great pitchers has been exceptional fortune.  To the extent Verlander has a defining feature that makes him the true unrivaled leader in AL pitching, it's been his ability to pitch deeper into games than his rivals (which may be a function of how atrocious the Tigers' other starters were for most of the season, forcing Leyland to leave Verlander in to give his bullpen a rest). 

Verlander has been a very good pitcher, but for nearly the entire season, he was only marginally separated from any others in Fielding Independent Pitching. Jered Weaver and Justin Masterson were both his equal at the start of September. Masterson then got rocked, courtesy of a grand slam from the Tigers that destroyed his numbers (FIP doesn't take kindly to grand slams); Weaver is still his equal. Verlander benefited from a lot of defensive help (not merely from good defensive range, but through things like OF assists and Alex Avila throwing out baserunners -- both of which he got huge benefits from in the game I saw him pitch (where he was about as unhittable as Mike Pelfrey)). Verlander should win the Cy Young, but he owes his magnificent season to defensive freak plays and a significant aberration in strand rate.  While you can't write off strand rate entirely (obviously, Verlander's ability to strike batters out makes a huge contribution here), it's a significant aberration from the past performances of one Justin Verlander (who, unlike Cliff Lee, did not generate a sudden ability to strike batters out in the 2010-2011 offseason).

A good comparison would be Verlander's 2008 season.  It is, by any account, his worst season in the major leagues. But it's the one that most mimics batters' performance against him this year.  The line drive/ground ball/fly ball percentage is virtually identical. 

But his numbers were dreadful.  Three factors killed him -- 1) a high walk rate -- his walk rate was nearly double what it was in 2011; 2) a BABIP difference (unaffected by Verlander's superior strikeout ability, heavily affected by defense and luck) -- .296 in 2008 (the completely unsustainable .236 in 2011) and 3) a strand rate that was a career low of 65%. 

Of these three factors, I think everyone can acknowledge that the first is heavily influenced by narrative.  Greg Maddux, Tom Glavine, Jamie Moyer...there are plenty of pitchers who were known to live outside the strike zone but get the calls that would turn their games up a notch -- in Maddux and Glavine's case, to legendary status, in Moyer's, to a very impressive career.  Because Verlander is hyped, he will get calls he shouldn't. It's either a fact or a piece of widely-accepted apocrypha that accomplished pitchers will have a more expansive strike zone.  Verlander is in that range now. In 2008, probably not so much.  Given that Verlander posted a career-low walk rate (by a significant margin)

The second is one that's easily to write off entirely as luck, but that's worse than never looking at BABIP in the first place.  BABIP can tell you, at a glance, whether things are likely to be sustainable. But it's risky to read too much into it out of context.  Here, however, the context shows plenty to tell you that there is no reason to believe that his numbers are earned.  His 17.7% line drive rate is only 1% better than 2010, his ground ball rate is down 0.8% from 2010, and his fly ball rate absorbs both reductions -- with infield fly rate increasing by 1% as well.  Thankfully, he's in a stadium where fly balls are things of beauty (though he still got stung by an abnormally high HR/FB%, which is the only reason you can feel any pity for Justin Verlander -- that said, an abnormally high rate for Verlander would still be a phenomenally fortunate James Shields -- who has never dropped below 9.8%).  But there's simply no explanation for a decrease of .060 in BABIP when the line drive rate drops by 1%.

The strand rate is mostly Verlander's doing.  There's only 5 starts (of his league-leading 34) where he left a game after retiring one or two batters in an inning.  (There's likely a game or two in that mix where he put a runner on and then was yanked, but too lazy to check 34 different games).  The other factors likely contribute, but it's hard to exclude them.  It's unlikely he's going to combined with his relievers to strand 80% again next season, since I can find nearly no one that fortunate (with the awesome exception of Bob Wickman's 2005 season where he managed a 92.5% strand rate -- god bless small sample sizes and pitch-to-contact closers).  Roy Halladay, a pitcher who you still have to rank higher than Verlander career-wise, has managed a 73% strand rate. A pitcher who compares to Verlander in terms of strikeouts -- Tim Lincecum -- even with the benefit of facing pitchers in about 8-9% of at bats against, has never hit 80% and has a career percentage of 76% (markedly better than Verlander's). Kerry Wood (again, mostly in the NL) only once hit the 80% plateau and has an average of about 76%. 

I agree that Verlander is unrivaled for the Cy Young award this year.  But when the narrative is one of a pitcher truly unparalleled in his dominance by any measure -- the numbers don't bear it out.  There are a handful of pitchers in the AL who, given his combination of defensive support, a Cy Young caliber reputation (sorry, Justin Masterson, James Shields), run support (sorry again, all other Cy Young contenders other than CC Sabathia*), could have achieved what Verlander did.

*ESPN seems to be about the only data source that will allow you to search by run support, but it only allows for a run support average.  Let's just say on one level that I don't get it -- and on another level that it is nonsense of the highest degree.  Justin Masterson is a great example of someone who is emphatically not going to win the Cy Young and someone who got screwed by a team that didn't score runs -- his record in games where he allowed 2 earned runs or fewer is 7-5.  Read that again.  In the 22 starts where Justin Masterson allowed 2 earned runs or fewer, he won 7 games.

Fangraphs lists Verlander with 144 runs scored in support of him, Justin Masterson with 98, yet ESPN has Masterson with a greater run support average. So I took a broader view, maybe ESPN is just looking at the number of runs scored by a team in the entire game that those pitchers started??? Nope. That would skew the numbers even more. The Indians scored a total of 113 runs in ALL of Masterson's starts.  (In comparison, the Tigers scored 155 in Verlander's starts).  Thus, even setting aside Verlander's 35 more innings pitched, ESPN's run support percentage is just inexplicable. 

The average run support with the fangraphs numbers (RSx9 divided by innings pitched):
Verlander: 5.16
Masterson: 4.08

No comments: