Thursday, January 13, 2011

Can Beckett Bounce Back?

In an article released earlier today by Jon Tomase of the Boston Herald, he discussed Josh Beckett’s chances of returning to form.  If you’re a Red Sox fan, you’re probably not going to like his conclusion.

“Virtually every pitcher to struggle like Beckett did last season not only was never the same, but almost to a man failed to produce a single, solitary above-average season thereafter.

The results are actually kind of staggering. And while they have no bearing on whether Beckett will bounce back in 2011, they should at least give Red Sox fans pause following last year’s 6-6, 5.78 ERA disaster that can only partially be chalked up to injuries.

So what are we talking about? Thanks to the magic of, it was easy to sort for pitchers in their 30s who posted ERAs above 5.75 while pitching at least 125 innings.”

I don’t want to be too critical of Tomase because he actually did some statistical research, and posed a compelling argument.  If you dig below the surface of his argument, you find that it's flawed and full of empty calories.   First of all, the thresholds he chose for age, ERA, and innings pitched (pitcher of at least 30 with an ERA above 5.75 with at least 125 innings pitched) seem rather arbitrary; perhaps even convenient.  Maybe that’s because they are.  I don't think it's a coincidence that Beckett just happened to be a 30 year old pitcher who put up a 5.78 ERA in 128 innings last season.  What would happen if we changed the criteria to pitchers who are at least 29 years old and produce 5.50 ERA in at least 100 innings pitched?  Would there be more successful turnarounds?  We don’t know.  He doesn’t answer that question for us.  Tomase purposely chose those thresholds to help him prove his hypothesis.  It’s the oldest stat trick in the book.  Jack Morris supporters have been doing this for years.*

*  Jack Morris supporters often tout that he was the “Winningest pitcher of the 1980s.”  While this is true, what they don’t tell you is that the great pitchers of the 1970s (Seaver, Carlton, and Palmer) were done being productive by 1985, while the great pitchers of the 1990s (Maddux, Clemens, and Johnson) were just getting their started in the middle of the decade.  Morris was lucky in that his prime just happened to coincide during a lull between two great crops of pitchers.  The 1980-1989 time period that was chosen was purely arbitrary.  Why didn’t they choose the 1977-1986 or 1985-1994 decades? Two reasons.  One, it doesn’t have a catchy nickname like the “eighties”.  Two, those decades, while just as viable as the 1980-1989 decade, don’t help Morris’s case.  Therefore, his supporters don’t bring it up.  They’re manipulating the statistics to promote their agenda, while ignoring the statistics that do not.   

My second issue with Tomase’s analysis is with the three pitchers he identifies as comparables:  Jack Morris, David Cone, and Dave Stewart.  Let’s take a quick look at how the four pitchers fared during the seasons in question.

                             Age      IP       ERA       FIP       K/9        K/BB   fWAR
Josh Beckett          30     128       5.78       4.54      8.18        2.58       1.4
Dave Stewart        37     133       5.87       5.49      7.49        1.79        1.0
Jack Morris           38     153       6.19       4.51      6.07        1.58        1.8
David Cone           37     155       6.91       5.44      6.97        1.46        1.1   

The first thing that pops out at me is the difference in age between Beckett and his three comparables.  Morris, Cone, and Stewart were all in their late-30s during the twilight of their careers at the time they produced their “disaster” seasons.  It’s shouldn’t be a surprise they “failed to produce a single, solitary above-average season thereafter.”  Beckett, on the other hand, had just turned 30, entered the season as a pitcher in his prime, and was coming off of a 5 WAR season.  Injuries and bad fortune played a significant role in his performance.  Considering his age and true talent level, it would appear that Beckett is much more likely to bounce back than the three above named pitchers.    

The second thing I see is that Beckett had far better strikeout, walk, and ground ball rates during his “disaster” season than his counterparts.  For those of you whom are unaware, strikeouts, walks, and the ability to induce ground balls are three of the most accurate indicators for predicting future pitching performance.  While Beckett’s walk rate was higher than it had been in recent years, it was still slightly better than the league average.  Beckett’s strikeout and whiff rates remained above average and within his statistical norms.  His ground ball rate came in at 45.8%, which is actually above his career norm of 44.7%.  In theory, if he were to repeat his peripheral rates in 2011, he will likely have a much better ERA.

So why did Beckett struggle last season?  Two simple reasons.  The first reason is that Beckett appeared to be incredibly unlucky with regards to home runs allowed.  In 2010, he allowed 20 home runs on 141 fly balls for an HR/FB rate of 14.2%.  While HR/FB rates aren't very predictive, most major league pitchers typically produce a rate between 9-11.5% depending on their home ball park.  If we were to regress Beckett's HR/FB ratio to the league norm of 10.5%, Beckett should have only been expected to give up 15 home runs, not 20.  This poor luck on fly balls, probably added between 8-12 runs to his year end total. 

The second reason for Beckett's struggles was due to extraordinary poor luck on balls in play.  In 2010, he had an obscenely high .349 batting average on balls in play (BABIP) despite allowing line drives only 19% of the time.*  Based on his batted ball statistics, Beckett was expected to give up 120 hits (versus 151 actual) for an expected BABIP (xBABIP) of .296.  31 hits is a huge variance.  As most of you can imagine, these unexpected additional hits are a huge reason why Beckett's ERA ballooned last season.

* The batting average on line drives in play is approximately .720.  The batting average on ground balls is .240.  The batting average on fly balls in play is around .150. 

What caused this variance?  Well, as I mentioned above, luck probably played a significant role, but that wasn't the only culprit.  Poor defense was also a major factor.  The Red Sox were plagued by injuries last season, and most of their best defensive players spent significant portions of the season on the disabled list.  As most of you know, when a team replaces a starting position player with someone from either the bench or the minor leagues, the replacement player tends to be of lower quality on both sides of the ball.  As the FIP/luck dragon video I posted the other day demonstrated, ERA is largely dependent on the quality of fielders behind the pitcher.  As the quality of fielders decreases, the likelihood for additional unexpected hits increases.   This, in turn, results in higher ERAs. 

Judging by Beckett's 4.01 xFIP (which is FIP with an HR/FB rate adjusted for league norms), it appears that he was a much more effective pitcher than his ERA would lead us to believe.  While his injuries certainly didn't help his cause last year, his 5.78 ERA was largely brought upon by poor luck and bad defense.  If Beckett were to repeat his 2010 strikeout, walk, ground ball, home run, and batted ball numbers in 2011, it's very likely that his ERA will bounce back to rate more inline with his 2010 xFIP than his 2010 ERA.  Based on my analysis, Beckett looks like a pretty good bet to return to form in 2011.


  1. @fdog - Thanks man! Take a look at Bill's article from the Platoon Advantage (link under our blog roll). Its pretty outstanding.