In case you had not heard, yesterday the Milwaukee Brewers released struggling southpaw Randy Wolf. Yesterday, Jaymes noted the news, and effectively summarized Wolf’s legacy in Milwaukee while looking forward:
Thus, looking back, was the Wolf signing a good one? Probably not. But it’s not an all-time bad decision, either. Wolf isn’t going to go down between Suppan and Jeffrey Hammonds in Brewers infamy. He was paid based on past production, and like every other pitcher signed as a free agent in baseball, underperformed his next contract. He provided the Brewers 200 innings a year as a solid #4 pitcher at a time when they had no one in the minors capable of filling that role. Now, the Brewers do have the arms to fill those spots, and with no postseason aspirations to finish the year, they won’t be pressured into having a “veteran presence” in their rotation. They can give Wolf a chance to latch on somewhere else — maybe even a contender, who knows — and since it’s still August, he’d still be eligible for a postseason roster.
There is something about pitchers like Wolf. So long as MLB teams insist on working starting pitchers into rotations with five spots, there will be a need for warm bodies that can fill out those rotations. Of course, this is under-selling Wolf at his best; at times, these veterans can put together strong campaigns. In Wolf’s case, his 2011 performance was one significant reason the Brewers were able to maintain a consistent rotation, one-through-five. Yet, as good as Wolf was in 2011 — by my run prevented ranking at Bernie’s Crew, Wolf was the 27th best pitcher in 2011 NL — his 2012 performance fell notably short. Even in a league/park environment of approximately 4.52 runs/9 IP, Wolf’s 2012 tally of 94 runs allowed was approximately 23 runs below average.
Unfortunately, Wolf’s 2012 combination of 5.95 runs average / 5.69 IP/G is extreme enough to wipe out his moderate successes in 2010 and 2011; in 2010, Wolf helped Yovani Gallardo anchor a poor rotation with a performance fit for a moderate/low #3 starter (#48 in 2010 NL), and in 2011, Wolf joined Gallardo and Shaun Marcum as one of the team’s three best starters. Wolf’s 2010 and 2011 campaigns looked good for the money — a 10+ year veteran that works 428 average innings over two seasons is absolutely worth $18.75 million. But, as Jaymes suggests, even if Wolf’s contract was not an all-time bad deal in Brewers’ history, 142.3 extremely below average IP can immediately overtake the good years in terms of value; 23 runs below average for $9.5 million is the type of gamble a team makes when they sign an aging veteran, even if 428 IP / 202 R accompany that one bad season.
Greinke / Wolf
In March, I wrote a poorly titled article, “Randy Wolf is a Better Pitcher than Zack Greinke.” In the article, I explored the bounds of Fielding Independent Pitching stats, looking specifically at how pitchers are able to employ their fielders to prevent hits. I asked the question, of course, because although Zack Greinke posted peripherals that were extremely better than Wolf’s in 2011, Wolf ultimately outperformed the young ace-in-waiting. While in passing, we’re frequently tempted to say, “well, Wolf was simply luckier than Greinke,” I wanted to look deeper into those pitchers’ historical performances with their fielders. It turns out that Wolf, in his career, has typically used his fielders to help outperform his K/BB/HR shortcomings, while Greinke has typically underperformed his expected fielding efficiency.
Ultimately, I asked the following question: Do pitchers that learn to pitch in front of efficient defenses learn how to limit hits? I asked this question because Wolf came of age in front of relatively reliable fielders with the Philadelphia Phillies, while Greinke cut his teeth in front of inconsistent Kansas City Royals fielders. Was there some element of trust that Wolf was able to establish with his fielders? If he was able to learn that a batted ball in play need not be an uncontrolled lightning bolt of misfortune, was he able to learn how to limit the damage?
I gather that there are a whole host of pitchers similar to Wolf: there are MLB pitchers with value beyond FIP. Or rather, with valuable traits that FIP does not capture.
On the whole, Fielding Independent Pitching ratios are extremely important. The rates at which a pitcher strikes batters out, allows walks, and surrenders home runs will in turn frame their percentage of batted balls in play. Furthermore, there are elements of FIP that are simply valuable on their own. For instance, pitchers with above average strike out rates have a higher ceiling for performance regressions as their careers advance; pitchers with average-or-worse strike out rates have to walk the tight-rope of limiting the damage. For obvious reasons, too, limiting home runs and walks has valuable consequences for pitchers.
During his first two years in Milwaukee, Randy Wolf provided strong examples of how a pitcher can limit the damage. For instance, even though he only struck out 142 of 936 batters faced in 2010, he maintained relatively average home run and walk rates. By walking 87 batters and allowing 29 home runs, Wolf did not provide peripherals that were great by any stretch of the imagination, but he limited walks and home runs to acceptable rates. His 2011 campaign accelerated these trends; even though his strike out rate declined further, to 134 K in 903 BF, Wolf walked 66 and limited his gopher ball count to 23. These walk and home run totals helped Wolf to further limit the damage.
FIP can be a misleading statistic in this regard. Pitchers with low strike out rates will be penalized when their ratio between BB, HR, and K is calculated (likewise, pitchers with high strike out rates might gain an undue advantage in terms of expected pitching performance). Even if their BB and HR rates are average (or better), a low K rate will mean an inflated FIP. When that FIP is placed on a runs average or ERA scale, our expectation will be to say, “since this pitcher does not strike out a lot of batters, they will have a low ERA or allow a lot of runs.” FIP, in this way, allows strike outs to overshadow BB and HR. As a result, pitchers like Wolf will frequently appear “lucky” if they have a good season. For example:
2010 NL: 0.81 FIPratio, 3.59 FIPConstant, 4.40 runs average; 2010 Wolf: 1.64 FIPratio, 5.23 expected runs average (4.46 actual)
2011 NL: 0.72 FIPratio, 3.45 FIPConstant, 4.17 runs average; 2011 Wolf: 1.07 FIPratio, 4.53 expected runs average (4.03 actual)
Before park adjustments, comparing Wolf to the 2010 and 2011 National League suggests that Wolf allowed approximately 30 fewer runs than expected. One might suggest that using FIP to predict future performance, we could expect Wolf to eventually regress. In a basic way, this is true; but, that misses the point of Wolf’s value in 2010 and 2011. Wolf’s strike out rates from 2008-to-present pretty much guarantee that his walk and home run rates will be overshadowed. In this regard, we might simply hold a type of fatalism over pitchers like Wolf: their clocks will be ticking, and so long as they strike out batters at below average rates, we will expect their performance to regress. Unfortunately, this will cause us to miss the full value of their performances.
(Wolf and Greinke are extreme examples of this. Among 50 qualified starters in the 2011 NL, Greinke boasted the 12th best WAR, Wolf the 38th best WAR. However, Wolf boasted the 28th best ERA among those qualified 2011 NL starters, Greinke the 34th best ERA. Despite all of its benefits in separating fielding and pitching performances to some degree, FIP can also paint misleading pictures of pitching value. To the extent that one might rely on FIP before ERA to judge pitching value, those judgments might fall short of portraying actual on-field value from time to time. We might use Greinke’s FIP to show that over time we expect him to be a markedly better pitcher than Wolf, and therefore more valuable; but, that says little to nothing about Grienke’s 2011 value).
As for my original article, Greinke certainly gets the last laugh in terms of 2012 pitching performance. In Milwaukee, his 123 IP / 49 R performance was approximately 12-13 runs better than his league/park, which placed him on pace for a Top 20 spot among 2012 NL starting pitchers. However, both Greinke and Wolf allowed more hits on batted balls in play than their expected defensive efficiency suggested:
2012 Wolf: approximately 457 batted balls in play, 147 expected hits, 158 actual hits
2012 Greinke: approximately 342 batted balls in play, 110 expected hits, 113 actual hits
Of course, when a set of fielders allow hits on batted balls in play 32% of the time, the term “defensive efficiency” is highly debatable and out of place. It would be an understatement to say that both Wolf and Greinke pitched in front of an inefficient defenses in Milwaukee.
Even if pitchers can learn how to use their fielders by pitching in front of surehanded fielders during their early seasons, there is no remedy for working in front of an extremely inefficient unit.
For further study, I’d like to present a list of pitchers that outperformed their 2010-2012 FIP by approximately 5%. In the spirit of my analysis of Wolf above, I think we can study these pitchers to determine exactly how pitchers overcome deficient K, BB, or HR rates to improve their runs allowed. This is a quick-and-dirty list, so I used FanGraphs’ FIP and ERA. I then surveyed K/9, BB/9, HR/9, groundball%, and left-on-base% for keys to outperforming FIP.
For instance, a pitcher might be able to outperform their FIP specifically by limiting walks, or perhaps home runs. Even with a low K%, a pitcher might be able to strand runners by allowing a high percentage of groundballs. There are numerous other reasons, I suppose, which I why I present this list for study:
Rodrigo Lopez (Keys: 5.09 K/9 / 2.66 BB/9; 38% GB, 70.2% LOB)
Barry Zito (Keys: 38% GB, 71.1% LOB)
Jeremy Guthrie (Keys: 5.35 K/9 / 2.55 BB/9; 40.9% GB, 70.6% LOB)
Tommy Hunter (Keys: 4.80 K/9 / 1.97 BB/9; 42.8% GB, 73.3% LOB)
Bronson Arroyo (Keys: 5.20 K/9 / 2.06 BB/9; 40.9% GB, 74.9% LOB)
Kyle Kendrick (Keys: 44.3% GB, 71.5% LOB)
Joe Saunders (Keys: 44.0% GB, 71.5% LOB)
Jonathan Sanchez (Keys: 1.03 HR/9, 41.3% GB, 72.9% LOB)
Wade Davis (Keys: 76.2% LOB)
Jeremy Hellickson (Keys: 81.4% LOB)
Ervin Santana (Keys: 6.71 K/9 / 2.96 BB/9, 41.3% GB, 73.9% LOB)
John Lannan (Keys:0.77 HR/9, 53.1% GB, 71.7% LOB)
Mike Leake (Keys: 6.15 K/9 / 2.37 BB/9, 48.3% GB, 72.2% LOB)
Jair Jurrjens (Keys: 40.7% GB, 73.2% LOB)
Jeff Niemann (Keys: 6.95 K/9 / 2.88 BB/9, 45.7% GB, 72.4% LOB)
Carlos Zambrano (Keys: 0.78 HR/9, 44.7% GB, 72.3% LOB)
Jeff Karstens (Keys: 5.53 K/9 / 1.87 BB/9, 43% GB, 73.1% LOB)
Dillon Gee (Keys: 0.95 HR/9, 48.4% GB, 71.5% LOB)
Aaron Harang (Keys: 1.02 HR/9, 73.4% LOB)
Ricky Romero (Keys: 0.90 HR/9, 54.7% GB, 71.7% LOB)
Ted Lilly (Keys: 7.34 K/9 / 2.36 BB/9, 74.5% LOB)
Jason Vargas (Keys: 5.62 K/9 / 2.48 BB/9, 72.7% LOB)
Trevor Cahill (Keys: 0.85 HR/9, 57.5% GB, 73.7% LOB)
Clay Buchholz (Keys: 0.85 HR/9, 49.9% GB, 76.1% LOB)
Jhoulys Chacin (Keys: 0.92 HR/9, 51.5% GB, 72.9% LOB)
Clayton Richard (Keys: 0.85 HR/9, 50.1% GB, 72.1% LOB)
Mark Buehrle (Keys: 4.73 K/9 / 1.95 BB/9, 44.5% GB, 72.4% LOB)
Matt Harrison (Keys: 0.73 HR/9, 48.7% GB, 74.1% LOB)
Brett Myers (Keys: 6.82 K/9 / 2.45 BB/9, 48.6% GB, 73% LOB)
Wandy Rodriguez (Keys: 7.42 K/9 / 2.95 BB/9, 47.2% GB, 72.7% LOB)
Ian Kennedy (Keys: 7.92 K/9 / 2.56 BB/9, 76.6% LOB)
Here is a companion list of really good pitchers that also outperformed their FIP by approximately 5%. Since these are mostly good pitchers anyway, it’s difficult to locate specific reasons for their performance (except to say, “hey, these guys are really good pitchers”).
Matt Garza (Keys: 7.87 K/9 / 2.81 BB/9, 42% GB, 72.9% LOB)
Tommy Hanson (Keys: 8.39 K/9 / 3.06 BB/9, 41% GB, 74.4% LOB)
Tim Hudson (Keys: 5.90 K/9 / 2.63 BB/9, 59.1% GB, 75.3% LOB)
Shaun Marcum (Keys: 7.53 K/9 / 2.37 BB/9, 74.6% LOB)
Ryan Vogelsong (Keys: 0.69 HR/9, 45% GB, 80.1% LOB)
Hiroki Kuroda (Keys: 7.08 K/9 / 2.15 BB/9, 48.5% GB, 75.8% LOB)
C.J. Wilson (Keys: 7.84 K/9, 0.60 HR/9, 49.6% GB, 72.4% LOB)
R.A. Dickey (Keys: 2.19 BB/9, 0.77 HR/9, 51.2% GB, 76.4% LOB)
Johnny Cueto (Keys: 2.46 BB/9, 0.63 HR/9, 47.5% GB, 77.6% LOB)
Jordan Zimmermann (Keys: 1.81 BB/9, 0.85 HR/9, 43% GB, 76.9% LOB)
Gio Gonzalez (Keys: 0.63 HR/9, 47.9% GB, 76.2% LOB)
Cole Hamels (Keys: 2.22 BB/9, 47.2% GB, 80.5% LOB)
Matt Cain (Keys: 2.28 BB/9, 0.71 HR/9, 74.2% LOB)
David Price (Keys: 0.75 HR/9, 46.2% GB, 77.9% LOB)
Jered Weaver (Keys: 0.84 HR/9, 78.9% LOB)
Felix Hernandez (Keys: 50.4% GB, 76% LOB)
Justin Verlander (Keys: 40.5% GB, 75.7% LOB)
Clayton Kershaw (Keys: 42.9% GB, 76.2% LOB)
Of course, in the spirit of my Greinke analysis above, I also thought it’d be a good idea to provide a list of pitchers that underperformed their 2010-2012 FIP by approximately 5%. Again, some of these pitchers are simply really good, and it’s difficult to find particular stats that provide keys to their underperformance. I suppose, in this regard, simply playing in front of a poor defense can unduly impact pitchers. However, some of these pitchers also have average or worse left-on-base percentages, which brings another question: could a pitcher’s ability to effectively repeat his mechanics from the stretch impact his ability to strand runners?
Francisco Liriano (Keys: 3.98 BB/9, 69.2% LOB)
Zack Greinke (Keys: 0.81 HR/9, 68.2% LOB)
Cliff Lee (Keys: 0.84 HR/9)
Jaime Garcia (Keys: 69.9% LOB)
Doug Fister (Keys: 70.3% LOB)
Tim Lincecum (Keys: 3.57 BB/9)
Chad Billingsley (Keys: 3.34 BB/9, 70.9% LOB)
Brandon Morrow (Keys: 3.55 BB/9, 38.6% GB, 68.1% LOB)
Anibal Sanchez (Keys: ????)
Yovani Gallardo (Keys: 3.18 BB/9, 0.96 HR/9)
Daniel Hudson (Keys: 0.84 HR/9, 39.6% GB)
Jake Peavy (Keys: 0.93 HR/9, 38.3% GB, 70.2% LOB)
Jon Lester (Keys: 3.32 BB/9, 0.86 HR/9)
Dan Haren (Keys: 1.09 HR/9)
Scott Baker (Keys: 1.12 HR/9, 35% GB)
Justin Masterson (Keys: 6.86 K/9 / 3.36 BB/9, 69.1% LOB)
Josh Beckett (Keys: 1.15 HR/9, 70.5% LOB)
Rick Porcello (Keys: 0.87 HR/9, 66.9% LOB)
Jason Hammel (Keys: 3.03 BB/9, 0.92 HR/9, 69.2% LOB)
Derek Lowe (Keys: 3.14 BB/9, 68.8% LOB)
Gavin Floyd (Keys: 0.97 HR/9, 70.8% LOB)
John Danks (Keys: 0.91 HR/9, 70% LOB)
Ubaldo Jimenez (Keys: 4.07 BB/9, 0.81 HR/9, 70.9% LOB)
Max Scherzer (Keys: 1.19 HR/9, 39.6% GB)
Ricky Nolasco (Keys: 1.05 HR/9, 68.1% LOB)
Bartolo Colon (Keys: 1.08 HR/9)
Jonathon Niese (Keys: 0.99 HR/9, 70.6% LOB)
James Shields (Keys: 1.16 HR/9)
Chris Volstad (Keys: 5.71 K/9, 1.05 HR/9, 66.6% LOB)
Charlie Morton (Keys: 5.79 K/9 / 3.40 BB/9, 66.4% LOB)
Luke Hochevar (Keys: 1.01 HR/9, 65.6% LOB)
John Lackey (Keys: 3.07 BB/9, 0.91 HR/9, 66.6% LOB)
Chris Narveson (Keys: 3.40 BB/9. 1.06 HR/9, 69.1% LOB)
Jeff Francis (Keys: 1.04 HR/9)
Joe Blanton (Keys: 1.41 HR/9. 68.1% LOB)
Bud Norris (Keys: 3.84 BB/9, 1.16 HR/9, 68.1 LOB%)
Homer Bailey (Keys: 1.14 HR/9)
Nick Blackburn (Keys: 4.10 K/9, 1.48 HR/9, 67.7% LOB)
Brian Matusz (Keys: 3.63 BB/9, 1.48 HR/9, 35.9% GB, 67.8% LOB)
Jason Marquis (Keys: 3.08 BB/9, 1.22 HR/9, 66.8% LOB)
Jake Arrieta (Keys: 3.87 BB/9, 1.21 HR/9, 67.2% LOB)
Kevin Correia (Keys: 1.26 HR/9, 68.3% LOB)
Josh Tomlin (Keys: 4.95 K/9, 1.37 HR/9, 37.3% GB, 66.5% LOB)
Edinson Volquez (Keys: 5.32 BB/9, 1.07 HR/9)
A.J. Burnett (Keys: 3.56 BB/9, 1.23 HR/9)
Randy Wells (Keys: 3.37 BB/9, 1.08 HR/9)
J.A. Happ (Keys: 4.27 BB/9, 1.16 HR/9, 38.4% GB, 69.7% LOB)
Roberto Hernandez (Keys: 5.15 K/9, 66.1% LOB)
Kevin Millwood (Keys: 1.15 HR/9, 69.8% LOB)
Ultimately, I think there are a lot of different projects we can undertake to determine overall pitching value and to test the limits of Fielding Independent statistics. Insofar as pitching always occurs in an environment that includes park factors, specific fielders, etc., we will always encounter elements that obscure our understanding of value. Baseball is not only great because the game on the field is immensely quantifiable and accessible to our minds, but also because the game also includes inexplicable and improbable performances, too.
Baseball-Reference. Sports Reference, LLC., 2000-2012.
JSOnline. Journal Sentinel Inc., 2012.