One of the problems with ranking baseball players is that any stat you choose as the basis for those rankings is going to have some blind spots, biases, or shortcomings that influence the rankings. We saw this problem in the recent arguments about the AL MVP, where camps of writers rallied around either traditional or “advanced” stats to make their respective cases about Miguel Cabrera and Mike Trout. To some extent, the same type of disagreement occurs when pitchers produce great seasons for poor teams; should their poor W-L record be held against them when their other statistics are exceptional? From any perspective, the stats chosen to rank baseball players will require some assumptions about what is valuable to the ballgame, and how players should be measured according to those standards of value.
Revisiting 2011 Pitching Rankings
In 2011, Zack Greinke pitched his first season in the National League, working for the Division Champion Milwaukee Brewers. After an injury sidelined Greinke during the early season, he returned to strike out 201 batters in 171.7 IP, walking 45 and allowing 19 home runs. Despite this performance, Greinke allowed 82 runs, mimicking his fortunes with the 2010 Kansas City Royals. That season, Greinke worked 220 innings, and despite striking out 181 against 55 walks and 18 homers, the righty allowed 114 runs alongside the Royals’ defense. Over just two seasons, Greinke allowed 45 more runs than his fielding independent performance suggested. Adding in his exceptional 2009 campaign, and Greinke’s three-year forecast suggested that one might expect greatness for this starter.
Greinke was only one of 20 National League starters that deviated from the fielding independent performance of their ranking spot in 2011. Ranked #40 in the NL, Greinke’s runs allowed made him a card-carrying middle rotation starter, but those 16 #3 starters averaged a FIPratio of 0.78, a thoroughly average ratio between K/BB/HR that resulted in an expected runs average of 4.23 in 2011; Greinke’s -0.28 FIPratio was indeed from another planet, suggesting an otherworldly runs average of 3.17. Even if Greinke ranked as a middle rotation starter in 2011, there were key elements in his performance than suggested otherwise.
The trouble with ranking pitchers by either runs average or Field Independent Pitching is that there will be deviations either way. In 2011, 71 NL starters worked more than 100 innings; 20 of those pitchers posted Fielding Independent Pitching ratios that deviated from their actual ranking. Basically, that fact undermines a ranking system; 28% of the ranked pitchers were arguably placed in the wrong spot. And yet, if we ranked those 71 starters by their Fielding Independent Performance, we’d have deviations heading in the other direction.
On November 3, 2011, I published “Notes on Rankings” on JSOnline, part of my series of 2011 NL SP rankings. I like what I wrote there on deviations, so I’ll cite myself here (bad form, I know. It’ll be over soon, I promise):
As I noted yesterday, judging by three-year FIP, there are twenty 2011 NL starters that deviate from their 2011 ranking spots over the course of multiple years (2009-2011). This can be due to basic ineffectiveness, breaking out after a couple of seasons of limited use, and bad luck; or, it can be due to basic pitching style (such as low K% pitchers, or extremely high K% pitchers).
Johnny Cueto improved on his previous performance level. Mat Latos, Ubaldo Jimenez, Chad Billingsley, and Chris Carpenter regressed from previous performances. Jhoulys Chacin, Ian Kennedy, Madison Bumgarner, Mike Leake, Dustin Moseley, Brandon Beachy, Jaime Garcia, and Dillon Gee each compiled 50% or more of their 2009-2011 IP total in 2011, suggesting that their three-year FIP is not necessarily indicative of their overall ability, and was compiled in limited playing time.
Tim Lincecum and Zack Greinke have FIP that are almost too good to actually match in any regard; certainly, Greinke received his fair share of defensive inefficiencies during his starts, but at some level, his strike-out ratio is so high that it ceases to be indicative of a fielding-independent performance that he can actually maintain in terms of runs allowed. Ricky Nolasco fits that description to a lesser extent; he is one of the poster-children of great FIP pitchers that never materialized on the diamond.
On the other end of the spectrum, it is arguable that Jeff Karstens, Joe Saunders, and, to a lesser extent, Chris Volstad produce strikeout ratios that are too low to produce a FIP that adequately reflects their on-field performance. Edinson Volquez, at the low end of the FIP spectrum, simply allowed so many walks and home runs that his FIP wholly deviated from his overall ranking.
These are the twenty pitchers with three-year FIP performances that deviate from their 2011 ranking spot. This simply means that their ranking is not necessarily indicative of their overall ability, either due to lack of previous playing time, career improvement or career decline, or outliers of the statistic (JSOnline.com, Journal Sentinel, Inc., 2012).
The fun part of the off-season is that we get to completely nerd-out and work on issues like this. So, I ask a simple question: “How did those 20 pitchers with FIP that deviated from their 2011 ranking perform in 2012?” As though the baseball gods expected this question, they blessed us with a hilariously even spread of results; nine of those starters posted runs averages that were more valuable in 2012 than 2011, nine of those starters were less valuable in 2012, and two of those starters either worked in the American League (Ubaldo Jimenez) or only served as an emergency starter in 2012 (Dustin Moseley). Interestingly enough, 14 of those 20 starters worked fewer innings in 2012, including notable injuries to Chris Carpenter, Brandon Beachy, Jaime Garcia, Jhoulys Chacin, Chad Billingsley, Dillon Gee, and Jeff Karstens.
Not surprisingly, the pitchers that ranked as deviations from middle-to-low rotation spots in 2011 improved in 2012, whereas the top-ranked deviations were worse in 2012. In some regard, this suggests that these rankings were corrected from 2011 performances to 2012 performances (although it’s still arguable whether these performances convey these pitchers’ true value). One major exception is Chris Volstad, who was a low-rotation starter in 2011, and less valuable in 2012 after being traded to Chicago. The other major exception is Johnny Cueto, who posted a great-but-injury-shortened season for the Reds in 2011, only to improve to the top spot in the 2012 National League. Tim Lincecum completely flipped sides of the rotation, and his 2012 performance will undoubtedly be a deviation from his FIP once again.
There are a lot of interesting stories throughout this list. Mat Latos underperformed in PetCo Park last year, but a move to Great American Ballpark didn’t stop him from improving his value. Even in injury-shortened seasons, Garcia, Billingsley, and Beachy were much more valuable in 2012 than in 2011 (in terms of basic runs prevented). Greinke and Nolasco finally improved their runs allowed value against their FIP, although Nolasco remains the poster-child for under-performing Fielding Independent Pitching stats — Nolasco has allowed 77 more runs than his FIP projects from 2009-2012. Even Madison Bumgarner pitched an average season, making his five run regression is misleading.
There will be more deviations in the 2012 rankings, I am sure of it, and we’ll have more questions about projected pitching value for 2013. If anything, looking at a pitcher’s runs allowed, as well as his Fielding Independent Pitching performance, gives us the ability to think about the different ways in which pitchers provide value for their club.