I’m the first to admit I’m not a full-on sabermetrician.
That’s not to say I don’t use my fair share of “modern” statistics when actually evaluating players – sure, I like hearing about homer totals and batting average as much as any fan, but telling me that CC Sabathia is good because he won 21 games last year will get you nowhere. My use of advanced stats, though, typically goes only as far as ERA+ for evaluating pitchers (with its non-advanced complements ERA, WHIP, K/9 and BB/9) and OPS+ for hitters (throwing in OPS, IsoP, IsoD, and XBH/K for good measure). Now, that’s not because I don’t recognize that advanced metrics have their place, or because I don’t understand that they can sometimes quantify things that are hard to see with my eyes. But, generally, I prefer to speak a language we can all appreciate – and advanced stats simply aren’t everybody’s thing.
That all said, in my daily/weekly/monthly perusal of stats, I noticed something developing over the course of the 2010 season that is worth discussing here, particularly as the debate as to who should be in the Cubs’ rotation and whom the Cubs should acquire for that rotation heats up.
So what did I notice?
Randy Wells was performing much better on advanced paper than he seemingly was on the field.
While we were all seeing a guy regress from his 2009 form, with reduced control, a straighter fastball, and short outing after short outing, the advanced stats saw something else. They saw a guy who was, consistently, one of the best pitchers on the 2010 club. Of course, that’s a bit of a tallest midget medal, but still, it was surprising, given his sometimes 5.00+ ERA, and constant failure to escape the early innings unscathed.
The basic statistics backed up what we were seeing: Wells’ ERA ballooned from an impressive 3.05 in 2009 to 4.26 in 2010 (a dominant July – 1.83 ERA in 5 starts – helped curb that final ERA) and his WHIP went from 1.276 to 1.400. His peripherals were a bit of a mixed bag, but certainly didn’t indicate improvement: while his K/9 increased from 5.7 to 6.7 (arguably a bad sign for a control pitcher), his BB/9 and H/9 both jumped in 2010. Unsurprisingly, his ERA+ fell from a truly great 142 in 2009 to a minimally above average 102 in 2010.
Clearly, you would be wise to conclude, Wells had regressed.
Unless you dug into the advanced stats. It turns out, Wells was a little lucky in 2009, and was unlucky in 2010. And in both years, he was pretty much the same pitcher.
One of the best measures of a pitcher’s luck – BABIP (batting average on balls in play) against – supports this theory. In 2009, Wells had a BABIP of .294, which is slightly better (or more lucky) than average (about .300). In 2010, that number rose to a very unlucky .320. What that means is, either Wells was giving up a ton more line drives in 2010 than in 2009, or balls that weren’t finding holes in 2009 were finding holes in 2010. In reality, it was probably a combination of the two, but at a minimum, it suggests that Wells wasn’t as bad as we thought.
Wells’ FIP (fielding independent pitching) tells the same story, as it was 3.88 in 2009, and only slightly higher – 3.93 – in 2010. Some folks swear by FIP, as it attempts to judge a pitcher only on those things he can control, and not the things that are dependent on the fielding behind him; but I haven’t yet been persuaded that it is quite comprehensive enough (it literally considers only home runs, walks, and strikeouts). Still, over a large enough sample size, FIP is a useful tool, among others.
Finally, considering the most in vogue sabermetric stat – Wins Above Replacement, which attempts to calculate how many wins a player was worth over a theoretical replacement level player – Randy Wells was actually better in 2010 than he was in 2009. His 2009 WAR was a solid 3.1, but in 2010 it was an even better 3.3. The difference is mostly explained by his worse luck in 2010, his increased strikeout production in 2010, and the simple fact that he threw 29 more innings in 2010.
In fact, based on WAR, Wells was the 22nd most valuable starting pitcher in the NL last year – ahead of guys like Jaime Garcia (3.2), Tim Hudson (2.7), and Jonathan Sanchez (2.6). What are the chances you’d be turning up your nose at having any of those guys in the 2011 rotation? Wells was second on the Cubs, behind only Ryan Dempster (3.5), and was even more valuable – if you believe WAR – than Carlos Marmol (3.1 – making Marmol, by far, the most valuable reliever in the NL last year – as if you actually needed statistics to tell you that one).
Oh, and as for the two hot names on the Cubs’ target list? Fausto Carmona and Matt Garza? You guessed it – the Cubs were better off with Wells, who out-pitched both Carmona (2.7 WAR) and Garza (1.8 WAR).
So is WAR the end-all-be-all of player evaluation? I guess technically, it’s supposed to be. But no. Is FIP? BABIP? Nah. Statistics are a useful way of quantifying what we see, but they are not a total replacement for what we see. I don’t need to know his FIP to know that John Grabow was garbage in 2010, just like I don’t need to know Randy Wells’ stats to know that he wasn’t quite as effective, on the whole, in 2010 as he was in 2009.
But these advanced stats do tell me that I – and others – have dramatically underestimated Randy Wells’ value in both 2009 and 2010. Through his first two full seasons, Randy Wells has proved to be a very good number three starter, and there is little reason to believe he won’t continue to be that in 2011.