So, last night I made my way to the Cell to see whether Trevor Cahill's performance would be “the story,” or if it might be the latest White Sox bullpen conflagration. As it turned out, the game didn't exactly cooperate. Early on, Cahill struggled in his first game after agreeing to a five-year, $30.5 million deal on Monday, while the Sox's bullpen did not combust late, instead aiding an extra-innings win over Oakland, 6-5 in 10.
So, my memes were already shot to hell, but it makes sense to talk about Cahill nevertheless, because there are some general points about evaluating pitchers worth making, especially the perils of taking any of the multitude of well-designed interpretive metrics out there as the final word for evaluating a pitcher's quality.
As fields go, sabermetrics is as guilty as any when it comes to picking its favorites on the basis of who does what, why and how. Mathematics, raised to the level of immutable logic as a matter of faith, winds up becoming less of a language that describes an assembly of events on the diamond in a way that we can easily aggregate and summarize them, and instead becomes the vehicle for absolute pronouncements -- some which of which have the virtue of being true.
There is nothing better than absolute certainty, after all, and so as long as you can hum a few bars, add a mysterious floating head and some flames (both preferably green) and voila! You are the Wizard of Oz, something less than Ozzie Smith when it comes to playing the field, and yet quite determined to be oracular in matters mathematical. It may not matter that you may not be absolutely certain. Instead, you stick to what you know, because you've got numbers to prove it.
Take Cahill, already something of a disappointment for the predictive punditocracy. Cahill upset the analytical applecart from the very moment of his being drafted in 2006, because he was a second-round high school pitcher picked by the so-called “Moneyball” A's -- you know, the team that didn't pick high school pitchers, or didn't pick them early at any rate. Cahill didn't throw especially hard, then or now, usually topping out around 92 or so, sitting around 90, hardly the stuff of Nolan-esque legendry or drool-worthy stuff.
As it turns out, it should have been anyway. In 2009, as nothing more than a 21-year-old rookie, Cahill upset statheads his latest time by posting a 4.63 ERA for the A's. BIS' Component ERA (referred to as ERC) said that was a 4.79 season in its interpretation of his performance, while Baseball Prospectus' SIERA, aimed at anticipating future work, judged Cahill's rookie season as worthy of a 5.08. Per FIP, FanGraphs reports that season was a 5.33 campaign.
That doesn't sound like much to base a future on, of course, except that Cahill followed it up by finishing fourth in the AL in ERA in 2010 as a sophomore, despite a SIERA of 3.90 and an FIP of 4.19. At least ERC had jumped the other way by evaluating his performance as a 2.81 -- instead of over-performing, somebody's metric was finally suggesting that maybe, just maybe, Cahill was actually something like this good.
When a pitcher outperforms his metrics, statheads usually run for the usual suspects, clothing performance with wailing about BABIP, or scurrying toward an observation about his HR/FB percentage. They could do the former, but not the latter, because Cahill's HR/FB rate wasn't better than league-average, inconveniently enough. But at least there are his strikeouts, or the lack of them, a sin with which you can condemn anyone to statistical sub-worthiness. With 4.5 K/9 and an 11.6 percent strikeout rate as a rookie, followed by 5.4 K/9 and 15.1 percent K-rate in 2010, Cahill was below average, perhaps even in Lake Wobegone.
Of course, there's the additional problem that different databases wind up with different classifications of what is or is not counted as a fly ball, but that's one of those snaggy-nasty details that might get people to wondering about how far we can stretch the data we do have to make all-knowing pronouncements.
Get hung up on these facts -- or factoids -- and you run the risk of not being ready to rush to an absolute condemnation or endorsement of the A's decision to give Cahill a five-year, $30.5 million deal. Committing to this extent of erasure, of arbitration cases-to-be that never will be, to faith in Cahill's performance, seems like a huge leap of faith in a pitcher who hasn't punched people out, and who has gotten the benefit of a lower-than-expected BABIP.
The problem with these kinds of broad strokes is that they risk missing the trees for the forest. Cahill is a strike-thrower, yes, and one armed with a hell of a sinker. He's showed a plus curve in the minors, and hasn't really thrown it all that much in the majors (less than 14 percent of the time last year). It would be crazy to get too hung up on how much he has or hasn't thrown that curve this year -- not that some statheads aren't willing to try -- except that pitch selection is defined by the opposing batter's strengths and weaknesses, as well as whether Cahill's trying to execute on those pitches from the stretch or a full windup. But in the rush to judgment over whether the money has been spent badly or well, or if Cahill's going to be something or nothing, it's easy to lose sight of the fact that maybe he's just a 23-year-old talent who has already pitched well and is nevertheless a couple of years removed from what would constitute a normal career peak.
Cahill's mediocre velocity marks and equally tepid strikeout rates do not reflect something he does have going for him already. He got looking strikes 32 percent of the time last year, against a MLB average of 28 percent, an improvement from a league-average rate as a rookie in 2009. Where he generated swinging strikes just 11 percent of the time in 2009 (where 15 percent is average), he ticked that over to 16 percent in 2010. He's been doing that while overwhelmingly relying on a sinker/fastball/change mix, perhaps not all that surprising given his youth.
You want more curves? The kid's 23: Give him time, and count on his club to let him grow as he must, not as it might need -- having already put down more than 30 large on the proposition, it's safe to say the team is betting on his career, and not just his 2011 success rate with breaking stuff.
Cahill pitched last night, of course, and didn't have much to say about his breaking stuff. After the game, he and A's manager Bob Geren talked about execution on fastballs and fastball location. Geren speculated about whether Cahill was having trouble with his grip in the first two innings, when he walked three and allowed an Earl Weaver special -- a three-run homer that put the Sox in the driver's seat. Cahill observed that, early on, he “didn't really know where the ball was going. I was trying to get my fastball outside to righties, inside to lefties,” but “I was leaving everything down the middle.” As for the question of being able to hold onto the ball on a cold spring night in Chicago, “It was a lot harder ... I couldn't really get a hold of the ball.”
Single samples being what they are -- essentially meaningless -- the outcome instead seems to suggest that it would be nuts to jump to some sudden conclusions, about Cahill's curve, or his future. He's beaten expectations consistently and well in his five-odd seasons as a pro, and he might continue to. For statheads, it's important to remember that by getting too stuck on the general truth -- that what is true for the population as a whole is true for everybody, leading to too-quick guesstimates of impending doom -- analysts too can get left holding the bag more than we'd like to admit.
Take Matt Cain of the Giants, for example. My fellow statheads have been burning electrons on the subject of Cain's impending doom for a good four seasons now, and yet Cain keeps managing to stand that proposition on its head and rank among the more effective starters in the league in each subsequent season.
It gets especially ridiculous when you start positing how the Giants' defense must be the answer for Cain being able -- this is the team that has employed ex-catcher Pablo Sandoval at third, DH types like Pat Burrell or Aubrey Huff in the outfield corners and aging leather-less mediocrities like Freddy Sanchez or Miguel Tejada or Edgar Renteria in the middle infield. Naturally, the analysis crew will get it right, eventually: Cain is doomed to go the way of all flesh, the same as you and me, so predict failure long enough and consistently enough, and you're guaranteed to be “right.”
This won't be the first or last time a stathead prefers the security of theory over the virtue of accepting anomalies. As one colleague put it to me more than nine years ago, he “couldn't wait for Livan Hernandez's arm to fall off,” all the better to prove what we “knew” to be true about pitch counts to be immutably so. Livan is still pitching, of course, and if anything has been better than ever, even without reaching six strikeouts per nine since 2004, that at a time when MLB-wide strikeout rates have moved in the opposite direction, past 7.0 K/9.
Livan may yet get to 200 wins, which might seem surprising, but he is at 166 already. As long as there is a need for human beings capable of throwing 30 starts in a season and throwing strikes -- and you can already define that need as permanent -- he might last as the right-handed universe's answer to Jamie Moyer, deathless and dutiful and competent as an innings-eater, if something short of all-powerfully awesome.
To bring this back to Cahill, it's worth keeping in mind that what we think we know can be dangerously misinformational. In his own way, like Cain or Livan, he's beaten expectations and projections. In his age-23 season, he might continue to. The vast preponderance of data suggests that he might not, but against that you have his youth, his limited repertoire and the absence of any knowledge of when and whether he might expand it and how effective he'll be, and I think we've come up with another reason to watch and learn.
If Cahill succeeds where Livan has or Cain has, it's our job to learn from how he beat that “preponderance” of interpretative data, and enjoy the results as baseball fans in the meantime.