Diamondbacks Spring Training Stats: A cautionary tale from last year

I went back and looked at the numbers for our hitters and pitchers in 2014, and compared that with how they performed up until the end of April. I figured any carry forward from spring, would be most obvious in those early games, rather than over the entire season as a whole. And? Nothing. Some played better in spring, some played worse, and yes, some did play in the first month of the season at the level they did in pre-season games. But trying to figure out in advance who will fall into which category? You're a better Nostradm

Wildly over-performing (hitters' division): Martin Prado

Ah, yes, we thought as the Cactus League came to an end in 2014. This is the Martin Prado for whom we traded. He has spent most of the pre-season ripping the cover off the ball. He led the team in spring hits, going 22-for-50, though it was a little weird he didn't have a single walk over that time. Still, good things happen when you put the ball in play, right? Only, once the season started, the gods of BABIP got into mid-season form as well, with Prado's balls in play no longer becoming hits at a .500 clip. Who would have guessed it? Anyone with a basic knowledge of sabermetrics and a faith in regression, I imagine.

Spring training: .440/.451/.620 = 1.071 OPS
Through April 30: .263/.317/.342 = .659 OPS

Wildly under-performing (hitters' division): Paul Goldschmidt

Remember last year when we were concerned about Goldie? Nope, me neither. Which is probably another indication of how objecting the analysis of spring training numbers is. They're more like a Rorschach blot, where what you see in them reveals as much about your own prejudices and biases, than about the subject. But there's good sense in not over-reacting when a good player - in this case, one coming of an MVP runner-up campaign - doesn't appear to be firing on all cylinders during the pre-season. Because the near 1,500 major-league PAs Goldschmidt had to that point are a better indication of true talent than the 61 he had last spring.

Spring training: .241/.311/.370 = .681 OPS
Through April 30: .323/.368/.516 = .884 OPS

Pretty freaking accurate (hitter's division): Mark Trumbo

But something about blind squirrels, and there were a couple of nuts to be found here: Chris Owings' early OPS was within 30 points of his spring tally, and slugger Trumbo came considerably closer than that. Indeed, even things like his K-rate (25%) proved to be not far off what he did in April. That could augur well for this April, since his current OPS (excluding last night's game) was 1.045. However, it's worth pointing out that in 2013, he had a pretty wretched spring, posting a .570 OPS, but that was not reflected in his first month of the regular season, with an .862 OPS over those opening 26 games.

Spring training: .233/.298/.465 = .763 OPS
Through April 30: .210/.264/.506 = .771 OPS

Wildly over-performing (pitcher's division): Josh Collmenter

Not that Josh threw badly last April: indeed, he pitched well enough in relief to get himself moved to the rotation. It's more a case of Collmenter having been untouchable in spring, not allowing a run over his last 14 Cactus League appearances. That's an ongoing streak now covering 25.2 consecutive scoreless innings and dating back more than two years, to March 12, 2013. Perhaps - and I'm just speculating - his funky overhand mechanics allow him to hit the ground running? Whatever the reason, his line over the past three spring trainings has been excellent: four runs over 33.2 innings of work, with a K:BB ratio of 21:3. Yet his career April ERA is 4.60.

Spring training: 12.2 IP, 5 H, 0 R, 0 ER, 0 HR, 2 BB, 7 K, 0.00 ERA
Through April 30: 31 IP, 28 H, 13 R, 13 ER, 4 HR, 9 BB, 21 K, 3.77 ERA

Wildly under-performing (pitcher's division): Brad Ziegler

Unlike Goldschmidt, yes, I do remember concern being expressed about Ziegler's horrendous numbers last spring, where he had a K:BB ratio of 1:6. No, that's not a mis-print. At the time, James described the numbers as "somewhat worrisome," but did point out Ziegler "spent spring trying to improve his slider and changeup." And that's one of the issues with pitching analysis: this is when players will try out new pitches, adjust their mechanics, etc. If something doesn't work - and in this case, whatever Ziegler was tinkering with, clearly didn't - it will be discarded once Opening Day comes around, and normal levels of service will resume. Raw numbers won't tell you this.

Spring training: 6.2 IP, 13 H, 12 R, 12 ER, 1 HR, 6 BB, 1 K, 16.20 ERA
Through April 30: 14. IP, 8 H, 2 R, 2 ER, 0 HR,.7  BB, 11 K, 1.26 ERA

Pretty freaking accurate (pitcher's division): Trevor Cahill

Of course, sometimes a crappy spring is simply the overture to a crappy season, and so it proved in 2014 for Cahill. The problem that would plague Trevor throughout the season was not the stuff, and that was his control, and that was reflected in a spring K-rate very similar to the first month. While the walk-rate was relatively restrained, the hits were much more frequent: all told, the combination WHIP in spring was 1.82 compared to 1.69 in Cahill's early starts, and through the murky spectacles of hindsight, definitely felt like a portent of dark things to come.

Spring training: 22 IP, 34 H, 18 R, 17 ER, 4 HR, 6 BB, 22 K, 6.95 ERA
Through April 30: 26.2 IP, 30 H, 22 R, 21 ER, 3 HR, 15 BB, 29 K, 7.09 ERA

What has this taught us. Some players perform better in spring than at the beginning of the regular season. Some players perform worse. Some perform at about the same level. And there is basically no way to tell who might fall into which category. anyone will fall. The sample sizes we're looking at are small enough that the noise of random variation overwhelms any signal, and that's even if it wasn't for all the additional static, like pitchers trying out new weapons, etc. So, whether you're excited to see players doing well, or concerned they aren't, you're probably better off adopting phlegmatic indifference to all spring training stats.

This has been a public service announcement...