Spring Training Report: Stats Don't Matter

I feel compelled to write this, as there is a lot of hand-wringing in A's fandom about how a particular player has been performing. I'm not writing about one player in particular, so much as the idea of what spring training is supposed to be about, and what spring training shouldn't be about.

Spring Training is about getting players back into playing shape and getting them used to playing every day. Some guys, of course, show up in the Best Shape of Their Lives, whereas other guys spend the winter lounging around. In either case, having not played baseball for up to 4 months, it's necessary for players to warm up to playing 162 games with scattered off days. The 6 weeks leading up to Opening Day allow position players to refocus their hitting eye, work on baserunning, and generally hone their baseball skill set. For pitchers, it's about building stamina: starters gradually build their pitch count towards 90-100 pitches, and both starters and relievers may experiment with new grips, mound positions, or do things that aren't very amenable to adjustment mid-season. Finally, player development staff qualitatively evaluate players to see who has made positive adjustments, needs more adjustments, or just doesn't have the skillset to crack the major league roster.

What Spring Training is not about is comparing small samples between players in a fundamentally biased environment. Not only do position players get roughly 60-80 PA, they are collected in an environment that is completely different than the A's home park, and the road stadiums they most commonly play in. Granted, the addition of Houston yields the A's more games in a hitters' park than before, but the idea that 80 PA is complete enough to be drawing any conclusions from is just unsound. Indeed, when I hear Bob Melvin say that the A's might be relying on Spring Training stats more heavily than other teams, I cringe. I suppose the team may have a system devised to properly couch spring training data - Pitch F/x, for example, might be very useful to identify release point corrections - but again, the problem is more fundamental: crap data in, crap data out.

Given what I have said about Spring stats, let's do some comparisons. Fortunately, MLB.com keeps spring training stats - one of the few sites that does. Unfortunately, they only have basic stats, so we'll stick to the best available: OPS. So, to illustrate my point, let's look at the top 5 OPS leaders during Spring Training for the last 3 years, excluding 2013:

2010:

Sean Rodriguez (1.373)

Aaron Hill (1.371)

Mitch Maier (1.344)

Jose Bautista (1.343)

Mike Aviles (1.243)

2011:

Kila Ka'aihue (1.306)

Michael Morse (1.239)

Miguel Cabrera (1.213)

Alex Gordon (1.187)

Travis Buck (1.165)

2012:

Albert Pujols (1.287)

Chris Young (1.232)

Lorenzo Cain (1.193)

Billy Butler (1.161)

Delmon Young (1.156)

To be sure, there are legitimate MLB players on these leaderboards, and some stars. But there are also guys that are barely holding on to MLB roster spots. Travis Buck? Trying to make the Padres, but probably starting in Tucson. Kila Ka'aihue? Reassigned to minor league camp by the Diamondbacks. Mitch Maier? After years of kicking around with the Royals, he was picked up by the Red Sox. Unfortunately, his chances of making the Boston roster after spraining his wrist appear to be minimal.

For comparison's sake, here are the OPS leaders during that same time period during the regular season:

2010:

Josh Hamilton (1.044)

Miguel Cabrera (1.042)

Joey Votto (1.024)

Albert Pujols (1.011)

Jose Bautista (.995)

2011:

Jose Bautista (1.056)

Miguel Cabrera (1.033)

Ryan Braun (.994)

Matt Kemp (.986)

Prince Fielder (.981)

2012:

Miguel Cabrera (.999)

Ryan Braun (.987)

Mike Trout (.963)

Buster Posey (.957)

Andrew McCutchen (.953)

You see my point. These are the best players in the world, and the names here, and their respective stats, make a lot of sense. Sorry, but Travis Buck (or really anyone else) isn't going to have a 1.165 OPS over a whole season. Eric Sogard and Shane Peterson won't lead the A's in hits were they to have 650 MLB PAs in 2013. Hiroyuki Nakajima might not be as good of a player as the A's had hoped, but 80 Spring Training PAs having just come over from Japan won't tell us that. 650 PAs of consistent play, however, will go a long way to settle out some of the positional battles the A's do have right now.

Drawing any conclusions from 80 PA of data is a fool's errand. Whether it's for fantasy purposes, or for actual games, players ought to be judged on their track record of MLB or MiLB performance. Jason Parks, of Baseball Prospectus fame, runs a series called Prospects Will Break Your Heart. To borrow that a little bit, I'd say Spring Training Statistics Will Break Your Heart.