<
>

Beware of spring training statistics

With the dawn of a new spring training comes optimism and excitement for a slate of baseball-filled months. The freshly cut grass, the neatly chalked lines and the carefully manicured infield dirt are just a few of the aesthetic wonders we come to appreciate. Oh, and the stats. Each spring training at-bat, whether by a 15-year MLB veteran or a doe-eyed rookie, is meticulously recorded.

Perhaps the most appealing part of the return of baseball is the mountain of numbers populating tables on various baseball websites. The notion that the future of a highly touted prospect or a once-injured player on the mend can be foretold by his spring training performance is as much a part of baseball as peanuts and Cracker Jack. The numbers, though, come with a lot of pitfalls that should be made apparent.

The first issue is that, in general, the sample sizes are very small. Last year's leader in spring training at-bats was Hunter Pence, who took 85 of them for the Phillies. Overall, 984 players took at least five at-bats last spring, but only 114 of them took even 60 at-bats -- less than 12 percent. You would be just as accurate flipping a coin to decide these players' fates as opposed to relying on their spring training statistics.

Russell Carleton of Baseball Prospectus, after doing some mathematical heavy lifting, decreed that the only statistic that stabilizes before at least 100 plate appearances is strikeout rate, defined as strikeouts divided by plate appearances. Among other rate stats, the only others that stabilize prior to 200 PAs are walk rate, home run rate and isolated power. But, as evidenced by Pence's leading AB total, even the players with the most playing time don't even approach half the requisite PAs in a given spring.

It is worse for pitchers as the sample sizes are even smaller. The leader in spring innings pitched last year was Randy Wolf at 30.1, more or less half of a season for an everyday reliever. In total, 868 pitchers threw at least two innings during 2012 spring training, but only 84 of them -- less than 10 percent -- crossed the 20-inning plateau.

To make matters worse, the statistics are made all the more unreliable by a host of other factors, perhaps most important of which is that many of the players deciding games every day are among their team's least important players, or minor leaguers. For example, in Phillies camp last year, Pete Orr had 67 PAs, Scott Podsednik had 63, Hector Luna had 57, Lou Montanez had 58, Tyson Gillies had 24 and Tuffy Gosewich had 22, combining for 291 spring training PAs. The same group (actually just Orr and Luna) combined for 123 for the Phillies during the regular season with Montanez, Gillies and Gosewich never getting a taste of the majors during the season.

During the regular season, you can expect that a vast majority of the batters a pitcher faces will be of major league quality. That is not the case during spring training, when many veterans are shuffled out of the game by the fifth inning, particularly in late February and early March. For most of spring training, at least half of every game is filled with veteran journeymen and young prospects, very few of whom ever make an impact at the major league level in the same season.

On any given day, both pitchers and hitters might be focusing on one particular area of their game on which to work. Maybe Justin Verlander hasn't felt good about his fastball, so he will make a point to throw nothing but fastballs in his spring training outing. During the regular season, Verlander would never do that, no matter how much confidence he has lost in his fastball. Additionally, the art of game-calling tends to take a back seat, so the cat-and-mouse game between the pitcher and hitter becomes a glorified version of batting practice. Hitters, rather than trying to make the best contact, might be trying to perfect swing mechanics or practice situational hitting.

Defense muddies the water even more. In the infield, you might have an unfamiliar shortstop and second base combination. Washington Nationals fans, used to Ian Desmond and Danny Espinosa teaming up, saw Zach Walters and Will Rhymes there Sunday. Younger catchers learn to manage the game and develop a rapport with a revolving door of pitchers. Younger center fielders learn to direct traffic in the outfield while some players playing in the corner outfield for the first time get used to hitting the cutoff man. To say spring training defense is sloppier than regular-season defense would be a massive understatement. Such defense affects the stat lines of both the pitchers and the hitters, and in such small samples, is dangerous if left unnoticed.

If you relied on spring training stats in 2012, you probably thought Eric Hosmer (1.127 spring OPS) was poised for a breakout year, but the Royals first baseman finished with a .663 regular-season OPS. Likewise, you might have thought Francisco Liriano finally turned a corner (2.33 spring ERA), but the lefty finished with a 5.34 ERA with the Twins and White Sox. When it comes down to it, there are simply too many factors simultaneously affecting a very limited set of data. You are better off simply enjoying the stats for what they represent -- the return of baseball -- than anything else.

Bill Baer runs the Phillies blog Crashburn Alley. You can follow him on Twitter @CrashburnAlley.