It’s only five days until the New York Yankees’ first spring training game against the Atlanta Braves. Baseball is almost back, and I’m absolutely giddy with excitement, as are all of you I’m sure.
However, I want to take this time to remind you all about the perils of taking spring training statistics too literally.
Every year, there are bad players that have fantastic springs and good players that have terrible springs. Usually, these bad players end up being bad in the regular season, and vice versa for the good players. Every once in a while there are cases in which spring training performance is indicative of season success. The problem is, these cases are almost impossible to separate from the cases of random variation and small sample size.
If Cervelli has a monster spring training, don’t take that to mean that he’ll bring that into the regular season. (Image: John Munson/THE STAR-LEDGER via USA TODAY Sports)
Yep, I said it. Small sample size. It’s a phrase that’s sometimes overused and can certainly be frustrating for the everyday fan. We want to believe that breakout performances are indicative of a player’s true talent. We want to weave stories to explain the stats. However, it’s simply the case that there is a ton of noise – that is, random variables that affect the outcome – when the sample is too small.
Not only is spring training too small of a sample to really learn much about a player, but there are other variables that make it unreliable as well. Here are just a few:
- Pitchers use this time to try out new pitches and experiment with different approaches
- Minor leaguers and other players that won’t touch the majors any time soon get a chance to play.
- Players aren’t giving 100% because they don’t want to risk injury or because they are rehabbing from injury.
Overall, it’s just a different playing environment in spring training, so most of the stats that come out are very distorted and come with too much noise.
Nevertheless, some stats are more useful than others, even in spring training. You want to take any stat you find with a grain of salt for the reasons listed above, but here’s a list of just a few stats that are worth keeping an eye on, though not worth drawing conclusions from:
- Fastball velocity – Again, you need more than just one start for this to be relevant. But if a pitchers’ velocity is consistently down during all of spring training, especially towards the end, something may be wrong. Key word: may.
- Strikeout rate (batters and pitchers) – K% is about as reliable a statistic as you are going to find in spring training, which isn’t saying much. Because of worse competition and the fact that pitchers are experimenting more, strikeout rate won’t be perfect, but a big swing in one direction or another just might be indicative of something larger.
- Walk rate (batters and pitchers) – This is less reliable than strikeout rate, but again, big surprises one way or another should be investigated further.
And that’s pretty much it. I know that’s not a lot to draw from, but that’s sort of the point. Spring training is about preparing for the season, not predicting what will happen during the season. These players aren’t trying to win the games, but hone their skills. Consequently, the stats that we should look at are the ones that most accurately measure the pure skill of the player. So don’t use ERA or RBI or even batting average; look at velocity, strikeouts, and walks, and even then, take everything that you see with a grain of salt.