clock menu more-arrow no yes mobile

Filed under:

ST Correlations, Volume One

(Not sure if there'll be a volume two, but just in case, there you go.)

Every March, the same debate rages on with moderate fury - what do we do with spring training statistics? Sometimes they're used as an indication that big things are on the way (Beltre!), whereas other times they're tossed aside without a second thought. Are they really as meaningless as people such as myself like to say they are?

Earlier this afternoon I thought I'd take my first stab at answering that question by examining the correlation between individual batting averages in ST and the following season. The data pool includes all 246 players who picked up at least 50 at bats last year in both the exhibition and regular seasons. I did my best to remove park factors by using Baseball Prospectus' translated batting average, and while it's far from perfect, I think it's better than using the raw number. The inherent assumption here, of course, is that ST ballparks are neutral across the board, but there's not much I can do about that.

Anyway, to the results:


Statistically speaking, the correlation is significant at the 1% level, but it's still a low r value, far closer to zero than to one. For those of you in possession of The Hardball Times Baseball Annual 2006, the correlation here is roughly equivalent to the season-to-season correlation of a hitter's ability to turn outfield flies into doubles/triples (I only cite this because it's one of the only "popular" examples of similar work in the field of baseball of which I'm aware). It's real, but it's weak.

Put another way, the year-to-year correlation for batting average is around 0.40, so as far as projecting (translated) BA is concerned, spring training is approximately half as useful as the previous season. Which is to say, you can use it, but it's not going to take you very far.

So that's that. While it's not a picture of complete and total randomness, it's not much, and you'd really just be better off deferring to the previous season entirely when it comes to talking about what you expect to see in the coming months. Leave the spring training batting averages to the people who love to make retrospective arguments (Bret Boone hit .215 last spring! Of course he was going to collapse!). While in some cases there might be something there, it's just too difficult to separate what's real from what's a fluke, and in the end it's just not worth the attempt.

If there are more of these, they'll be posted in the coming days. If there aren't, then great, that saves me a lot of time. Either way, I'm just glad that legitimate baseball begins in earnest this Sunday.