Friday, March 12, 2010

One Spring Statistic that Matters

Most of us know that spring training statistics are not very meaningful. Every spring there is a young player who bats .375 in 40 at bats. When the season starts, he is either in Triple-A or struggling to hit hit major league breaking pitches. Every spring, there is a veteran pitcher with an ERA of 7.50. Once the real games begin though, he pitches seven strong innings in the season opener.

Spring training numbers are generally unimportant for a few reasons. First, they are based on small sample sizes. We've all seen weaker hitters get off to strong starts for the first three weeks of the regular season, only to crash to earth by the middle of May. Players can go hot and cold in spring training as well.

Spring performance is especially tricky because the competition is weaker. There are a lot of players used in spring training games, such as career minor leaguers and rookies not yet ready for the majors, who will not be playing once the season starts. It's a lot easier to pile up hits against minor league pitchers than against major lerague pitchers every night.

Finally, players are not always playing to win in spring training. A pitcher may be working on a new pitch or delivery. A batter might be trying a new stance or a different approach. Veteran players, already guaranteed a roster spot, might simply be going through the motions just trying to stay in shape while avoiding injury.

So, it's easy to see why most spring numbers are generally meaningless. However, John Dewan, owner of Baseball Info Solutions, did a study on spring statistics a few years ago and concluded that one statistic does have some meaning:

A hitter with a positive difference between his spring training slugging percentage and his lifetime slugging percentage of .200 or more correlates to a better than normal season.

So, if a player has a .400 career slugging average during the regular season and he is slugging .650 during spring training, you might expect him to slug significantly higher than his usual .400. in the coming season. There is no correlation for other statistics such as batting average, on-base average and ERA.

Dewan says that at least 35 spring training at bats are required for his theory to have meaning.
Nobody has 35 at bats yet but if, in a couple of weeks, Magglio Ordonez is still slugging .929 and Don Kelly is at .882, it might be worth noting. We'll visit this again later.


  1. Interesting! I do have one question: where can you find spring training stats?

  2. OK, disregard my first question: I found them.

    I do have two other questions though. First, does the reverse hold true as well? If a players SLG is significantly worse than his career norm, does this correlate to a worse-than-average regular season?

    Second, what if the increase in SLG during spring training can be almost entirely attributed to an increase in BA? Take Austin Jackson, for example, who is currently slugging .650 in spring training (compared to a minor league career SLG of .410). This is a difference of +.240, but this is aided by an abnormally high .450 BA in spring training. Would it be more useful to use ISO as the indicator instead of SLG? For AJax, his spring training ISO is .200 compared to a ML career ISO of .122. What do you think?

  3. I don't think there is any evidence that the reverse is true.

    Your second point is a good one. If it's an increase in power that matters, then ISO should work better. It also could be a combination of power and average increasing that is important. I don't have an answer but would love to see some research on ISO.

    For those who could not find the stats, they are here:

  4. Good point, Lee. I had forgotten about that.



Blog Archive


My Sabermetrics Book

My Sabermetrics Book
One of Baseball America's top ten books of 2010

Other Sabermetrics Books

Stat Counter