“Spring training stats don’t matter.” We often repeat that. And, in a general sense — the sense in which most of us think about stats — they kind of don’t. A guy who hits .395 in the spring is not necessarily poised for a batting title. A pitcher who puts up a spring ERA of 1.04 is not bloody well likely to keep that going. Even in a broad “guys who do well in spring will do well in the regular season” assertion is not very well borne out by the numbers.
But that doesn’t mean that there are no predictive spring training stats. Dan Rosenheck of the Economist believes he has determined that there are:
Yet in spite of all these caveats, the claim that spring-training numbers are useless is wrong. Not a little bit wrong, not debatably wrong—demonstrably and conclusively wrong. To be sure, the figures are noisy. But they still contain a signal. At the MIT Sloan Sports Analytics Conference held in Boston on February 27th-28th, I presented a study (see slides) that explained how to extract the statistical golden nuggets buried in this troublesome dataset, and offered some lessons this example provides for the practice of quantitative sports research more broadly.
The stats that are predictive: peripherals like strikeouts per at bat for hitters and K/BB ratios for pitchers. The predictions one can make from such things are not spelled out in 50-foot neon letters and numbers — thus making them not easily consumed by the majority of us who are, at best, analytical dilettantes — but there is a signal above all of the spring training noise to be had if you’re looking for it.
Rosenheck’s larger takeaway: whenever you hear someone assert something unequivocally like “spring training stats are meaningless,” don’t buy it. Because it’s quite possible they just haven’t looked hard enough. And maybe don’t care to.