May 2010

Out on a limb: Weakness in numbers

By Jordana Bieze Foster, Editor

Sometimes it’s hard to tell when it’s spring in New England, particularly when there are frost warnings in late April and the local hockey and basketball teams are still playing well into May. But  one telltale sign is when the air is abuzz with baseball statistics.

Baseball fans everywhere love their stats, and Red Sox Nation is no exception. And it’s not just the numbers themselves, although some – like Josh Beckett’s 7.22 earned run average and David Ortiz’s .154 batting average in April – have been popular topics of discussion. Like savvy practitioners analyzing an abstract, Sox fans are also keenly interested in just how those numbers are calculated.

This spring, the spotlight was on the Ultimate Zone Rating (UZR), developed by sabermetrician Mitchel Lichtman in 2001 for ranking a fielder’s defense. Technically, it is “the number of runs above or below average a fielder is in range runs, outfield arm runs, double play runs and error runs combined.”

This measure takes into account ballpark dimensions, field surfaces, direction and velocity of balls put into play, batter handedness, number of base runners and outs for every ball fielded. But until very recently, it didn’t account for “quirky parks and portions of parks,” such as left field at Boston’s Fenway Park.

That was bad news for former Sox left fielder Jason Bay, when he became a free agent at the end of last season. At the time, his UZR was a below-average -13.8. Now that the formula has been revised, his UZR is a much more respectable +1.9.

That might help Bay save some face, but it doesn’t mean he gets a do-over on free agency. He won’t get a second shot with teams that might have passed on him, or offered him less attractive terms, because of questions about his defense.

It’s worth noting that, even as UZR’s relative merits have been dissected and analyzed in the last month, most discussants acknowledge that the Red Sox and other major-league front office staffs are unlikely to rely solely on UZR or any other single measure in evaluating a player’s defense. Nor have any teams done away with scouting either.

UZR is attractive because it is easy to implement. The same can be said of many measurements used by researchers and practitioners to assess lower extremity function. But trying to sum up a complicated process in a single number is always going to be problematic.

There will inevitably be variables not accounted for in a given statistical model, although the effects of not accounting for them may vary. To be fair, the recent changes to the UZR calculation affected only a small number of major-league players. Similarly, the shortcomings of measures that you use regularly in your practice may affect only a small number of your patients. But that doesn’t mean you shouldn’t be aware of the possibility. Don’t discount your clinical experience; sometimes numbers really do lie.

If only that were the case for David Ortiz.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.