Rick Nelson

Baseball and test escapes
October 26, 2005

I recently commented on peculiar scientific theories and how they might help propel our local baseball team into the post-season championship rounds. The Red Sox did survive the regular season, so perhaps those theories were successfully applied on the team's behalf. But the Red Sox promptly fell to the White Sox three games to none in a best-of-five division series.

However, there's a scientific explanation to that as well, as described in the Wall Street Journal by sports columnist Sam Walker ("Blame the Umpires," October 14). A subscription is required to access the column, but the argument, based on statistics and quality control, basically goes like this (I've adapted it for the test-engineering community):

The White Sox are known for their pitchers' prowess, while the Red Sox rely on their hitting. For the best chance of success, Red Sox hitters need high-quality pitches--ones that pass within the strike zone, defined by the width of home plate and the distance from a batter's knees to chest.

An umpire has only a few milliseconds during which a ball traverses home plate to apply his "test" of the pitch, so it's not surprising that results can be somewhat inconsistent. Unfortunately for the Red Sox, the three umpires who worked home plate during the series tend to exhibit an uncharacteristically high number of "test escapes": pitches out of the zone that they nevertheless call strikes. Knowing this, Red Sox hitters had to swing at low-quality pitches or risk striking out looking. Consequently, they performed poorly and were able to score only 11 runs in three games.

Evidence of the test escapes resides in statistics: during the regular season, average earned runs scored per game with the same three umpires calling balls and strikes averaged only 7.3, vs. a league average of 8.6.

It seems that Major League Baseball has a serious test correlation problem.

Posted by Rick Nelson on October 26, 2005 | Comments (0)