Monday, June 30, 2008

Mets demonstrate regression to the mean.

Statistics was about my least favorite field of study. What you learn is that, while wonderful things can happen, they are very unlikely; and, if they do, they are very, very unlikely to occur again anytime soon. Perhaps the most depressing statistical concept is regression to the mean. This tells you that, if someone has a history of doing badly, then suddenly has a run of success, that run will almost certainly end and s/he will return to her/his pattern of failure.

Back in May, the Mets beat the Yankees twice, at Yankee Stadium. One game of the three game series was rained out and rescheduled for this past Friday as part of a doubleheader, the first game to be played at Yankee Stadium and the second at Shea. Anyway, after the Mets swept their abbreviated series, I posted this celebratory piece, with the cautionary note, "just don't backslide". By writing this, of course, I was challenging the notion of regression to the mean, and regression prevailed, in a big way, as the Mets then dropped four straight to the Braves.

So, now, as I savor the Mets having won their season series with the Yanks, I see them regressing once again, dropping tonight's game to the Cards, 7-1, as John Maine has his turn in the starting rotation's frequently used barrel. The Mets thus also blew a chance to help the Cubs, a team I may have jinxed, who were swept this weekend by their Southside rivals.

Tim Marchman bets on regression to the mean in his column in the New York Sun about Ollie Perez.