Buster Olney still doesn’t understand small samples

In his blog today, Buster Olney is back on the “Coors Field East” drum he’s been beating all season. It’s gotten tiresome. Olney ran with the issue after the Yanks’ first homestand — which we’ll get to in just a second — and revisited the topic because of the five home runs the Yanks and Twins hit last night. Both of these instances constitute minuscule samples, and it leads Olney to some wrongheaded analysis.

However, after five homers were hit in the Yankees’ win over the Twins on Monday, there have now been 63 homers in 17 games in new Yankee Stadium: 32 by the Yankees, 31 by the Yankees’ opponents. Last year, the Yankees’ pitchers allowed 68 for the entire season, and the Yankees’ hitters mashed 92, for a total of 160. So at the current rate, there will be more homers hit in new Yankee Stadium by July 17 — the first home game after the All-Star break — than there were during the entire 2008 season in old Yankee Stadium.

First off, extrapolating from a small sample is no better than analyzing a small sample by itself. All it does is magnify the random effects experienced in a sample of just 17 games. As a for-instance, just look at then-Tigers first baseman Chris Shelton in 2006. He hit nine home runs through the first 13 games of the season. If you extrapolated from that small sample, Shelton would have hit over 100 homers that season. Obviously that’s a stupid analysis, and it’s exactly what Olney is doing when he looks at homers in the new Stadium.

Yes, there have been an inordinately high number of balls flying out of the new Stadium, but to look at the numbers now and decide that it’s a bandbox is silly. There are circumstances which can arise in small samples which skew the results. The most notable is the first series at the Stadium. In the four games against Cleveland, both teams hit 20 home runs. In the two games against Oakland both teams hit six, meaning there were 26 homers hit over six games, or 4.3 per game. Now look at the 11 games since, wherein the Yanks and their opponents combined for 37 home runs, or 3.36 per game.

The home run total doesn’t figure to be a statistic which will correct itself. The Yankees have 64 remaining home games, and it’s a virtual lock that more than 97 balls will leave the park in that span (160 balls left Yankee Stadium last season). However, to take the numbers in such a small sample is a fool’s game. The Yankees have barely played a fifth of the total games at Yankee Stadium so far. The first series at the Stadium sticks out to us because of the 20 home runs that left the park, but that’s a ridiculous pace which can never be sustained, even in the most hitter-friendly park. The number has been trending downward since, as any statistician would have predicted.

Olney asks at the end of his bit, “Is it still too early to say the new place plays small?” Again, considering the Yankee Stadium schedule isn’t even a quarter done, the answer is yes. It is way too early. Please, Buster, let’s wait until July to determine how many home runs will leave the park at that point.