Monday, May 15, 2006

Continuing with the previewing of the NCAA regionals, the University of Arizona has received the No. 2 seed nationally (behind UCLA) and a lot of "buzz", thanks in good part to a 10-game winning streak to close the season (and, presumably, Texas's two-and-out collapse at the Big 12 tournament).

As shown in this U of A game-by-game log, the Wildcats defeated Washington, Stanford, Oregon, and Oregon State twice each, and Arizona State and Cal once each, during the streak.

One often hears about the importance of momentum heading into play-off competition, and the 2005 MLB World Series champion Chicago White Sox provide some anecdotal support for the proposition. The Chisox won their final five games of the regular season, then went 11-1 in the post-season, for a 16-1 close-out, overall.

In late September 2005, in fact, the baseball website Hardball Times published a piece analyzing many years of data to assess the degree to which September success carried over to championships in October, above and beyond overall season record. The article appears to show that any benefit to teams from momentum carryover is fairly limited.

I operate another website, in conjunction with my statistics teaching, called the "Hot Hand," where I investigate sports streakiness more generally. Demonstrating that an observed streak -- whether it's a baseball/softball hitting streak, basketball shooting streak, winning streak, or losing streak -- has any significance beyond chance is difficult. After all, if you flip a coin hundreds of times, you'll occasionally obtain streaks of several heads or several tails in a row.

The author of the Hardball Times piece raised the following question for the 2005 MLB post-season, which can also be applied to the Arizona Wildcats entering the 2006 NCAA softball tournament:

What does all this mean for this year? It means that some people will pick the hottest team (say, the Indians or Yankees) to win it all, because they have momentum on their side. Based on this study, I'd say that the postseason is still a relatively random event.

One obvious difference is that the MLB play-offs include only eight teams, meaning that ability levels are fairly similar and thus the role of chance is increased. In the 64-team NCAA softball field, most of the favored teams will likely advance on, in predictable fashion. It's only at the Women's College World Series (or perhaps just at the later stages of it) that the teams' ability levels will be sufficiently similar to allow a large role for chance and randomness.