Home Unfiltered Articles Players Baseball Prospectus
Basketball Prospectus home
Click here to log in Click here for forgotten password Click here to subscribe

Click here for Important Basketball Prospectus Premium Information!

<< Previous Article
The Clipboard (03/11)
Next Article >>
Premium Article Mid-Majors Prevail (03/14)

March 14, 2011
Seeding by Numbers
This Year's Field

by Kevin Pelton


While much of the discussion after yesterday's NCAA Tournament selection show focused on who did not make the tournament--notably Colorado, St. Mary's and Virginia Tech--where the chosen 68 were seeded will surely ultimately have more relevance than at-large teams that were unlikely to advance past the opening round anyway.

A year ago, as fans debated whether West Virginia deserved a No. 1 seed (the Mountaineers would advance to the Final Four anyways as the second seed in the East), I investigated the recent history of seeds through the prism of my colleague Ken Pomeroy's ratings. The findings confirmed some conventional wisdom, helping explain the common 5/12 upsets and demonstrating what an advantage top seeds have in the opening round. Over a four-year span, the average performance of each seed lined up as would be expected.

That is certainly not the case when we look at the 2011 field, again using Pomeroy ratings. Because a handful of key teams underperformed compared to their statistics, and because the selection committee's criteria differs from the objective methods Prospectus favors, the distinction between seeds this year is fine at best.

What we see, instead, is a few groupings. The top four seeds are all, on average, relatively similar. In part, this reflects a year in which the NCAA has not been especially top-heavy. With a handful of teams trading the top spot in the polls throughout the regular season, the gap between the top four teams and everyone else is a small one. The selection committee and the Pomeroy ratings agree on three of the four No. 1 seeds. The difference is that, by the numbers, the last top seed should have been Texas--which actually ended up on the four line.

From seeds five through nine, the average difference is difficult to spot. The difference, in terms of Pythagorean winning percentage, is only .026. By comparison, from 2006-09, the difference between those seeds was .035. There is another group of relatively similar teams, on average, between the 10 line and the 12 line before talent begins quickly dropping off among the automatic qualifiers from smaller conferences.

Another way to look at the data is to compare the best and worst team at each seed.

This perspective highlights some of the unusual seeds. It also reinforces how unusually good the teams on the four line are this year. Every four seed is better than a pair of two seeds (both North Carolina and Florida). Washington is an uncommonly strong No. 7 seed, part of a top-heavy East region (five of the Pomeroy ratings' top 15 teams are in the East). The gap between the best and worst teams at the same seed grows as we enter the double-digits and encounter some odd placement. No two teams are more underseeded than Utah State and Belmont, No. 12 and No. 13 in the Southeast. In fact, only one team seeded lower than a five (the aforementioned Huskies) has a better Pomeroy rating than either the Aggies or the Bruins. While the committee was fair to mid-majors overall, it's hard to see where those two teams were seeded as anything but disrespect to high-achieving champions of poor conferences.

When you combine the relatively flat distribution of quality by seed with a handful of teams that were underseeded, you get the potential for a lot of upsets in the early rounds of this year's tournament. Nate Silver of FiveThirtyEight.com noted on Twitter that his model projects 9-10 first-round upsets instead of the usual seven or eight.

As the brackets are revealed, there is also a rush to determine which regions are strongest and weakest. Surprisingly, the numbers indicate that they are relatively balanced. If we average the Pomeroy ratings for all 16 teams--assuming that the better team will win each of the "First Four" play-in games and eliminating the lesser team from consideration--the Southeast has the strongest rating at .848, with the other three regions between .821 and .829.

The problem with this method is that it implies that the distinction between the respective 16 seeds are more important than almost anything else. But while North Carolina Asheville is much stronger than the other 16s, that doesn't change the odds of Pittsburgh moving by a noticeable amount. Even the Southeast's terrific pair of 12 and 13 seeds is unlikely to have much long-term impact. We can weight instead by the number of games each seed would be expected to play within the region--four for Nos. 1 and 2, three for Nos. 3 and 4, two for Nos. 5-8 and a single weight for everyone else--to get a better measure of which region is most difficult at the top. Using this weighted average, all four regions are essentially the same. The Southeast is still on top at .883, but the weakest region (the West) is not far behind at .876.

Most of the time, when analysts discuss the regions, what they are really considering is how difficult the path is for each No. 1 seed to reach the Final Four. This, too, can be simulated. My simplistic method merely assumed that the top seed would play each seed as expected en route to Houston (first the 16 seed, then the 8, then the 4 and finally the 2) and assigned them all approximately the average Pomeroy rating for a No. 1 seed (.870).

From this view, the Southeast is in fact the easiest region, with a typical No. 1 seed having a 36.8 percent chance of advancing to the Final Four. As tough as the bottom of the Southeast is, the region also features the weakest No. 2 seed (Florida) and also has a subpar pair of 8 and 9 seeds. Add in a BYU team that has struggled without Brandon Davies and Pitt will be tough to beat. Despite a tough possible Sweet Sixteen matchup against Kentucky, Ohio State (32.0 percent) has the second-easiest path to Houston because North Carolina also rates poorly for a team on the two line. Kansas' schedule gives an average No. 1 seed a 31.0 percent chance of reaching the Final Four. The most difficult road, then, belongs to Duke--which does not match conventional wisdom. Even before accounting for the fact that the Blue Devils might have to face San Diego State in Anaheim (see Silver for more on the importance of travel in the NCAA Tournament), potential matchups against a tough Aztecs team and Texas in the Sweet Sixteen mean an average No. 1 seed would have just a 26.4 percent chance of winning the West region.

An issue as relative trivial as which region the teams were slotted could be the difference between reaching the Final Four and falling short in one out of 10 scenarios. That's what we should be worrying about in the wake of Selection Sunday.

Kevin Pelton is an author of Basketball Prospectus. You can contact Kevin by clicking here or click here to see Kevin's other articles.

0 comments have been left for this article.

<< Previous Article
The Clipboard (03/11)
Next Article >>
Premium Article Mid-Majors Prevail (03/14)

State of Basketball Prospectus: A Brief Anno...
Tuesday Truths: March-at-Last Edition
Easy Bubble Solver: The Triumphant Return
Premium Article Bubbles of their Own Making: Villanova, Temp...
Tuesday Truths: Crunch Time Edition

Premium Article Mid-Majors Prevail: Or Did They?

2011-03-18 - Premium Article Start/Sit Projections: Week of March 21
2011-03-16 - The NBA Perspective: A Tournament Preview
2011-03-15 - Premium Article Tournament Preview: Kansas in the Southwest
2011-03-14 - Premium Article Seeding by Numbers: This Year's Field
2011-03-11 - Premium Article Start/Sit Projections: Week of March 14
2011-03-10 - Premium Article Mixed Results: Reevaluating Boston's Trade
2011-03-09 - Premium Article Tournament Preview: Pac-10 Tournament

Basketball Prospectus Home  |  Terms of Service  |  Privacy Policy  |  Contact Us
Copyright © 1996-2017 Prospectus Entertainment Ventures, LLC.