Full disclosure: Every year I love the prospect of a 16 winning its first-ever game over a one-seed. Loyal readers will recall my irrational exuberance regarding Oral Roberts in 2006.
This is not without reason, though. If you run the numbers, the chances of all of the one-seeds getting to the Final Four are typically roughly equivalent to one of the top-seeds losing in the first round. Yet somehow history has betrayed the odds. The Apocalypse Scenario played itself out in 2008 when all of the one-seeds made the Final Four, while 16-seeds are 0-108 against one-seeds all-time. And there have really only been four cases where the outcome was in doubt in the final minute – none since Western Carolina took Purdue to the wire in 1996.
Even with this enormously long drought, I make it a point to watch each 1/16 game from the beginning, knowing that there’s the potential for history, and that if it occurs, I’ll have been along for the ride from start to finish. Usually, I need only maintain this ritual until the first or second media timeout when the game is obviously out of reach. But every once in a while there’s an Albany/UConn in 2006 or a Holy Cross/Kansas in 2002 where my vigilance is rewarded with a mildly interesting game in the second half. On Thursday and Friday, those are the moments I live for.
I’ll continue that tradition this season, once again believing this is the year. And this really could be the year. I say this every year, of course, but this time I mean it. For reals. The crop of 16 seeds in the 2012 tournament appears to be the strongest ever. I went back through my ratings going back to 2003 and then the SRS ratings at College Basketball Reference for data before then to find the 16-seeds that were ranked the highest. These are the ten-highest ranked 16-seeds of all-time:
Year Team Rank 2012 Lamar 102 1993 Wright St. 105 2009 E. Tenn. St. 111 2006 Oral Roberts 115 2012 Vermont 120 2012 UNC Asheville 121 1986 Montana St. 124 2008 Portland St. 127 1987 Penn 130 1985 Southern 130
Yep, 2012 has produced three of the strongest 16-seeds in the history of the tournament. Consider, too, that comparing the raw ranking of a 2012 team to teams from the ‘80s and ‘90s isn’t totally fair since there were fewer Division-I teams years ago. A ranking close to the top 100 is more impressive now with 345 teams in the mix than it was when there were 50 fewer teams. Also consider the absurdity of Lamar and Vermont facing off in a play-in game just to earn the right to slay a one-seed. Giving #213 Norfolk State a 15-seed and #165 LIU a free pass to the round of 64 weren’t the committee’s best moments on Sunday.
People have noticed the spuriously high probabilities for the low seeds to advance past the round of 64 in the log5 analysis, including this e-mailer:
Hello Mr. Pomeroy,
According to your NCAA Tournament log5 analysis, there’s only about a 50.9% chance of all eight teams seeded #1 or #2 reaching the round of 32. Seems ridiculously low. Even if you ignore Duke/Lehigh completely, there’s still about a one in three chance that one of the other 7 bites the dust. Do you think this has more to do with unusually strong #15 and #16 seeds or is it mostly due to log5 underrating a high seeded team’s round of 64 chances?
I’ll agree that a 49% chance for a one- or two-seed to lose in the first round is stunningly high. I fear log5 overstates the chances of heavy underdogs. There were only two cases this season of a team winning a game when given less than a 5% chance of success.
The estimate isn’t so far off though. Based on the Vegas lines and assuming Kentucky and UNC have a 99% and 96% chance of winning its first games, respectively, the chance of a one- or two-seed falling is about 38%. And the chance of just a one-seed falling is roughly 15%, which is much, much greater than the chance of all four ones getting to the Final Four, which is about 1% this season.
I get my hopes up for this every year, though, and for 27 consecutive years I have been disappointed. Of the other teams on the list of the highest-ranked 16-seeds, the only one that made a feeble run at an upset bid was 2009 East Tennessee State which gave Pitt a minor scare. For instance, 1993 Wright State found itself on the unhappy end of a 43-point thrashing from Indiana. Thus, it was probably in the top 100 when it entered the tournament, even better than Lamar.
Despite the awful track record, I’ll be watching each of the 1/16 games from the opening tip, with a special eye on whomever emerges from the Lamar/Vermont game to take on UNC. This will be the year.