NCAA tournament goes without 12-5 upset for 1st time since 2007
When the public fills out their brackets for the 2016 NCAA tournament, they may have to throw away an old, tried-and-true method for identifying upsets.
For the first time since 2007, this year's tournament didn't have No. 12 seed beat a No. 5 seed and advance to the Round of 32.
Buffalo came close against West Virginia, pulling close late but ultimately fell 68-62. Earlier Friday, Wyoming put up far less of a fight against Northern Iowa. Thursday saw Stephen F. Austin hang tight but fall 57-50 to Utah, and Wofford get edged 56-53 by Arkansas.
It perhaps shouldn't be surprising for four favorites to sweep four underdogs, but this year is a stark contrast to recent seasons. In each of the last two years, three 12-seeds were victorious, and the tournament's history has far more 12-seed success than makes sense on the surface:
| Year | 12-seed over 5-seed | Total 12-seed wins |
|---|---|---|
| 2015 | 0 | 0 |
| 2014 | 3 | 4 |
| 2013 | 3 | 4 |
| 2012 | 2 | 3 |
| 2011 | 1 | 3 |
| 2010 | 1 | 2 |
| 2009 | 3 | 4 |
| 2008 | 2 | 4 |
| 2007 | 0 | 0 |
| 2006 | 2 | 2 |
| 2005 | 1 | 2 |
| 2004 | 2 | 2 |
| 2003 | 1 | 2 |
| 2002 | 3 | 5 |
| 2001 | 2 | 3 |
| 2000 | 0 | 0 |
| 1999 | 2 | 3 |
| 1998 | 1 | 1 |
| 1997 | 1 | 1 |
| 1996 | 2 | 3 |
| 1995 | 1 | 1 |
| 1994 | 2 | 3 |
| 1993 | 1 | 2 |
| 1992 | 1 | 2 |
| 1991 | 1 | 2 |
| 1990 | 2 | 3 |
| 1989 | 1 | 1 |
| 1988 | 0 | 0 |
| 1987 | 1 | 2 |
| 1986 | 1 | 2 |
| 1985 | 1 | 2 |
| 1984 | 1 | 5 |
| 1983 | 1 | 5 |
| 1982 | 0 | 0 |
| 1981 | 0 | 0 |
| 1980 | 1 | 1 |
The 12-5 upset, while often reliable, has been overstated to some degree. A study by Benjamin Morris from FiveThirtyEight found 5-seeds under-perform expectations but not a great deal more than any other top-six seed.
But the 12-5 matchup is usually the first to include an at-large team instead of a low-seeded conference champion, and the gap between fours and fives is historically larger than the gap between fives and sixes, so it's really the first reasonable upset on the board.
The absence of such an upset this year doesn't mean the trend has ended or the committee is seeding differently or anything like that. It's one year and as much as the last three years don't show that 12-seeds are actually better than 5-seeds, a single-year sample doesn't tell us much at all.
But it probably means you called at least one upset incorrectly.