One month away from Selection Sunday, the NCAA tournament committee revealed a sneak peek at its top 16 seeds during the season. A lot can change between now and March 11, but we know No. 2 Virginia, despite a one-point overtime loss at home to Virginia Tech on Saturday, is the top overall seed, projected to be in the tournament’s South region. No. 6 Cincinnati (2-seed), No. 4 Michigan State (3-seed) and No. 15 Tennessee (4-seed) are also projected to share that region when the madness starts in March.
No. 1 Villanova was rewarded with a top seed, No. 1 in the East, as were No. 5 Xavier (top seed in Midwest) and No. 3 Purdue (top in West). The rest of the released bracket was as follows, with AP rankings in parentheses:
|Seed (overall seed)||South||East||West||Midwest|
|1||Virginia (1)||Villanova (2)||Purdue (4)||Xavier (3)|
|2||Cincinnati (8)||Duke (7)||Kansas (6)||Auburn (5)|
|3||Michigan State (11)||Texas Tech (10)||North Carolina (12)||Clemson (9)|
|4||Tennessee (13)||Ohio State (14)||No. 5 Arizona (15)||Oklahoma (16)|
There are a couple of takeaways from this, but the most harrowing is the RPI appears to have too much influence in the seeding process.
Over the summer, the NCAA Division I Men’s Basketball Committee altered the definition of a quality win, placing more emphasis on winning away from home. For example, if a team beats a squad ranked in the top 30 of the RPI at home, that is considered a Group 1 win, the same as a victory on the road against a Top 75 RPI team or a Top 50 RPI team on a neutral court.
|1||1 to 30||1 to 50||1 to 75|
|2||31 to 75||51 to 100||76 to 135|
|3||76 to 160||101 to 200||136 to 240|
|4||161 to 351||201 to 351||241 to 351|
Oklahoma has six Group 1 wins — home wins over Kansas and Texas Tech, road and neutral-site wins over Wichita State and USC, plus a two-game sweep of TCU — which, according to WarrenNolan.com, is more than 10 of the other teams in the top 16.
Kansas is 19-6 (8-4 in conference play) and trails Texas Tech for the Big 12 lead by a game. Plus they have home losses to Arizona State, Texas Tech and Oklahoma State, a loss to Washington on a neutral court plus just got handed a 24-point loss at the hands of Baylor, a team that was 4-7 in the Big 12 entering the contest. But because the Jayhawks lead the nation with nine Group 1 wins, they were rewarded with a 2-seed in the West.
Rewarding teams with quality wins based on location isn’t an issue, but the fact that the matrix depends on the RPI to begin with is a major problem. RPI is one of the worst indicators of team strength. Last season, if you used RPI to fill out the final 63 games of the 2017 NCAA tournament, you would have ended up with 39 correct selections, the second-worst result among 75 rating systems tracked by Kenneth Massey. RPI ended up 36th out of 61 systems (38 correct) in 2016 and 40th out of 64 systems in 2015 (43 correct).
While the RPI is the metric used on the team sheets distributed to the selection committee to gauge the relative strength of teams under consideration, other metrics such as those from Ken Pomeroy and Sagarin are also available to the committee throughout the season and during selection week. However, those aren’t factored in to the Group 1 wins, even though they should be — each was better at predicting future wins in the NCAA tournament last year than RPI.
|2017 Men’s NCAA Tournament||Correct (out of 63)||Rank (out of 75 systems tracked)|
Sagarin and Pomeroy agree on 13 of the 16 teams that deserve to be considered among the top 16 seeds in this year’s tournament with Auburn, Tennessee, Clemson, Wichita State, Creighton and Arizona only appearing in one of those rankings’ top 16 slots.
|2017-18||RPI rank||Pomeroy rank||Sagarin rank|
|North Carolina (19-7)||9||10||11|
|Texas Tech (21-4)||12||8||10|
|Michigan State (24-3)||14||6||4|
|Ohio State (22-5)||16||11||15|
|Wichita State (19-5)||19||18||13|
|West Virginia (18-7)||34||16||8|
Gonzaga is perhaps the biggest snub of the group. Ranked seventh by both metrics, the Bulldogs are also considered the 11th-best team in the nation based on the composite ranking of 31 different methods. Gonzaga also outscores opponents by almost 20 points per game (12th best) after adjusting for strength of schedule.
Last year, the seventh-best team according to Pomeroy was West Virginia, a 4-seed, who, coincidentally, lost to top-seeded Gonzaga in the Sweet 16. Sagarin’s No. 7 team leading into the 2017 tournament, Florida, made it to the Elite Eight. In other words, if the committee is going to discount teams like Gonzaga on Selection Sunday, it’s worth remembering them for an upset or two in your bracket.