NCAA Tournament Seeds: Do the Metrics Agree with the Committee?

Last March, I took a look at how closely the Committee's seedings coordinated with three advanced ratings indices: (1) RPI, (2) Ken Pomeroy's ratings, and (3) Jeff Sagarin's ratings in an effort to see how closely the Committee lined up with the advanced consensus.

That was fun. So let's do it again.

Last year, the most egregiously misseeded teams all played to form except for Northwestern. To wit:

• #10 Wichita State (should've been #5) beat #7 Dayton, then narrowly lost to #2 Kentucky (65-62);
• #6 Maryland (should've been #10) lost to #11 Xavier 76-65;
• #9 Seton Hall (should've been a #1 in the NIT) lost to #8 Arkansas 77-71;
• #11 Providence (should've been a #3 in the NIT) lost to #11 USC 75-71 in the First Four;
• #5 Minnesota (should've been #8) lost to #12 MTSU 81-72;
• #8 Northwestern (should've been #11) beat #9 Vanderbilt 68-66, then narrowly lost to #1 Gonzaga 79-73; and
• #9 Virginia Tech (should've been #12) lost to #8 Wisconsin 84-74.

Here are the findings with a team's actual seed listed first and its expected seed listed second. The list covers all teams from Virginia through Syracuse; San Diego State through Texas Southern will not be included as automatic qualifiers at the bottom of the bracket.

CORRECTLY SEEDED (15)
Virginia (1), Villanova (1), North Carolina (2), Purdue (2), Cincinnati (2), Michigan (3), Gonzaga (4), West Virginia (5), Clemson (5), Ohio State (5), Houston (6), TCU (6), Texas A&M (7), Nevada (7), Creighton (8)

MISSED BY ONE (14)
Kansas (1 - 2), Duke (2 - 1), Michigan State (3 - 2), Tennessee (3 - 4), Wichita State (4 - 5), Arizona (4 - 5), Kentucky (5 - 4), Florida (6 - 7), Arkansas (7 - 8), Seton Hall (8 - 7), Florida State (9 - 10), Texas (10 - 11), UCLA (11 - #1 NIT), Syracuse (11 - #1 NIT)

MISSED BY TWO (8)
Xavier (1 - 3), Texas Tech (3 - 5), Auburn (4 - 6), Miami (6 - 8), Butler (10 - 8), Oklahoma (10 - #1 NIT), St. Bonaventure (11 - #2 NIT), Arizona State (11 - #2 NIT)

MISSED BY THREE (6)
Rhode Island (7 - 10), Virginia Tech (8 - 11), Kansas State (9 - 12), Providence (10 - #2 NIT), Alabama (9 - #1 NIT), NC State (9 - #1 NIT)

MISSED BY FOUR (0)

MISSED BY FIVE (1)
Missouri (8 - #2 NIT)

Here are some additional quick hits of note:

1. The hilarious seeding saga of Wichita State gets another entry. Last year, the Shockers should've been a 5. This year, the Shockers should be a 5. Their actual seeds? A 10 and a 4. Ha!

2. Word on the street is that the NCAA had it in for USC, Louisville, and perhaps even Oklahoma State, wanting to keep them out of the Tourney due to recent investigations. It doesn't hold water for Okie St (expected 15 seed), but it sure does for USC (9) and especially Louisville (8). There was no notable outlier rating for either Louisville or USC, either. They were straight up snubbed.

3. Saint Mary's also got the shaft (expected 9 seed).

4. A few automatic qualifiers to keep an eye on:

• Loyola (IL) received an 11 seed as the top auto-bid among the non-majors. Their resume would've given them...an 11 seed, even as an at-large. Look out, overrated Miami.
• Davidson and San Diego State are both properly seeded as 12 seeds.
• New Mexico State also got a 12 seed and earned a 14 seed based on its resume. Quite strong.

5. In addition to USC and Louisville as mentioned above, here are the best consensus teams to miss the field:

• Middle Tennessee State would've earned a 13 seed in a pure meritocracy.
• After Louisville, USC, and Saint Mary's, Penn State and Notre Dame were the next best teams to miss the Tournament. Both should've been on the 12 line.
• Maryland and Marquette both would've been 13 seeds in a pure meritocracy.

6. I've mentioned outliers a bit above, but they bear additional mentioning here:

• RPI absolutely ruined the chances of both Penn State and Notre Dame. Nittany's RPI was 69th; the consensus of their KenPom and Sagarin ratings would've put them at 33.5 on the 9 line. Notre Dame's RPI was 81st (!) but their KenPom/Sagarin consensus was 28th, on the 7 line. Tough for the Irish.
• Baylor experienced a similar fate: 100th in RPI but consensus 32 (8 seed) elsewhere.
• Rhode Island should've been a 4 seed according to RPI, but on the bubble according to both other metrics. Fortunately for them, Oklahoma isn't any better, thus, the big winner here is Duke who figures to enjoy a smooth path to the Sweet 16.
• There may be something funky going on with RPI and the SEC. LSU and Mississippi State posted notably similar scores in KenPom (61st, 62nd) and Sagarin (70th, 65th), but both schools fell way down the list in RPI (92nd, 91st). The similarity of their scores in all three systems makes the RPI score stand out even more.

7. Finally, a nod to the teams with the greatest consensus across the three ratings systems:

• Unsurprisingly, the teams with the least variation were Virginia and Villanova. Virginia nabbed the top spot twice and one second place finish while Villanova took the opposite spot each time. Let's call this one point of variation for each.
• The only other squad with just one point of variation? Michigan, who came in 11th, 10th, and 11th in RPI, KenPom, and Sagarin, respectively.
• Teams with only two or three points of variation: Duke (duh), Kansas (OK), Purdue (sure), Cincinnati (fine), and...Seton Hall? The Pirates were 28th in RPI while coming in 26th in both of the other metrics. Weird but cool.
• Remarkably, UCLA came in at 45th, 48th, and 44th across the three systems. Nobody else has that type of consistency so far down the chart except for Marquette (50th, 53rd, and 51st) who was left out in the cold.