"My rule was I wouldn't recruit a kid if he had grass in front of his house.
That's not my world. My world was a cracked sidewalk." —Al McGuire

Friday, February 18, 2022

Predicting the Top-16 Reveal

Marquette was last included in the 2019 Top-16 Reveal
Screenshot from CBS Sports

This Saturday will bring the NCAA's official Top-16 Reveal. When even the Selection Committee is doing bracketology, it means we're getting to the serious part of the season. We decided to dig into the history of the Reveal, using all of the data from the first five Reveals, including both NET and RPI data as both were the relevant metrics of their time. First, here are the teams year by year, in descending order from the Top Overall Seed at the Reveal to the 16th and final Revealed team. The teams below the gap are teams that were ranked in the Top-16 of the current metric at the time but were left out:

2021 Reveal NET 2020 Reveal NET 2019 Reveal NET 2018 Reveal RPI 2017 Reveal RPI
1 Gonzaga 2 Baylor 2 Duke 3 Virginia 1 Villanova 2
2 Baylor 1 Kansas 4 Tennessee 4 Villanova 2 Kansas 3
3 Michigan 3 Gonzaga 3 Virginia 1 Xavier 3 Baylor 1
4 Ohio State 7 San Diego State 1 Gonzaga 2 Purdue 9 Gonzaga 11
5 Illinois 4 Duke 6 Kentucky 5 Auburn 5 UNC 4
6 Villanova 9 Dayton 5 Michigan 6 Kansas 6 Florida State 6
7 Alabama 10 Louisville 7 UNC 8 Duke 8 Louisville 5
8 Houston 5 West Virginia 8 Michigan St 9 Cincinnati 11 Oregon 10
9 Virginia 6 Maryland 10 Purdue 11 Clemson 4 Arizona 9
10 West Virginia 17 Florida State 14 Kansas 18 Texas Tech 14 Virginia 14
11 Tennessee 11 Seton Hall 13 Houston 7 Michigan St 21 Florida 7
12 Oklahoma 20 Villanova 15 Marquette 21 UNC 13 Kentucky 12
13 Iowa 8 Auburn 16 Iowa State 13 Tennessee 10 Butler 8
14 Texas Tech 15 Oregon 19 Nevada 14 Ohio State 22 West Virginia 36
15 Texas 21 Butler 11 Louisville 15 Arizona 15 UCLA 21
16 Missouri 34 Michigan St 12 Wisconsin 12 Oklahoma 19 Duke 17
  Colorado 12 Arizona 9 Virginia Tech 10 Rhode Island 7 Xavier 13
  Loyola Chi 13     Texas Tech 16 Nevada 12 Cincinnati 15
  Colgate 14         Seton Hall 16 Creighton 16
  USC 16                

Here are our key takeaways:

The Reveal Reinforces the Current Metric: The first thing that stands out is the number of teams ranked in the top-16 of the RPI/NET that are also in the Top-16 Reveal. 67/80 (83.8%) teams in past reveals were in the top-16 of the current metric. When you go outside the top-16, 78/80 (97.5%) were still in the top-22, so it's rare for teams to show up below that ranking. In the instances teams outside did show up, they were massive outliers in the closest alternate metric. For RPI #36 West Virginia, given a 4-seed in 2017, they were ranked #3 in kenpom at the time. For NET #34 Missouri, given a 4-seed in 2021, they were ranked #2 in RPI at the time. If you aren't in the top-22 of the current metric, you had better be elite in the companion metric.

The Top Two Seeds Lines Are Top-11 Teams: This has happened without fail. There have been a handful of double-digit teams to crack that top-8, but none lower than 11th. Worth watching that NET as there are some outliers perceived to be fighting for the 2-line. The Top-11 teams are also pretty safe when it comes to inclusion. 52/55 (94.5%) teams ranked in the top-11 were included in the Reveal, whereas just 15/25 (60%) of the teams from 12-16 were included. This calculus played in when evaluating Villanova, Texas Tech, and Duke for the final 2-seed.

Only Elite Mid-Majors Need Apply: Only 10/80 (12.5%) teams from outside the traditional Top-6 conferences have been included in the Reveal. 6/10 (66.7%) were ranked in the top-5 of the current metric. The exceptions had immaculate records, including 2017 Gonzaga (#11 in RPI, 25-0), 2018 Cincinnati (#11 in RPI, 22-2), 2019 Houston (#7 in NET, 22-1) and 2019 Nevada (#14 in NET, 22-1). If you had 3+ losses outside the Top-6 leagues, you were left out. And obviously Colgate in 2021, who was a COVID scheduling NET outlier.

Replacement Teams Are From Top-6 Leagues: There have been 13 teams from outside the Top-16 of the current metric to be included and all of them came from one of the traditional Top-6 leagues. Don't expect the Selection Committee to look to a mid-major if they throw a curveball.

We have a little discussion on our selections for the Top-16 Reveal. In accordance with past history, Gonzaga is the only team from outside the traditional Top-6 leagues. 14/16 teams selected were also in the Top-16 of NET, with the exceptions being #20 Wisconsin and #31 Providence. Here is commentary on a line-by-line basis:

1-Seeds -- Gonzaga, Arizona, Auburn, Kansas: The first three seem clear. Gonzaga is the top overall seed because despite the resume, their quality metrics are by far the best in the field. Both Arizona and Auburn seem like obvious picks, which left four teams for one spot. Kansas has the best average of Resume and Metric rankings and leads the nation in Q1+2 wins, which allowed them to edge out Baylor, Kentucky, and Purdue for the final 1-seed.

2-Seeds -- Baylor, Kentucky, Purdue, Villanova: The first three feel obvious and would more likely show up on the top line than the line below. Villanova gets the last 2-seed due to their #5 NET ranking and a better Resume+Metric Average than any team below them.

3-Seeds -- Texas Tech, Duke, Tennessee, Wisconsin: Tech and Duke were considered for the 2-line, but the metrics just didn't justify moving them up despite excellent quality wins. Tennessee's win over Kentucky helped solidify their selection here and while their Q1 isn't as good as some others, they have no losses outside Q1A. Wisconsin was on our 4-line previously, but with their NET moving up to 19 after the Indiana win, there is now precedent for them being placed this high (2021 Oklahoma, 2019 Marquette, and 2018 Michigan State were all ranked lower) and their 8 Q1 wins are second only to Baylor.

4-Seeds -- UCLA, Illinois, LSU, Providence: Both UCLA and Illinois seemed like they had to be included. Both have the numbers and enough Q1 volume to warrant inclusion without any truly bad losses. LSU does have a Q3 loss but we moved them up because no one around them has a resume that is significantly better enough without some serious blemishes. Providence sticks as the last team in. While their NET ranking doesn't seem to warrant inclusion, their single digit resume average is good enough that it seems likely they'll be included. PC fans have long said their is no comparison to their resume, but I would point to 2017 West Virginia. While they had diametrically opposite resumes (WVU had monster wins and narrow losses, PC has narrow wins and big losses), both excelled in the metric that was not the NCAA's chosen of the day. On the team sheets, WVU had the aforementioned kenpom rank of #3 while Providence has an average resume metric of 8.5.

Also considered -- Alabama, Texas, Houston: The Selection Committee always discusses a few teams that are just on the outside looking in. These three all have pros and cons. Alabama has a brilliant top-end resume and SOS, but their two Q3 losses and borderline metrics weren't good enough to put them in. Texas has similar quality and is ranked #15 in the NET, but too much of their record is puffed up by Q4 games. Houston has excellent numbers, but despite the #4 NET (which would be the highest ever left out of the Reveal) they have zero Q1 wins, zero wins over the field, and too many total losses to be included from outside a power conference.

And last...Marquette: The Golden Eagles had a legitimate chance to get in this discussion, but the loss to Butler and narrow win over Georgetown dropped their NET from 24 to 32. Had they beat Butler and held their 26-point lead against the Hoyas, they may have been able to get to the 4-line or at least been one of the alternates considered, but instead they are heading in the other direction. For now they hold on to our last 6-seed thanks to their Q1 wins, but it was a close evaluation between them and St. Mary's (the Gaels' 2-6 Q1 record just wasn't good enough). A win at Creighton so their Q1+2 record doesn't fall below .500 would be strongly advised.

Here's the entire S-Curve:

1-Seeds: 1-GONZAGA 2-ARIZONA 3-Auburn 4-Kansas


3-Seeds: 9-Texas Tech 10-DUKE 11-Tennessee 12-Wisconsin

4-Seeds: 16-Providence 15-Lsu 14-Illinois 13-Ucla

5-Seeds: 17-Alabama 18-Texas 19-HOUSTON 20-Ohio State

6-Seeds: 24-Marquette 23-Michigan State 22-Usc 21-Connecticut

7-Seeds: 25-St. Mary's 26-Arkansas 27-COLORADO STATE 28-Boise State

8-Seeds: 32-Seton Hall 31-San Francisco 30-Iowa State 29-Xavier

9-Seeds: 33-Miami 34-MURRAY STATE 35-Wyoming 36-Tcu

10-Seeds: 40-Memphis 39-Notre Dame 38-Wake Forest 37-Iowa

11-Seeds: 41-Davidson 42-Creighton 43-Byu 44-Michigan 45-Kansas State 46-San Diego State







Last Four Byes: Notre Dame, Memphis, Davidson, Creighton

Last Four In: Byu, Michigan, Kansas State, San Diego State

First Four Out: Indiana, North Carolina, SMU, Rutgers

Next Four Out:  VCU, Oklahoma, West Virginia, Belmont

Multibid Leagues

Big East: 7

Big 10: 7

Big 12: 7

SEC: 6

ACC: 5

MWC: 4

WCC: 4

Pac-12: 3

American: 2

Atlantic-10: 2

No comments: