The Selection Committee released their top-16 on Saturday morning, and for the second year in a row, our projections from 6 days before the release were more accurate than the day before. While David Worlock reported the Selection Committee didn't begin their meetings until Wednesday, they seemed to again disregard the results of Monday through Thursday. UNC's loss to So what were the lessons learned? Here's what we took away:
Conference Titles Matter: North Carolina was thought to be 7th, but finished 5th. In fact, the top-5 were all teams leading Power-6 conferences, and the team who was the highest compared to our projection was SEC leader Alabama. San Diego State also showed up after beating New Mexico Friday, the one result that did seem to matter from the week as San Diego State was at 19 on our S-Curve going in yet showed up 5 spots higher for the Committee.
NCSOS Matters: Iowa State's productive conference play (6 Q1 wins, including 3 Q1A on the road) wasn't enough to get them to the 2-line where we thought they deserved to be. Instead, they were the third 3-seed. This could be problematic for Big 12 bubble teams that largely played terrible non-conference schedules, and factored into the final spot in our new S-Curve.
Metrics Over Results: Tennessee ahead of Marquette despite a better collection of wins for the Golden Eagles, Alabama ahead of Baylor and Iowa State despite multiple more Q1 wins for the Big 12 teams, and Auburn and Illinois ahead of Wisconsin despite 6 Q1 wins for the Badgers and 6 combined for the Tigers and Illini. When it came to seeding, predictive metrics were highly prioritized.
Aztec Uprising: The highest NET team in the potential 6-bid Mountain West landed not just a protected seed, but 14 overall. A few things stand out. First, this likely means for the Mountain West, the Committee ignored BPI, which rates Mountain West teams significantly lower than the other predictive metric, kenpom. Second, this reinforces that NCSOS matters. The Aztecs #14 NCSOS was better than Auburn, Wisconsin, Creighton, Dayton, or Clemson that they finished ahead of.
Ultimately, this all led to a pair of major decisions. The first was a change not to make. After the UConn rout of Marquette and Purdue loss at Ohio State, many bracketologists moved UConn to #1 overall, but we're sticking with Purdue. For now, they still have the better NCSOS, metrics, and top-end results. Putting UConn ahead feels like a prisoner of the moment move, not a review of the entire season, though if the metrics did shift in UConn's favor that change would be made.
Second, on the bubble, Gonzaga moved into the field. The Committee has historically punished teams on the cut line with weak NCSOS, and Ole Miss was slated to be the last team in with a #324 NCSOS while Gonzaga is 300 spots better at #24. They still have to go through Dayton, but there's just enough there to put Gonzaga on the right side of the cut. More than other bubble teams, though, they can't afford slip-ups because their remaining conference slate doesn't give them much opportunity to improve (they really need the win at St. Mary's).
As far as how the Committee did, we try to remember that when you come up with the official seed list, you can't be wrong. Marquette at #7 overall was a surprise because they had the better body of work than North Carolina or Tennessee, but given that information and the rout at UConn, they've now fallen to #8 on our S-Curve. Alabama was even more shocking at the top of the 3 line, as was the separation between Alabama and Auburn, who came in with virtually identical resumes including splitting the season series with each other. That speaks strongly to the importance of conference championships because that's really the only place the Tide stands out over the Tigers. The same goes for San Diego State, who doesn't have the overall resume of teams they are ahead of but moved into the Mountain West champs position Friday night (likely the contingency the Committee mentioned). Finally, Illinois was a mild surprise. Every number indicated they belonged in the top-16, but so often the Committee has valued overall resumes that compare to Clemson rather than metrics with minimal top-line wins like Illinois.
Here's the full S-Curve and bracket:
Multibid Leagues
Big 12: 10
SEC: 8
Big East: 6
Big 10: 6
Mountain West: 6
ACC: 4
Pac-12: 2
WCC: 2
No comments:
Post a Comment