Again, they use resume metrics to pick the field, and predictive to seed.
I get that and I'm saying there are situations where they need to look at that a little deeper. The problem is that like a lot of mathematical formulas the WAB and SOR get weird at the margins. When talking about bubble teams the underlying problem is that you are comparing apples an oranges. Resume metrics attempt to bridge that gap. So, for example, if you look at Auburn (#9 SoS per KenPom, 17-16 on Selection Sunday) and compare them to say Iowa (#28 SoS per KenPom, 21-12 on Selection Sunday) the metric works. As I see it the question is (or at least should be) what is harder:
- Going 17-16 on Auburn's schedule, or
- Going 21-12 on Iowa's schedule?
Auburn's schedule was tougher but Iowa's record is better. I'd argue that one for Iowa. Both went 0-fer against high end OOC opposition and Iowa lost at ISU (#2 seed, still in) while Auburn had the aforementioned four losses to high end teams. The metric works well because you are comparing reasonably comparable schedules and reasonably comparable records.
The problem is at the margins. It is a lot harder to compare Auburn (#9 SoS per KenPom, 17-16 on Selection Sunday) to Miami (#251 SoS per KenPom, 31-1 on Selection Sunday). With Iowa/Auburn you are comparing a slightly better record against a slightly tougher SoS. With Auburn/Miami you are comparing a much better record against a vastly easier SoS. The metrics may say it is Miami but they are flat out wrong. They are wrong because what did Miami accomplish that Auburn didn't? Let's look:
- Each had one loss in the bottom two quadrants: Auburn at 6-0 in Q4, 4-1 in Q3, Miami at 15-1 in Q4, 10-0 in Q3
- Each had three wins in Q2: Auburn at 3-2, Miami at 3-0
- While Auburn wasn't good against the best, they at least tried and had SOME wins in Q1: Auburn at 4-13, Miami at . . . well the RedHawks didn't see a Q1 game before Selection Sunday.
If your metric thinks that a team with a dreadful Q4 loss and ZERO offsetting Q1 wins should get in over a team that has Q1 wins and does NOT have Q4 losses then your metric is wrong.
Then they moved down when they used the predictive metrics to seed them.
Bringing up KenPom and any of those other predictive metrics is pointless in bringing up the team's merits of getting in. If Miami not only got in, but was a 10 seed or something
I already can make the seeding argument because the plain fact is that Miami was MASSIVELY overseeded as a #11. Here are the NET and KenPom rankings for the #11 and #12 seeds (sorted by NET and NET is first followed by team, then KenPom, then Torvik then seed):
- #36 NCST, 36, 41 - 11 Seed
- #37 SMU, 48, 48 - 11 Seed
- #42 Texas, 31, 38 - 11 Seed
- #43 VCU, 42, 46 - 11 Seed
- #45 USF, 47, 51 - 11 Seed
- #54 Akron, 70, 71 - 12 Seed
- #56 McNeese, 65, 67 - 12 Seed
- #64 Miami, 90, 87 - 11 Seed
- #72 UNI, 76, 75 - 12 Seed
- #75 High Point, 85, 86 - 12 Seed
By NET Miami is right in the middle of the #12 seeds but even that is overly generous to the RedHawks because their #90 KenPom is worse than any of the #12 seeds as is their #87 Torvik ranking. Per NET they should be a #12 seed but if you seeded based on KenPom or Torvik they'd be no better than a #13.
Just because the committee seeded them as low as they possibly could seed an at-large doesn't mean they were over-seeded. They were massively over-seeded.
The question here is how the hell did SMU get in
Maybe they did it to protect their emotional 'feel good' decision to put Miami in? SMU's NET isn't too bad but their KenPom and Torvik are weak. Not as weak as Miami, obviously but still very weak by legitimate at-large standards. Including SMU and putting them against Miami an hour from Miami's campus was awfully convenient.