NCAA Power Index

The NCAA DIII Championship Committee met on March 19^th and they made a recommendation that PairWise, otherwise known from this point forward as the NCAA Power Index (NPI), should be adopted as the selection metric for at-large bids across all team sports. The recommendation states that this should begin this fall (2024).

Within the rationale given, the committee found that when set to the “default” setting of 75/25 that NPI produced a similar championship field as to what the selection committees produced. The Championship Committee suggested that changes to the parameters or dials (as they call them) be considered after the first year of use but after that only at set intervals. They also need to look at whether each sport will share the same dials or if each sport can have their own settings.

Another recommendation that doesn’t impact women’s volleyball is that an access ration of 1:6 be used in team sports to determine the size of the NCAA Tournament. In other words, a bid should be available to every 6^th team that participates in the sport. This would impact women’s volleyball except that they will continue to cap the bracket at 64 teams.

I have previously written about PairWise (or NPI), which you can find here. In summary, I don’t like this direction. I don’t feel it’s a better system than what we have although it will be much easier to implement. Everyone will also see the results every week as opposed to the last couple of weeks of the season. I do believe with the 75/25 setting that Strength of Schedule (SOS) becomes less important for selection and that we’ll see entirely undeserving teams get at-large bids. My suggestion would be to use NPI for 80% of the selections and then use it and human interaction for the last 20%. In women’s volleyball, this would typically mean 16 at-large bids determined by NPI with 4 going through human eyeballs.

The meeting notes mentioned that they looked at 2021 and 2022 data. Through the excellent work of my Latvian Hackers, I was able to access the data for 2019 and 2021 so let’s see how the brackets would have differed. (I will be having a hard conversation later with Anastasija as to why I don’t have the 2022 data.)

In 2019 there was one Pool B bid and 20 at-large bids (Pool C). There was no mention of how the Pool B (Independent Teams) bids would be determined but I assume it would be through NPI. I guess they could also eliminate the Pool B bid altogether but I doubt that happens.

The team back in 2019 that I thought would get an at-large bid that didn’t was Mount Union. The selection committee agreed on 19 teams with me but took Ithaca instead of Mount Union. NPI had Mount Union as the 25^th at-large team and Ithaca as the 23^rd at-large team. So, the 2019 NPI results would have neither team getting a bid and would further eliminate UW-Eau Claire (21^st at-large team) and Wittenberg (30^th at-large team). Replacing Ithaca, Wittenberg and UW-Eau Claire would have been MIT (15^th at-large team), Franklin & Marshall (16^th at-large team) and Chapman (18^th at-large team). One important note about those three eliminated teams was that they were strong SOS teams. In addition, only UW-Eau Claire had 20 wins. It was an interesting year because Wittenberg and Ithaca were teams that played outstanding schedules but didn’t have a lot of wins from those matches. Wittenberg only had 15 wins but had beaten 4 teams ranked ahead of it by NPI. I really didn’t like Ithaca that year because they only had one win against a quality opponent. They got their bid based on losing to a bunch of really good teams. That’s not a recipe for success with NPI. One thing to remember about NPI is that it eliminates regional rankings. There is no longer a case where a team blocks another from being considered. Because of that, some of the differences between what NPI would have selected and the selection committee are due to this reason. When looking at the three teams NPI favored, MIT had only one win over a team ahead of it, Franklin & Marshall none and Chapman one. Teams not selected were UW-Eau Claire (4), Emerson (3), Mary Hardin-Baylor (3) and Mount Union (4). What did these three teams have in common in addition to not being able to beat good teams? All had 20 or more wins.

Looking at the 2021 data, teams that were selected by the committee but rejected by NPI are Springfield (21^st at-large team) and Ithaca (35^th at-large team). Ithaca basically repeated their strategy from 2019 for another unexpected at-large bid. The two teams NPI favored were Plymouth State (Ugh) and UC Santa Cruz. UCSC was their 16^th at-large team while Plymouth State was 18^th. If I remember correctly, UC Santa Cruz was blocked by the regional ranking ordering (no one check me on that). Plymouth State was a team that played absolutely no one and didn’t have a significant win from their 22-win total. UCSC was 17-3 that year but neither team beat someone higher ranked by NPI. Springfield was 3-3 against teams ranked higher and even Ithaca was 4-12. That year I liked Denison (19-7 and 2-3 against higher ranked teams), Muskingum (17-10 and 3-7) and UW-Oshkosh (19-12 and 2-6). Again, you see the issue with record among these teams.

It should be pointed out that the 75/25 “default” dial relates to SOS and record. It can be changed to favor SOS more and that might be a partial solution but just from these examples you can see what kind of issues will be coming our way in November. Even with a different dial, there will be problems. Back to my 80% solution and that wouldn’t haven’t resolved MIT, Franklin & Marshall or UCSC. It would have only hindered Chapman and Plymouth State.

Here’s the thing. I like teams that play tough schedules, beat a few tough teams, don’t trip up against weaker opponents and show me they can score (Points/Set). You’ll notice that doesn’t jive with the current selection criteria. It also doesn’t seem to jive with NPI. In the end, what we have is new selection criteria without it being spelled out. We simply have a formula. Coaches will need to figure out how to game this new formula, which I’m confident they will do. The teams that play the tough schedule and win will not have an issue but the 6 spots on the backside of the at-large bids won’t necessarily go to the best teams. Maybe that also doesn’t happen today with the selection committee, but at least there is someone that can explain the reasoning. (Of course, they don’t but they could.)