Before this past weekend’s series against Connecticut, the Massachusetts hockey team sat firmly 19th in the NCAA Power Index (NPI). NPI is the mathematical system used by the NCAA to select and seed teams for the NCAA Tournament. NPI is primarily based on win percentage and opponent win percentage, with road wins weighted slightly more heavily than home wins. Rankings used in this article are NPI rankings unless otherwise noted.
The Minutemen (9-12-1, 12-9-1 Hockey East) traveled to Storrs on Friday and dominated the Huskies (17-10-5, 11-8-3 HEA) 5-1, who were No. 14 entering the matchup. The following game at home was a 3-3 tie that UMass won in a shootout. The Minutemen’s reward for taking five of six Hockey East points against a team ranked higher than them? Dropping to 20th in NPI.
I believe that NPI does as good a job at evaluating how good NCAA hockey teams are as if I were playing first-line right wing for the Colorado Avalanche.
The Minutemen’s NPI value did go up by 0.41 this weekend, which was the 17th biggest rise of the weekend. So NPI is stating that there were 16 teams in the country that had better weekends than beating UConn away and tying them at home.
Some ratings I agree with, like Ohio State, which beat and tied the then No. 1 Michigan State in East Lansing and went up a huge 1.02 NPI value. Other teams like Maine and Boston University also deserve to be up there with sweeps of strong Hockey East teams.
Other teams that are ahead of UMass make no sense whatsoever. Take Rensselaer Polytechnic Institute, with a win over No. 59 Brown and a win over No. 57 Yale.
The question I pose is, does that deserve to be valued as a better weekend than what the Minutemen accomplished? One might say that they won both their games, so it should be valued slightly better than UMass’. I would disagree with you.
I argue that the weekend that the Engineers (11-22-1, 8-13-1 ECAC) had should not be valued more than twice as much as the Minutemen’s. Rensselaer went up 0.83, which is 0.42 more than UMass, for a weekend I would say was not as good as the Minutemen’s.
It’s not just the Engineers’ weekend that was unfairly valued. Sacred Heart went up 0.77 for two one-goal victories over No. 45 Army.
The one that makes no sense to me is Ferris State. The Bulldogs (7-22-5, 6-18-2 CCHA) played twice at No. 51 Lake Superior State, resulting in a 5-2 win and a tie, which the Lakers (11-20-3, 8-16-1 CCHA) won in a shootout. They had essentially the exact same weekend that UMass had, with the second game being played on the road instead of at home. Plus, Lake Superior State is a considerably worse team than the Huskies.
In my mind, the weekend the Minutemen had is infinitely better than the one Ferris State did, but the NPI calculations didn’t agree. NPI has the Bulldogs rising 0.69, decidedly more than UMass. How does this happen?
At its core, it seems the NPI incorrectly values different conferences based on conference “atmosphere” — the spread of quality teams within a conference — rather than conference ability as a whole. Take the CCHA, the most extreme example of how your conference atmosphere can boost the entire conference’s NPI.
The CCHA has only nine teams. If you scan the CCHA standings, the top five teams are separated by just two points. Minnesota State sits on top with 51 points. Both Michigan Tech. and Bowling Green are at 49, in fourth and fifth place respectively.
The other four teams in the conference are a notable step down, creating a clear divide: five good teams and four weaker teams. Throughout the season, the good teams have split series with each other and tied a bunch of games, with no team ever breaking away, making it impossible to separate the teams by much in NPI.
They also all beat up on the four lesser teams, bumping their NPI up each time. When one of the teams in the top group beats a team in the weaker group, it boosts the entire top group because NPI has very little to separate them. Because they have only nine teams, they play each other and the lesser teams more often, giving them more opportunities to boost the conference.
This doesn’t happen in Hockey East. The entire field is close, from second to ninth in Hockey East separated by the same number of points as fifth and sixth are separated in the CCHA, 19. This is reflected in the NPI, and the entire conference is punished for being competitive.
The CCHA has been boosted by NPI at an unbelievable rate. They have four teams in the top 20 and one at 22nd. That’s the same number of top-20 teams as both the Big 10 and NCHC. This would all make a lot of sense if the CCHA were a comparable conference to the other top conferences, but it’s not.
In fact, the only conference that the CCHA has a winning record against in non-conference games is the ECAC. Its 2-8-5 against the Big 10, 11-17 versus the NCHC,1-4 against the AHA and most importantly for these comparisons, a winless 0-5-1 against Hockey East. The CCHA doesn’t win out of their conference so why are they being rewarded as if they are an elite conference?
For some reason, NPI is a great deal more lenient to the CCHA than HEA. Following the Minutemen’s tie, their NPI value dropped by 0.11, so a tie with No. 15 as No. 19 drops you by 0.11.
What would you expect for No. 17 St. Thomas after tying No. 46 Bemidji State that same night? A 0.2 decrease? 0.4? Wrong. The Tommies dropped by just 0.028 for tying with the Beavers, far less than UMass did for tying No. 15. I think that that is just wrong, plain and simple.
Even when you widen the scope out to the entire season, NPI still manages to improperly evaluate these teams.
Augustana is the current highest ranked CCHA team at No. 14. The Vikings have a very decent resume: a win over No. 10 Minnesota-Duluth, No. 16 Minnesota State and a great conference record. But if you compare theirs to the Minutemen’s, it’s clear that they don’t belong six spots ahead.
UMass’ best three wins on the year were at No. 6 Providence, home versus No. 9 Cornell and at No. 15 UConn. Its worst three losses on the year were at No. 48 UMass Lowell, at No. 45 Army and at No. 41 Omaha.
Augustana’s best three wins on the year were at No. 10 Minnesota-Duluth, home versus No. 16 Minnesota State and at No. 17 St. Thomas. Its worst three losses on the year were at No. 56 Ferris State, at No. 51 Lake Superior State and at No. 46 Bemidji State.
So, at the top and bottom end, UMass is clearly better. Even the average rank of opponents in UMass victories is 35.7 against Augustana’s 42.6. In losses, the Minutemen’s average rank is 25.8 and the Vikings’ is 30.7. Even for the overall strength of schedule, the average rank of opponents UMass faced is 31.3, while Augustana’s is 37.2.
Augustana is the best of what the CCHA have to offer according to NPI, and yet UMass clearly has the better resume. Why should we use a system that has four teams from the CCHA ahead of the Minutemen when it’s not even clear that its best team is better?
There are many reasons why the NPI doesn’t make sense for ranking collegiate hockey, and it all starts with the move away from Pairwise.
Myles Donato can be reached at [email protected] and followed on X @myles_donato.
