May 9, 2025
PRINT
What the Shift to NPI Means for the NCAA Regionals Debate
by Adam Wodon/Managing Editor (@chn-adam-wodon)
After putting it on hold last year, the NCAA Men’s Ice Hockey Committee is planning to replace the Pairwise/RPI system with “NPI” starting next season.
I’ve probably spilled more virtual ink writing about the Pairwise in the last 30 years than anyone since the Benadictine monks transcribed bibles.
But I won’t be sad to see it go.
Not because it wasn’t a good system anymore — it was still doing fine. It’s just that, it’s boring now. There were so many changes made over the years, that essentially what we know as the Pairwise, became just the RPI (Ratings Percentage Index). The remaining components of the Pairwise were like Milton Waddams in Office Space, hanging around in the basement with his stapler for years, not realizing he’d been fired long ago.
The days of Pairwise craziness were exciting, breaking down the nuances, trying to talk fans off the ledge. Oh the fun we had with “TUC cliffs,” “bonus points,” the Pairwise Grid, “the Bentley effect,” and the list goes on. Let’s not even talk about how many times in the early days I wrote articles explaining to people — coaches, even — that “Yes, the Pairwise absolutely IS how the NCAA Tournament field is selected. I swear.”
So farewell old friend. You done good.
As it stands right now, moving to NPI is not that drastic a change. It’s still a method of taking a team’s winning percentage, and adjusting for strength of schedule. It’s good to see that latter part, however, done in a more iterative way, akin to KRACH (which the CHN Power Ratings are based on). I was calling for a change to KRACH as far back as 2004.
Personally, I would still like to see a “recency bias” in the formulation. The Pairwise once had a “Last 16” criteria, but it was scrapped long ago. There’s room for tweaks. But NPI will be fine, like Pairwise was fine. Certainly better than the so-called smoke-filled room, that no one wants.
For the nerds like me, there will be plenty of time to get into the weeds of how the NPI formula works in the coming months. But we’ll leave the math to another day.
For now, I’ll just tie this in to my other pet topic (which is also getting old and boring) — the NCAA Regional debate.
* * *
As the debate continues, I respect the college hockey people who want to see home-ice Regionals, and I don’t disagree with many of their points. But at the end of the day, I continue to believe that no formula is precise enough to use it for giving such an advantage.
The current neutral-site format eliminates that concern.
Yes, there are certainly issues with the current Regional format that we all know about, and are discussed ad nauseum. My continued belief, as I’ve also written plenty about, is that we can improve the current format, without having to switch to a home-ice format.
Switching to NPI, and seeing the difference with Pairwise, is a perfect opportunity to demonstrate the point.
If you run the calculation, the NPI and Pairwise are fairly close, but with enough movement here and there to affect which teams would get home ice.
And using one over the other is a choice, about as subjective as anything else.
The decision of “which formula to use” only underlines the imprecision that will determine home ice. We are not talking about standings points. Those are cut and dried. You know what it is. Schedules are relatively balanced.
Not so with these algorithms. Yes, everyone knows the formula in advance of the season, but that doesn’t eliminate the point.
Even beyond the Pairwise vs. NPI difference, the Committee is still tinkering with what home-road weightings to use, what overtime weightings to use, and so on. Those decisions are also applicable to the Pairwise/RPI. There’s not a lot of math involved in those decisions. It’s based upon what feels right. What seems right. The six current members of the Committee are going to look at different tweaks to the algorithm — whether that be RPI or NPI — and pick the one that passes their own personal smell test.
And no disrespect is intended in that remark. I believe their decisions will be based upon sound, rational, good-faith thinking. But it will be fairly subjective thinking, nonetheless.
Mathematically speaking, you may be able to make an argument that NPI is “better” than Pairwise. But there’s no real math argument to the other factors.
* * *
Even if you concede that atmospheres would be better with home-ice NCAA games, there is still the issue of fairness.
You can’t get 100 percent perfection in atmosphere or fairness. You have to weigh which is more important to you, and what your definition of “atmosphere” and “fairness” are, which could be credited to either system in varying degrees, if you wanted it to.
Obviously, I fall on the side of “fairness,” even with — yes, I know — the obvious unfairness of certain things in the current neutral-site system (which I adamently want to fix as well).
Whenever I have written about these points in the past, inevitably I get people angrily asking me some variation of: “Hey, genius, if the Pairwise/RPI/NPI is good enough to pick the field, why isn’t it good enough to pick home ice” — although they say it with less perfect grammar.
Well, I’ve addressed that question about a bazillion times (scroll down in that link), including in articles and numerous podcasts. But I’ll repeat things again in a nutshell:
Using an imperfect formula to choose the field is a better way than the alternative (smoke-filled rooms), 100 percent.
But using an imperfect formula to decide home ice for Regioanls is not better than the alternative — neutral sites. So, given the options, I go with neutral sites.
Anyway, I kinda wish we’d just stop talking about this already, but I doubt we will.
E96 $H65:D9 42A:E2= 2?5 E96 s2?:D9 4:EJ @7 w6C?:?8]k^Am
