2026 Bracketology: The Road to the Aud

[Tom Coughlin]

I remember that! Marian sat for like a month and had already turned their gear back in and had stopped practicing.

I think the word on the USCHO forum was that the West region committee reps managed to rank Marian ahead of Utica by placing Marian fifth in the west rankings while Utica was sixth in the east rankings.

I may not be happy with how the NPI is shaping up this year, but I will take an objective measurement over that BS any day of the week.

 

[Tom Coughlin]

The 2000 team won 22 games and did not make the NCAA tournament.

That was well before my time (at least as an SNC hockey fan)

 

[WCHornet]

I think the word on the USCHO forum was that the West region committee reps managed to rank Marian ahead of Utica by placing Marian fifth in the west rankings while Utica was sixth in the east rankings.

I may not be happy with how the NPI is shaping up this year, but I will take an objective measurement over that BS any day of the week.

There is definately something amiss with the NPI when a team plays only 20 games against the 85th SOS out of 91 in the country and is sitting at number 6. This has to be addressed.

 

[JSicedog]

There is definately something amiss with the NPI when a team plays only 20 games against the 85th SOS out of 91 in the country and is sitting at number 6. This has to be addressed.

Just as big of a question mark is how Babson is sitting ahead of Aurora.
Babson 18-3-2, 80.292 wgt win%, 70/91 SOS, 0.068 QWB
Aurora 20-4-1, 77.741 wgt win%, 59/91 SOS, 0.468 QWB (3rd highest QWB)

This shows us that QWB has very little effect on the NPI, and playoff wins won't make much of a difference. Babson doesn't have too strong of a resume yet they are ahead of a team that has 4 wins over Adrian. I don't see a scenario where Salve doesn't get in.

 

[WCHornet]

Just as big of a question mark is how Babson is sitting ahead of Aurora.
Babson 18-3-2, 80.292 wgt win%, 70/91 SOS, 0.068 QWB
Aurora 20-4-1, 77.741 wgt win%, 59/91 SOS, 0.468 QWB (3rd highest QWB)

This shows us that QWB has very little effect on the NPI, and playoff wins won't make much of a difference. Babson doesn't have too strong of a resume yet they are ahead of a team that has 4 wins over Adrian. I don't see a scenario where Salve doesn't get in.

This is a bigger loophole than the smoke filled room.

 

[Tom Coughlin]

Just as big of a question mark is how Babson is sitting ahead of Aurora.
Babson 18-3-2, 80.292 wgt win%, 70/91 SOS, 0.068 QWB
Aurora 20-4-1, 77.741 wgt win%, 59/91 SOS, 0.468 QWB (3rd highest QWB)

This shows us that QWB has very little effect on the NPI, and playoff wins won't make much of a difference. Babson doesn't have too strong of a resume yet they are ahead of a team that has 4 wins over Adrian. I don't see a scenario where Salve doesn't get in.

The difference in SOS is only 0.5, while the difference in Win % is 2.5.

Win % matters above all else even though SOS gets a 75% weight in the formula because there is a much higher variation in teams' Win %.

 

[JSicedog]

The difference in SOS is only 0.5, while the difference in Win % is 2.5.

Win % matters above all else even though SOS gets a 75% weight in the formula because there is a much higher variation in teams' Win %.

You should be rewarded for beating good teams at least, if not more, the same amount as you are penalized for losing to bad teams. That is not the case. The qwb is meaningless.

 

[Lord_Zubin]

Not to make it any worse but if u flip 1 or both their losses against anna Maria or conn college they'd probably in a position for a 1st round bye ahead of hamilton.

 

[elbojpb]

I will take an objective measurement over that BS any day of the week.

I don't believe we will ever see a "fair" or "objective" system for choosing the NCAA participants (cf. the CFP fiasco), but it gives us something to discuss on a dreary February day. An excerpt from Scott Huston's article on the NPI ranking system from the D3 Hockey News Site (emphasis mine)

---

Unfortunately, in our sport, there is a serious scheduling conundrum. With an extreme lack of cross-over play between the East and West region, many have found that the RPI (or any math equation) should not be in charge of setting the postseason field. I’ve looked at advanced calculations on how the math works, and the RPI does get more accurate when there is ample non-conference play. Having seven conferences out east playing amongst themselves allows for a pretty fair ranking (most of the time.) But when there are three conferences in the other region that have very little opportunity but to beat each other up, it’s nearly impossible to rank them along with the east teams when there is so little cross-region play. It’s a hard task, but in my opinion, only using math does not provide a fair shot for all, with many west teams from extremely competitive conferences getting the short end of the stick in these calculations (rankings).

Disclaimer: Huston is a Gustavus alumus.

 

[Tom Coughlin]

You should be rewarded for beating good teams at least, if not more, the same amount as you are penalized for losing to bad teams. That is not the case. The qwb is meaningless.

I totally agree. Unfortunately that isn't how the math shakes out. From what I recall, the committee can adjust the dials every other year. Maybe the success of Salve, etc, will cause some adjustment this offseason.

 

[Tom Coughlin]

I don't believe we will ever see a "fair" or "objective" system for choosing the NCAA participants (cf. the CFP fiasco), but it gives us something to discuss on a dreary February day. An excerpt from Scott Huston's article on the NPI ranking system from the D3 Hockey News Site (emphasis mine)

---

Disclaimer: Huston is a Gustavus alumus.

It is "objective" in the sense that the dials are set before the season starts. Everyone agrees (implicitly or explicitly) on what the criteria will be ahead of time. Nothing was stopping SNC or Aurora or any other team from scheduling 7 non-conference games against Hiram. While that probably would have guaranteed them an at large bid, it wouldn't have prepared them for the tourney. Definitely a catch 22.

 

[JSicedog]

It is "objective" in the sense that the dials are set before the season starts. Everyone agrees (implicitly or explicitly) on what the criteria will be ahead of time. Nothing was stopping SNC or Aurora or any other team from scheduling 7 non-conference games against Hiram. While that probably would have guaranteed them an at large bid, it wouldn't have prepared them for the tourney. Definitely a catch 22.

It wouldn't have guaranteed them an at large, though. Chatham did exactly that, has a better raw win% than Salve, and sits at 15. I fail to understand how Salve's SOS is marginally higher than Chatham's, but their weighted win % is substantially higher.

 

[Tom Coughlin]

It wouldn't have guaranteed them an at large, though. Chatham did exactly that, has a better raw win% than Salve, and sits at 15. I fail to understand how Salve's SOS is marginally higher than Chatham's, but their weighted win % is substantially higher.

"Weighted" win % is adjusted for where you played (home or away) and for whether the game finished in regulation or OT. As far as I know, it has nothing to do with who you played.

The difference in NPI for Salve and Chatham is almost entirely explained by 75% of the difference in SOS, 25% of the difference in weighted win %, and the difference in QWB. There must be another small adjustment that I am missing though.

 

[Tom Coughlin]

It wouldn't have guaranteed them an at large, though. Chatham did exactly that, has a better raw win% than Salve, and sits at 15. I fail to understand how Salve's SOS is marginally higher than Chatham's, but their weighted win % is substantially higher.

Also, Chatham's raw win % is worse than Salve's. 17-3 is 85% (Salve) while 19-4 is 82.6% (Chatham).

 

[JSicedog]

"Weighted" win % is adjusted for where you played (home or away) and for whether the game finished in regulation or OT. As far as I know, it has nothing to do with who you played.

The difference in NPI for Salve and Chatham is almost entirely explained by 75% of the difference in SOS, 25% of the difference in weighted win %, and the difference in QWB. There must be another small adjustment that I am missing though.

Chatham 8-1 at home (.88%), 9-3 away (.75%), 2-0 neutral (100%)
Salve 7-1 at home (.875%), 7-1 away (.875%), 3-1 neutral (.75%)

Two more away losses for Chatham results in Salve's weighted win% to be 4.5 points higher than Chathams?

Hypothetically speaking Chatham's SOS should get a boost this weekend with two games at Utica. If they were to sweep Utica, how are they not a top 5 team in the NPI?

 

[VTpucksouteast]

I don't believe we will ever see a "fair" or "objective" system for choosing the NCAA participants (cf. the CFP fiasco), but it gives us something to discuss on a dreary February day. An excerpt from Scott Huston's article on the NPI ranking system from the D3 Hockey News Site (emphasis mine)

The article he writes makes a good point though, GAC alum or not. The NPI gets more "accurate" the more variety in teams you play.

When one region has 66 teams and the other has 25, the region with less teams is going to get the short end of the stick, particularly the ones who play zero east teams.

 

[elbojpb]

My assumption was that the weighting percentages would be "tweaked" each year, but according to the NCAA Q&A document, the % could be tweaked after this (the initial) season, but then only after every 3rd season!?

Q: When would sport committees adjust dials/weights?

A: Sport committees will use their annual meeting to review the championship selections process. Any changes to the weight of any dial will be communicated to the membership via the annual meeting report, direct communication to the coaching association and posted on the sport committee’s landing page on NCAA.org. Once the weights are established, they cannot be changed for a minimum of three seasons. Dials will not be changed weekly during the season. Note: The Championships Committee will allow sport committees to reconsider dial setting weights after the first year (2026-27 academic year) to ensure the initial weight setting was appropriate. It is anticipated that these adjustments would be limited, if at all. Division III did not change any dial weights after the first year.

 

[Tom Coughlin]

The article he writes makes a good point though, GAC alum or not. The NPI gets more "accurate" the more variety in teams you play.

When one region has 66 teams and the other has 25, the region with less teams is going to get the short end of the stick, particularly the ones who play zero east teams.

I could be wrong, but I thought the cross region matchups had a greater impact on the old pairwise ranking system. That system compared teams in both head to head matchups and record against common opponents (where most West region teams would have few/no common opponents against East region teams). I think part of the appeal of the NPI (at least as advertised) is that the rankings are not impacted by (lack of) cross region matchups.

 

[PrezdeJohnson09]

I could be wrong, but I thought the cross region matchups had a greater impact on the old pairwise ranking system. That system compared teams in both head to head matchups and record against common opponents (where most West region teams would have few/no common opponents against East region teams). I think part of the appeal of the NPI (at least as advertised) is that the rankings are not impacted by (lack of) cross region matchups.

I still think the objectivity of the math is the way to go. The women's tournament, despite what Sugar prattles on about on X these days looks pretty darn good based on what has happened so far in the results. His prejudice against the NESCAC and Eau Claire being on the outside looking in right now is what is sparking his ire there.

The men's side however clearly has some flaws if/when Salve appears like they will be getting in. What i find the most troubling on the men's side is it just seems like SOS is not being weighted appropriately currently and the QWB seems to not be much of anything at all right now. If they can adjust the dials a bit to clean this up, this system can and will still work even with the limited cross-region play.

I've been following D-III hockey for 30 plus years, I assure you NO ONE should want to return to the committee days and bringing subjectivity into the room. You will never get unbiased comparisons and the regionality bias and games that each region's committee plays will continue year and after year. If the west does something one year or one week, the east will counter the next and so on and so forth. It will always be a constant game of gotcha and trying to skew the results to favor individual teams/schools or regions. We do not want to return to this, when there is and has already been proven in most years to be a better way.

 

[elbojpb]

I've been following D-III hockey for 30 plus years, I assure you NO ONE should want to return to the committee days and bringing subjectivity into the room. You will never get unbiased comparisons and the regionality bias and games that each region's committee plays will continue year and after year. If the west does something one year or one week, the east will counter the next and so on and so forth. It will always be a constant game of gotcha and trying to skew the results to favor individual teams/schools or regions. We do not want to return to this, when there is and has already been proven in most years to be a better way.

Huston's take on the exact same issue ... probably skewed by the fact that only one NCHA team reaches the tourney in this week's bracketology edition.

Opinion: The NPI is an extremely flawed metric for Division III hockey.
You cannot base the sport’s national tournament off math alone, when schedules are wildly different and there is so little cross-over play, if any between many conferences.
I know there are people who used to dread “the smoke-filled committee room,” who selected the field and seeding back in the day. But, in my opinion, it’s wildly better than this system. The NPI being used as a tool is more than okay, but it should not be the end-all, be-all.