2026 Bracketology: The Road to the Aud

[***JCB***]

There is one result that's keeping Salve in this field and it's the fact that their looss to St. Mikes doesn't count.

I used to chat with a couple of former posters who know way more about this than me and this is something they used to talk about regarding Norwich in the reverse with the old interlock. Sure maybe it helped their SOS a smidge but Norwich and others had two wins removed from their total every year by having St. Mikes and St. Anslem on their schedule. They always used to say this wasa contributing factor in the Cadets not getting an at large a couple of times they didn't win their league.

I may not be getting the mat exactly right but Salve would be somewhere around 56.5 if all their games were counted which would put them 15th.

 

[Tom Coughlin]

There is one result that's keeping Salve in this field and it's the fact that their looss to St. Mikes doesn't count.

I used to chat with a couple of former posters who know way more about this than me and this is something they used to talk about regarding Norwich in the reverse with the old interlock. Sure maybe it helped their SOS a smidge but Norwich and others had two wins removed from their total every year by having St. Mikes and St. Anslem on their schedule. They always used to say this wasa contributing factor in the Cadets not getting an at large a couple of times they didn't win their league.

I may not be getting the mat exactly right but Salve would be somewhere around 56.5 if all their games were counted which would put them 15th.

Including their D2 games and assuming the same SOS, Salve's NPI would fall by about 0.4 points (25% of a 1.667% reduction in win percentage). That would put them between Adrian and SNC (so likely the first or second team out).

 

[KrbFlint]

There is one result that's keeping Salve in this field and it's the fact that their looss to St. Mikes doesn't count.

I used to chat with a couple of former posters who know way more about this than me and this is something they used to talk about regarding Norwich in the reverse with the old interlock. Sure maybe it helped their SOS a smidge but Norwich and others had two wins removed from their total every year by having St. Mikes and St. Anslem on their schedule. They always used to say this wasa contributing factor in the Cadets not getting an at large a couple of times they didn't win their league.

I may not be getting the mat exactly right but Salve would be somewhere around 56.5 if all their games were counted which would put them 15th.

Losing to St. Mike's should be an automatic disqualifier.

*kidding...kind of

 

[***JCB***]

Including their D2 games and assuming the same SOS, Salve's NPI would fall by about 0.4 points (25% of a 1.667% reduction in win percentage). That would put them between Adrian and SNC (so likely the first or second team out).

Again I could be the one thats wrong but I don't think I am.

(0.25×83.33 W%)+(0.75×47.2 SOS) = 56.225 + QWB .275 = 56.5

Whatever I know it doesn't matter...

 

[Tom Coughlin]

Again I could be the one thats wrong but I don't think I am.

(0.25×83.33 W%)+(0.75×47.2 SOS) = 56.225 + QWB .275 = 56.5

Whatever I know it doesn't matter...

I agree with the math as you laid out, but the only thing that is changing in this hypothetical is the win percentage (which would drop by 1.7ish %). I think the difference between what you are calculating and the higher value displayed in the USCHO NPI rankings is due to this note:

*Team's NPI has been adjusted to remove negative effect from defeating weak opponent.

I do not know how this adjustment works.

 

[Inkblot Sports]

A few things that I think need to be clarified:

The 25/75 weighting of winning percentage and SOS is misleading, as there is no time when a team's overall winning percentage and SOS are combined in that way. The NPI doesn't even output an SOS figure; that's created after-the-fact by averaging the opponent NPI ratings (not something that ever happens within the NPI calculation).

Rather, the 25/75 W%/SOS is applied to each individual opponent, in which the SOS part is that opponent's NPI and the W% is 100 for win or 0 for a loss. This is where the quality win bonus is added in, if applicable (and it also applies to the half-win from a tie and the 1/3-win from an overtime loss in the regular season).

If you've got 9+ wins (adjusting for overtime and ties as well as home/away weighting) and the total for a win is less than your NPI, that win is discarded as long as that doesn't bring you below 9. This applies also to the win "parts" of ties and overtime games... but the loss parts are unaffected. (USCHO's weighted winning percentages, from what I can tell, account for home/away weighting and overtime, but not for excluded results.)

Reposting this since it's relevant to the last few posts.

 

[altazo]

Honestly, seeing that no result has really moved Salve Regina, I don't think Adrian will be on the right side of things come Selection Sunday.

And if St. Norbert wins, Aurora falling below the cutline also becomes a potential outcome

Would Aurora be penalized that much for a road loss at SNC? My gut tells me that if SNC wins the autobid, Aurora gets one of the last at-large berths.

Hamilton is in the midst of almost losing 3 in a row, and yet they stay planted at #2 (that's that NESCAC majik...)

 

[***JCB***]

Reposting this since it's relevant to the last few posts.

Is there a chance the NCAA numbers are slightly different than what USCHO publishes then?

 

[Tom Coughlin]

Reposting this since it's relevant to the last few posts.

Alright...so I can't just do a weighted average of win percentage and SOS?

 

[Tom Coughlin]

How is the SOS of East teams so much higher than West teams? I don't see a single West team in the top 10 in SOS. Is it just because the West doesn't have a single dominant team like Hobart? Or because the MIAC and WIAC beat each other up all season?

 

[Lord_Zubin]

I mean the NESCAC has most of the top 10 slots because well they were the best conference this year with their ooc record. As much as we applaud the top 4 of the ncha this year the other 6 teams in the conference really stunk and dragged them down this year as they went 20-24-3 against the other d3 west conferences and that includes all the ncha teams. Case and point the worst team in the NESCAC in terms of NPI is Wesleyan at 52. They're currently ahead of 5 NCHA teams and theyre 3 hundreths off of MSOE which wouldve made it 6 teams. Having half your league in the bottom third of the entire country is usually not good for ur SOS.

 

[Russell Jaslow]

Is there a chance the NCAA numbers are slightly different than what USCHO publishes then?

No. They both get them from the same source.

 

[altazo]

I mean the NESCAC has most of the top 10 slots because well they were the best conference this year with their ooc record. As much as we applaud the top 4 of the ncha this year the other 6 teams in the conference really stunk and dragged them down this year as they went 20-24-3 against the other d3 west conferences and that includes all the ncha teams. Case and point the worst team in the NESCAC in terms of NPI is Wesleyan at 52. They're currently ahead of 5 NCHA teams and theyre 3 hundreths off of MSOE which wouldve made it 6 teams. Having half your league in the bottom third of the entire country is usually not good for ur SOS.

I get all of this.

It's the same old argument as when we had the selection committees - the west has more parity. The bottom half of the NCHA and WIAC are stronger than they get credit for, and the MIAC is basically entirely composed of slightly above average NCAA D3 programs.

There is no MAC or MASAC to schedule geographically convenient non-conference games with in the west region.

 

[Tom Coughlin]

There is no MAC or MASAC to schedule geographically convenient non-conference games with in the west region.

This is the part that is a bit confusing to me. If records in the NESCAC are inflated because they beat up on the MAC/MASCAC in non-conference games, shouldn't that reflect poorly in their SOS? Or is it just that non-conference games are such a small part of the overall schedule that the more important factor is how the rest of your conference faired in non-con play?

 

[Tom Coughlin]

I think the NCHA non-conference record was also hurt by several teams playing multiple "non-conference" games against other NCHA teams. Instead of beating up on a Beloit-type team, Adrian and Aurora played a home and home non-conference series. Aurora had "non-conference" games against Dubuque and Lake Forest. Adrian and Trine also played a non-conference game. I didn't check the rest of the NCHA, but I'm sure there were other matchups like this. All pulling the NCHA overall record closer to .500.

Edit to add...Lawrence had non-conference games against Dubuque and Marian. With seven "non-conference games" effectively being additional conference matchups, that is 14 lost opportunities for the NCHA to boost their non-conference record.

 

[Lord_Zubin]

The NESCAC played 3 games this year against those 2 conferences so it's more likely ur 2nd point. You can have 7 ranked teams for ur non-conference schedule but when u have 10 games against teams ranked that low now that can largely be cancelled out. U can even use the salve argument by saying scheduling non-conference games between ncha teams shrinks the pool of ooc games the NPI looks at and makes the other results weigh more.

 

[Tom Coughlin]

The NESCAC played 3 games this year against those 2 conferences so it's more likely ur 2nd point. You can have 7 ranked teams for ur non-conference schedule but when u have 10 games against teams ranked that low now that can largely be cancelled out. U can even use the salve argument by saying scheduling non-conference games between ncha teams shrinks the pool of ooc games the NPI looks at and makes the other results weigh more.

Agreed that the 0.500 record in those 14 team games could easily work in the NCHA's favor depending on the season. But with 8 of the 14 team games involving members of the NCHA's "big four", it is likely pulling the overall non-con record in the wrong direction. The NCHA would be much better off from an NPI perspective if Adrian/Aurora/Trine scheduled Beloit or Hiram for those games instead of playing each other.

Edited to add: If you take those 8 team games involving Adrian/Aurora/Trine and assume a 6-2 record against actual non-conference opponents for those games (which I think is reasonable given the strength of those teams), that could easily boost the NCHA's overall non-conference record by a few percentage points. With only 70 total non-conference team games, every potential win (or loss) makes a big difference.

Moral of the story is that the NCHA should ban its members from playing non-conference games against each other.

 

[Tom Coughlin]

This also highlights the lack of conferences to pull non-conference games from in the West, as altazo alluded to. A few of the NCHA "non-conference" matchups came as a result of in-season tournament play. With only three conferences in the West, these tournaments risk using up a non-conference game on a conference opponent unless they are able to draw an East team.

 

[Lord_Zubin]

Agreed that the 0.500 record in those 14 team games could easily work in the NCHA's favor depending on the season. But with 8 of the 14 team games involving members of the NCHA's "big four", it is likely pulling the overall non-con record in the wrong direction. The NCHA would be much better off from an NPI perspective if Adrian/Aurora/Trine scheduled Beloit or Hiram for those games instead of playing each other.

Edited to add: If you take those 8 team games involving Adrian/Aurora/Trine and assume a 6-2 record against actual non-conference opponents for those games (which I think is reasonable given the strength of those teams), that could easily boost the NCHA's overall non-conference record by a few percentage points. With only 70 total non-conference team games, every potential win (or loss) makes a big difference.

Moral of the story is that the NCHA should ban its members from playing non-conference games against each other.

ironically adrian did play hiram 2x this year

 

[Tom Coughlin]

ironically adrian did play hiram 2x this year

Right, and they should have played them two more times instead of playing Aurora twice

But seriously, Adrian would be in position for an at large bid right now if they would have done that.