Polls and NPI 2026

[adamw]

I think that's wrong. QWB should be a minor factor. Which tool are you using, btw? I'm not familiar with anything like that on CHN before the You Are the Committee comes out.

Maybe adamw can clarify since he's here—how much NPI does Cornell gain if they win today, and are they locked if they do?

I usually use the CHN-native NPI customizer. Where you see the NPI stats, there's a thing on the right where you can input future games' results and see the change in NPI.

Thanks. I had never noticed that before. I think this tool is currently bugged, though. When I plug in Cornell winning today, their NPI goes down...?

Try it again. Sometimes it goes down to .44, sometimes it goes up to .51. I don't get it either, might be a bug or just cache issues.

I'll look into that. It does try to cache your results so that, after you hit submit the first time, when you go back, it's supposed to be in the state you left it so that you can adjust from there, rather than having to start over.  So maybe there's some sort of incorrect compounding going on.  If you spot anything, let me know.

I mean, that's the one thing I'm seeing right now. Dunno if it'd happen for other teams/stuff.

I know what's happening - but haven't fixed it yet.

If you "modify results" and click Cornell winning today - and submit. It switches your selection to Harvard (which can be seen by going back to the Customize tab - Harvard is selected). But if on the next try, you select Cornell again and hit submit, then that time it sticks.

So first try - you're going to see a worse NPI than they have now. Second try will be correct.

I'm also not sure that the script is accounting for that being a conference playoff game, and thus no more home/road weighting. Need to check that too.

[stereax]

I think that's wrong. QWB should be a minor factor. Which tool are you using, btw? I'm not familiar with anything like that on CHN before the You Are the Committee comes out.

Maybe adamw can clarify since he's here—how much NPI does Cornell gain if they win today, and are they locked if they do?

I usually use the CHN-native NPI customizer. Where you see the NPI stats, there's a thing on the right where you can input future games' results and see the change in NPI.

Thanks. I had never noticed that before. I think this tool is currently bugged, though. When I plug in Cornell winning today, their NPI goes down...?

Try it again. Sometimes it goes down to .44, sometimes it goes up to .51. I don't get it either, might be a bug or just cache issues.

I'll look into that. It does try to cache your results so that, after you hit submit the first time, when you go back, it's supposed to be in the state you left it so that you can adjust from there, rather than having to start over.  So maybe there's some sort of incorrect compounding going on.  If you spot anything, let me know.

I mean, that's the one thing I'm seeing right now. Dunno if it'd happen for other teams/stuff.

I know what's happening - but haven't fixed it yet.

If you "modify results" and click Cornell winning today - and submit. It switches your selection to Harvard (which can be seen by going back to the Customize tab - Harvard is selected). But if on the next try, you select Cornell again and hit submit, then that time it sticks.

So first try - you're going to see a worse NPI than they have now. Second try will be correct.

I'm also not sure that the script is accounting for that being a conference playoff game, and thus no more home/road weighting. Need to check that too.

Yeah, but when I choose Harvard it goes down to 54.76. If it was actually choosing Harvard over Cornell, it'd show that, and not the 55.43 number...

[adamw]

I think that's wrong. QWB should be a minor factor. Which tool are you using, btw? I'm not familiar with anything like that on CHN before the You Are the Committee comes out.

Maybe adamw can clarify since he's here—how much NPI does Cornell gain if they win today, and are they locked if they do?

I usually use the CHN-native NPI customizer. Where you see the NPI stats, there's a thing on the right where you can input future games' results and see the change in NPI.

Thanks. I had never noticed that before. I think this tool is currently bugged, though. When I plug in Cornell winning today, their NPI goes down...?

Try it again. Sometimes it goes down to .44, sometimes it goes up to .51. I don't get it either, might be a bug or just cache issues.

I'll look into that. It does try to cache your results so that, after you hit submit the first time, when you go back, it's supposed to be in the state you left it so that you can adjust from there, rather than having to start over.  So maybe there's some sort of incorrect compounding going on.  If you spot anything, let me know.

I mean, that's the one thing I'm seeing right now. Dunno if it'd happen for other teams/stuff.

I know what's happening - but haven't fixed it yet.

If you "modify results" and click Cornell winning today - and submit. It switches your selection to Harvard (which can be seen by going back to the Customize tab - Harvard is selected). But if on the next try, you select Cornell again and hit submit, then that time it sticks.

So first try - you're going to see a worse NPI than they have now. Second try will be correct.

I'm also not sure that the script is accounting for that being a conference playoff game, and thus no more home/road weighting. Need to check that too.

Yeah, but when I choose Harvard it goes down to 54.76. If it was actually choosing Harvard over Cornell, it'd show that, and not the 55.43 number...

at this point - I think it has something to do with remembering past selections and not clearing things out. Because if I hit "reset customizations" (which clears the local storage cache) and start over, it's fine.

I do need to have it account for league tournaments having no home/road weighting though

[GBR1234]

From CHN
Onward
Cornell got past Harvard on Sunday to complete the ECAC's four-some in Lake Placid. And according to our Probability Matrix, the Big Red are now a lock for the NCAAs. Other teams hinge on the bubble. Check out the current NPI, and use You Are the Committee to see what will happen based on next weekend's final results.

[GBR1234]

https://www.collegehockeynews.com/ratings/probabilityMatrix.php

[BearLover]

I can't remember in which thread people were talking about this, but per the CHN probability matrix, there is a good chance we end up with the 6, 7, 9, or 11 seed. Very small chance we end up with 8 or 10.

[BearLover]

I just listened to today's USCHO bracketology podcast. I believe the hosts stated several things that are factually wrong. However, they did mention an interesting caveat, which is that they aren't confident the NPI numbers currently available are 100% accurate because this is the first time college hockey has used NPI.

I do note that I spot-checked a few numbers and the USCHO and CHN NPI numbers match, so at the very least those two websites appear to be calculating it the same way.

[upprdeck]

You know what would fix the problem?  The NCAA creating an API so other sites dont need to figure it out on their own.

It would need to be a smart API so you could adjust game results for what if stuff.

[ugarte]

got a umass friend begging us to tcb to preserve their dwindling at-large chances. good position to be in imo.

[Trotsky]

I know what's happening - but haven't fixed it yet.

-- Congress

[Trotsky]

https://www.collegehockeynews.com/ratings/probabilityMatrix.php

The top 4 already being locked in seems wild to me.

Also, WTF Wisco?

Also, also, why does Clarkson not map to a seed?

Edit: oh, I get it,  scroll right.  The columns aren't necessarily the tourny seeds, they are the final NPI positions.

Very nicely done table, Adam.

[upprdeck]

Especially as 3 of them are already done playing.

The fact that the NCHC has 4 of the top 6 kinda makes you wonder though

Duluth basically finished .500 in league 3-7 vs the top teams

but did great vs Lindwood. alaska, Bemidji, Omaha, ASU, Miami it adds up.

would 5-6 more wins do much for Cornells NPI?

Just winning vs BU or UMass would have given Cornell a chance for a top 4 seed perhaps?

[stereax]

Especially as 3 of them are already done playing.

The fact that the NCHC has 4 of the top 6 kinda makes you wonder though

Duluth basically finished .500 in league 3-7 vs the top teams

but did great vs Lindwood. alaska, Bemidji, Omaha, ASU, Miami it adds up.

would 5-6 more wins do much for Cornells NPI?

Just winning vs BU or UMass would have given Cornell a chance for a top 4 seed perhaps?

Flipping the BU matchup in the NPI play-around-with-er gives Cornell 6 at 56.55 right now. Flip the first UMass one, 56.53. Flip both, 57.37 and 5th, 0.01 NPI points behind WMU.

Part of the pain of being in the Ivy League is that we schedule fewer games... so each loss means more.

[stereax]

Especially as 3 of them are already done playing.

The fact that the NCHC has 4 of the top 6 kinda makes you wonder though

Duluth basically finished .500 in league 3-7 vs the top teams

but did great vs Lindwood. alaska, Bemidji, Omaha, ASU, Miami it adds up.

would 5-6 more wins do much for Cornells NPI?

Just winning vs BU or UMass would have given Cornell a chance for a top 4 seed perhaps?

Flipping the BU matchup in the NPI play-around-with-er gives Cornell 6 at 56.55 right now. Flip the first UMass one, 56.53. Flip both, 57.37 and 5th, 0.01 NPI points behind WMU.

Part of the pain of being in the Ivy League is that we schedule fewer games... so each loss means more.

Funny enough, if I flip the Princeton 2-AM-bus game and the RPI tie to wins, we end up at 56.77. Flip the Union game with it, 57.47, 4th, and getting bad wins discarded. That's without the BU or first UMass game.

It's all a game of inches. No point sitting in the coulda-woulda-shoulda now. Good thing UMass is doing well because that helps us a lot.

[Weder]

Part of the pain of being in the Ivy League is that we schedule fewer games... so each loss means more.

I'll never be able to find it again, but I saw a clip the the other day of the Wisconsin coach pushing for more regular season games. Don't know how widely held that opinion is among other coaches, though.