11/14/2026 Cornell at Yale

Started by Sperris, November 14, 2026, 06:18:10 PM

Previous topic - Next topic

BearLover

Quote from: stereax on November 19, 2026, 10:33:16 AM
Quote from: BearLover on November 19, 2026, 08:14:29 AM
Quote from: stereax on November 15, 2026, 06:42:19 PM
Quote from: BearLover on November 15, 2026, 05:32:05 PM
Quote from: stereax on November 15, 2026, 02:16:43 PM
Quote from: The Rancor on November 15, 2026, 01:34:15 PM
Quote from: stereax on November 15, 2026, 11:53:44 AM
Quote from: Snowball on November 15, 2026, 11:15:48 AMThere was a fair amount of play to like in this game: Rousseau was consistent and had some marvelous saves. There was good play form many: Cato of course, Zad, Pirtle, Fegaras, Fisher, Ashton and Gorski, imho. Of the new guys I thought that Randle and Gardner were impressive, and I love the energy of Coco.

Something's off with Walshy and face-offs: he was 8 for 22 last night, and was 16 of 34 the night before. That alone had a big impact on the weekend. Upper body injury?

Consider this counterfactual: one of Walshy's shots goes in, we beat Yale, and everyone walks away feeling a little better about themselves.

Would that actually have been a good thing? I don't think so. We'd be talking about a win instead of confronting the real problems staring this team in the face. Sometimes a loss like this, one that stings for an entire weekend, is exactly what a team needs. Better to have the wheels come off the early in the season than to keep papering over the cracks. This sh— needs to get fixed, and it needs to get fixed now.

Casey's a great coach, he can get this done, I am looking forward to it.
Personally, I wouldn't have felt good about a win unless it was, as I mentioned previously, an "ass-blasting" of the highest order. Winning, say, 2-1 would have been nice for my mental health, sure, but it would mean we barely edged Yale, one of the worst teams in the ECAC. If we showed up and put down 5 or 6, after that miserable Brown outing, that would have shown me that the team saw what happened the night before and said "fuck that, THAT'S not happening again", put their big boy skates on, and demolished the team playing against them. We had a good first period against Yale, and then just petered out.

Sometimes a change in net can wake a room right the fuck up. Last year, the guys played way better in front of Keopple, even if he was a worse goaltender. Maybe let Rousseau sit a game, put Roest in, see what you have there and send a message.

Agreed with your assessment that a lot of the guys individually looked good. Glad to see Pirtle back in the lineup, he always provides a spark. The freshmen have been overall quite good as well.

In re faceoffs - there are new faceoff rules league-wide. I assume that's part of it. Also, having Walshy take 34 faceoffs on a night is probably part of the issue.

Anyways. It feels like the bulk of the issue is with the forward corps - the defense has been great overall and the goalie hasn't been the problem, either. We need scoring now, and for these forward lines to click. Cato, Z, and Kraft can't be all of our scoring.

Casey will figure it out.

Re: Goaltending. You can't (or shouldn't) sit the goaltender, who came here to be a starter- much like he who shant be named- if he's performing as needed just to send a message to the rest of the team. I get what you are saying: "Anyone can sit, even him" but the next thing that happens is a call to his player agent to type up a portal letter. If he was giving up 5 goals on 10 shots and getting pulled in the 2nd, maybe. But as most of us, and Casey have said: it's the offense.

Also, and I might be misremembering a Schefer quote, but they can't score if we have the puck. Not sure if that applies specifically, but always good to keep in mind.
Yea, you're probably right about that one, we don't need another portaling goalie 😅

The thing about Rousseau is he is a League Average Goaltender. Which is fine for Cornell - we're not generally a super leaky team! But as a corollary to that, he's generally not going to be stealing games for you. (That being said, letting in like 0.5 goals more than expected per game has been adding up and he's at -3.07 on the year, good for 108th of 133 on the list at CHN. Where Courns was letting in 1 goal a game in the fall last year, Rous is letting in 2. Still has a positive CHIP in the same range as Spunar (NoDak) and Smith (P) though.)
I don't think CHIP is a real stat. Goals saved above expected is probably the best we can do *if* the expected goal statistic is properly calibrated, which for college hockey it sounds like it isn't because it doesn't factor in type of shot, speed of shot, etc. (as compared to the NHL where the stat factors in all this and more). So I'm afraid save % might be the best publicly available stat we have to evaluate college goalies. From that perspective we have been a lot worse the national average. (FWIW, if you look at goals saved above expected we've been similarly bad.) Hoping this turns around, obviously.
Enjoy.

TL;DR G + A + +/- + TOI + shot location + SOS. Blend is proprietary.

The issue is, SV% on its own doesn't say all that much. Even I could have a high SV% if I 1) managed to stay upright on skates long enough to make it to the net and 2) had softies hit directly into my stomach. GSAx is generally a lot more useful - but, of course, shot tracking isn't that good in the NCAA.

Additionally, CHN is the only publicly available resource for advanced stats. If you'd like to toss me $1k a year to subscribe to InStat, which has different numbers (I believe RMS uses it), please feel free to. Otherwise, I'm going off what I have available to me.

( @adamw , btw, this specific link keeps redirecting me to some hacker site - trk8.kokamedia.com - if you have time to check it out?)

More advanced stats for the stat nerds:

SV% is .867, which is, I believe, fourth- or fifth-worst in the country. (Rousseau is a .888, the drop is because of ENGs.) His stats specifically - 2.56 GAA last year vs 2.37 this year, .896 vs .888. While his SV% drops from last year to this year, this is likely because of superior shot selection on the shots that beat him, as well as a defensive structure that helps isolate many of the lower-danger shots. 35% of the shots Rousseau faces are high-danger. For context, this year, Cournoyer's HDShF% is 23.4. Hauser? 16.4. Yegorov (BU), 29.7. Lacelle (Omaha), 29.5. Rocha (Alaska-Anchorage), 31.2. (I choose the last two as classic "teams being shelled".)

Our Corsi is 59.7%, good for 5th in the country; our Fenwick is 57.6%, good for 8th. These stats essentially mean we are dominating play, for the most part.

GF is 2.2, which is tied for eighth-worst. Shooting percentage is 7.6%, seventh-worst. PDO? 96.40. (For reference, anything below a 98 is considered unlucky, and over 102 is considered lucky.)

If we were an NHL team, our PDO would be second-worst in the league (between CHI and CGY). Corsi would be second-best (between CAR and NJD), ditto Fenwick (between CAR and UTA).

All of this suggests we are due for positive regression, and that we have been EXTREMELY unlucky to date.
Yeah, I've seen this CHIP stat before. But it's just a stat CHN made up. There's nothing in this article that explains why it's reliable or accurate. I can take issue with a bunch of different parts of it. For example, it treats TOI as a proxy for a player being good:

Quote"Conceptually speaking, time on ice measures something that we don't have enough data to otherwise measure," College Hockey News publisher Adam Wodon said. "If a coach is playing a someone a lot, it must mean something."
This seems wrong. In particular, it leads to "double-counting": a player with more ice time is going to score more, AND now they get an additional bonus from being on the ice more?

The above is just one point to show why we shouldn't necessarily care about a statistic that somebody made up. I'm not even faulting CHN for making up this statistic--I welcome attempts to better understand college hockey. But the burden is on them to show why we should place any value in this stat.

For example, I could create a goalie stat that assigns goalies a value based on 33% GAA, 33% save percentage, 33% W/L record. This stat would suck! There's no reason why anyone should care about it.

The high danger shot stats you posted are interesting--where did you find that?
I never said CHIP was a perfect stat - and I agree that adding bonuses just because a guy sees more ice can be a bad way to do it. Just that it's The Best We Have, and a useful shorthand.

The HD shots are also on CHN here! Scroll down to the goalie section. I believe they only have this since the 24-25 season.
yeah, i'm just not convinced a stat whose inputs don't make sense to me and whose weightings are not public information is "the best we have." kind of just seems like an arbitrary number. not trying to start another flame war with adamw so i'm happy to be wrong but i'm totally skeptical of any newly created stat/model like this

thanks for the hd number, that's interesting

stereax

Quote from: BearLover on November 19, 2026, 10:49:39 AMthanks for the hd number, that's interesting
ofc!
Law '27, Section C denizen, liveblogging from Lynah!
it's ster-e-ax, like stereo

Trotsky

I am surprised we are over 50% on FO%.  I'd have guessed we were well below.

Trotsky

#138
CHIP.

The inputs make sense:

  • Time on Ice
  • Shot Location
  • Strength of Schedule

The methodology makes sense, insofar as we have information:

  • Normalization
  • Multi-year data sets to generate the shot location expected value

I'd like complete transparency but perhaps CHN has ambitions to sell it as a proprietary metric.  It is more skeevy if it is secret, but that is their decision.

I like that somebody is making the effort, even if "Hobey Tracker" is more cringe than a week of The View.

Dafatone

Quote from: Trotsky on November 19, 2026, 01:11:24 PMI am surprised we are over 50% on FO%.  I'd have guessed we were well below.

We were well over going into this past weekend, I think.

Then the undercooked chicken struck.

stereax

Quote from: Dafatone on November 19, 2026, 02:11:22 PM
Quote from: Trotsky on November 19, 2026, 01:11:24 PMI am surprised we are over 50% on FO%.  I'd have guessed we were well below.

We were well over going into this past weekend, I think.

Then the undercooked chicken struck.
Pizza Hut.
Law '27, Section C denizen, liveblogging from Lynah!
it's ster-e-ax, like stereo

Trotsky

Quote from: stereax on November 19, 2026, 11:29:30 PM
Quote from: Dafatone on November 19, 2026, 02:11:22 PM
Quote from: Trotsky on November 19, 2026, 01:11:24 PMI am surprised we are over 50% on FO%.  I'd have guessed we were well below.

We were well over going into this past weekend, I think.

Then the undercooked chicken struck.
Pizza Hut.