The Athletic's Predictions (Rangers 14th)

[DialUp]

You're missing the point. Wysh has been over-the-top in his actual, written homerism. He's a known Devils homer so him putting that out there is pretty suspect to me.

He's an idiot and you should never, ever take anything Wysh says about the Devils or Rangers at face value. Period.

OK, I mean I knew the idiot part already, so the Devils homerism makes sense.

 

[TominNC]

And we are Rangers homers. These models and predictions do not affect the way the players play, so ultimately it doesn't matter. It's interesting to talk about, interesting when it goes right, interesting when it goes wrong.

An unbiased person could see that much of a difference in fortunes in the Devils and Rangers in one season?

Please

 

[Machinehead]

Goalies are overrated, a hot goalie can steal a series..which is it?

Both I guess?

Goalies can steal a series and are the most important players in a vacuum.

I think they're overrated not because they're not important but because the differences between the 50 goalies or whatever (as opposed to 300 forwards) is marginal, and because goaltending is pretty random.

I don't care who your goalie is, the old "getting hot at the right time" is a dice roll. The chances of a Vezina winner doing it over some rando is not as much as people think it is.

Andrew Hammond went like 25-3 over a stretch once and he wasn't an NHL goaltender. Matt Murray was fantastic in a Cup run twice(!) and he's f***ing terrible. Meanwhile, a slew of Hall of Fame goalies never won a Cup: Hank, Price, Luongo, Esposito, Giacomin. Hasek had to go to a video game Red Wings team to win it.

If there's no cap, then sure, play the margins. The cap world is kind of the thing that makes them overrated.

 

[DialUp]

An unbiased person could see that much of a difference in fortunes in the Devils and Rangers in one season?

Please

We went from out of the playoffs to ECF. The model doesn't play the game, it just takes inputs, does things, then spits out data. I suppose that is unbiased theoretically but the model maker might have biases (hint: shots on goal, possession, 5v5 goals) when designing the models. It's just one way to measure teams and predict. Rangers and St Louis betray the model using last year as an example. The devil's are model darlings. So is Edmonton! Lol. Take it as you wish.

NYR will be "ranked" higher by these analysts if they make the playoffs year after year doing it the same way they did it last year. Or maybe just be better with better players, which would be less stressful

 

[HFsNumber1Heel]

This model isn't too bad. So the Rangers are a WC team which is where we had them last year.

It's a similar situation where there is an unbalanced amount of wingers not slotted correctly. Can they overachieve like last year and add depth at the deadline?

Will anyone from Hartford progress and be ready for a callup?

A lot depends on Panarin. You know Mike and Kreider will do their usual stuff but if Panarin plays like he did against Boston they're in trouble with that line. Him and Trochek haven't looked good together so far.

 

[The Crypto Guy]

Before anyone goes crazy, Dom starts the Rangers article off stating that his model is not good at predicting prospect breakouts and therefore assumes all the kids will play at the level they were at last season, which obviously won't happen.

The model uses the last 3 years of data, and is therefore skeptical when a team makes a big jump like the Rangers did.

He does have a valid point about regression from Shesterkin and Kreider, seems inevitable.



Yeah he literally says he thinks his model will be proven wrong here, if the kids progress. Go off though.

Maybe it’s time he…finds a new model then? Why post something so flawed?

 

[TominNC]

Maybe it’s time he…finds a new model then? Why post something so flawed?

Thank you. When your model and look at it and it look f****d them yeah maybe don’t publish that. If you greatly underestimated the Rangers last year maybe think about what you missed. Instead we get “regression from the good year” and “none of the kids improve”

 

[mas0764]

Thank you. When your model and look at it and it look f****d them yeah maybe don’t publish that. If you greatly underestimated the Rangers last year maybe think about what you missed. Instead we get “regression from the good year” and “none of the kids improve”

Because three years of data is still more reliable than guessing about progress from kids.

Which I have in the past argued is 100% reliable to count on, but at the same time it’s hard to quantify.

I see no problem with saying “the stats say the Rangers will struggle at 5v5 though when Laf and Kakko improve they will be better and then my model will take that into account.”

It’s like you guys are offended that the model says they are 14th. Stats don’t lie so if the team doesn’t progress from it’s young kids then we can expect to struggle at 5v5 again.

The Rangers top two lines being based around Kreider-Zibanejad and Panarin-Strome/Trocheck haven’t been very good and don’t project to be very good, they are the backbone of a team that is mediocre and requires carrying by its goalie. Which is what I have been ranting about for some time. We need way more forward talent and if we want to be a contender guys like Kreider need to be more like our middle sixers than sacred cows who can’t be moved off the first line.

 

[mas0764]

So this model predicts Igor to regress but kakko, Laf, Chytil, Miller, etc to not improve? Got it.

It isn’t projecting that they won’t improve, it’s saying that it doesn’t know how to project their improvement.

As such the projections shouldn’t be seen as an ironclad lock of “this is how the teams will finish,” but rather how they are likely to finish absent big jumps.

In general a third year player who has posted 0.5 GSVA scores two years before isn’t going to have a meaningful change, but for Lafreniere we are obviously looking at different circumstances. But there is no model specifically for former first overall picks who struggled the first two years before making a huge leap their third year. The only time it has happened is in different eras of the NHL like with Joe Thornton twenty years ago. It’s inapplicable data and no better than guessing.

This season, besides Colorado, which teams would be considered not in the field?

Sometimes it probably is just one team. I’d have to go back and look.

 

[IDvsEGO]

Because three years of data is still more reliable than guessing about progress from kids.

Which I have in the past argued is 100% reliable to count on, but at the same time it’s hard to quantify.

I see no problem with saying “the stats say the Rangers will struggle at 5v5 though when Laf and Kakko improve they will be better and then my model will take that into account.”

It’s like you guys are offended that the model says they are 14th. Stats don’t lie so if the team doesn’t progress from it’s young kids then we can expect to struggle at 5v5 again.

The Rangers top two lines being based around Kreider-Zibanejad and Panarin-Strome/Trocheck haven’t been very good and don’t project to be very good, they are the backbone of a team that is mediocre and requires carrying by its goalie. Which is what I have been ranting about for some time. We need way more forward talent and if we want to be a contender guys like Kreider need to be more like our middle sixers than sacred cows who can’t be moved off the first line.

So you start off with a reasonable argument, and then say "the rangers top two lines being based around mika and panarin don't project to be very good.
They're solidly middle of the league, and it absolutely depends on the Right wings they have. Panarin is a top 15 player, mika a top 30, kreider doesn't project that high because models consider last year a fluke so they almost discard it.
If Kreider starts hot this season the models will immediately update with a projection of closer to 35-40g and the rangers top 6 would look to be in the top 10.
The issue the models have with the rangers is actually depth, the kids on the third don't project high enough yet, and the fourth is just bad , and theres not enough top end to balance.
If we run the kids in the top 6 (and they take a big jump) all of a sudden the models will love us.
Look at a team like anaheim or seattle. The models have so little data so they're basically making shit up.

 

[DialUp]

If we run the kids in the top 6 (and they take a big jump) all of a sudden the models will love us.

Is that right? How? They don't score goals yet. Are you saying they just need more ice time? I think they do, without it they will always be as shit as they are. We aren't winning the cup on the back of our top 4 forwards, we absolutely need others to step up.

There is a reason why all the teams the models have above us, are above us: their players are better as a whole. Every team would want Fox, Mika, Panarin, Igor, and probably even Kreider, but I highly doubt any team gives a damn about anyone else (until K'Andre has his expected breakout). At the end of year we'll see if that holds true. Either the 3rd line becomes actually good or they don't. If they don't, we will be ranked squarely in the middle of the pack again. Results be damned.

*Edit* now that I think about it, not many teams have more than 4 truly coveted guys. So who knows. My thoughts are as valuable as the models.

 

[mas0764]

So you start off with a reasonable argument, and then say "the rangers top two lines being based around mika and panarin don't project to be very good.
They're solidly middle of the league, and it absolutely depends on the Right wings they have.

Yeah that’s what I said, middle of the league isn’t “very good.” And yes if Lafreniere and Kakko exploded and were placed on those two lines, that is what would kick us up to more towards the elite tier.

Otherwise we are relying on goaltending to get there but teams with elite forwards are the far more often winners.

 

[TominNC]

Because three years of data is still more reliable than guessing about progress from kids.

Which I have in the past argued is 100% reliable to count on, but at the same time it’s hard to quantify.

I see no problem with saying “the stats say the Rangers will struggle at 5v5 though when Laf and Kakko improve they will be better and then my model will take that into account.”

It’s like you guys are offended that the model says they are 14th. Stats don’t lie so if the team doesn’t progress from it’s young kids then we can expect to struggle at 5v5 again.

The Rangers top two lines being based around Kreider-Zibanejad and Panarin-Strome/Trocheck haven’t been very good and don’t project to be very good, they are the backbone of a team that is mediocre and requires carrying by its goalie. Which is what I have been ranting about for some time. We need way more forward talent and if we want to be a contender guys like Kreider need to be more like our middle sixers than sacred cows who can’t be moved off the first line.

Stats lied last year

 

[mas0764]

Stats lied last year

No they didn’t. We were mediocre at 5v5 and were carried by a Hall of Fame caliber goaltending performance and hot power play. Those things are unlikely to both repeat at the same level. Luckily the kids should improve which the model also has a hard time predicting.

 

[TominNC]

So a reasonable person could assume excellent goaltending an improvement 5v5 because the kids mature.

 

[Leonardo87]

Yet another reason to add why I cancelled my subscription.

I know you guys don’t follow but they have one guy covering the Ducks and Kings. That’s like having one guy cover the Rangers and Devils. Lol. It just doesn’t work.

It’s a joke with their cut backs and articles like this aren’t going to help their site.

 

[DialUp]

Stats lied last year

All due respect, not sure you understand what the model says. Stats lie only if they were falsified.

 

[TominNC]

All due respect, not sure you understand what the model says. Stats lie only if they were falsified.

Ok the model was WRONG last year. Happy? I was responding to a post that said stats don’t lie.

 

[Kaapo Cabana]

Yes, it really is.

He's not just as wrong as he is right, that's just factually incorrect.

It's also not guessing because it's based on a statistical analysis. Perhaps what you are trying to say is that his statistics do not generate results that are better than guesses, but again, that's factually incorrect. His model produces a pretty good result actually.

But your lambasting of the Athletic as "holier than thou behind a paywall," makes me think you have another issue with the publication or with analytics in general, which is also unfounded.

Your general discrediting is an overreach. Statistics and analytics are not perfect but they are far from worthless and Dom's model isn't worthless either.

I completely disagree.

The athletic is a rag that employs reporters who only care about analytics and ignore every other facet of the game.

Analytics have their place, but they are a complimentary tool. That’s it.

It’s easy to spot the people who over use it in their analysis. Dom L is exhibit A

His takes are observably trash.

 

[Hunter Gathers]

OK, I mean I knew the idiot part already, so the Devils homerism makes sense.

Yeah, he wrote a whole f***ing book on the Devils, lol.

I completely disagree.

The athletic is a rag that employs reporters who only care about analytics and ignore every other facet of the game.

Analytics have their place, but they are a complimentary tool. That’s it.

It’s easy to spot the people who over use it in their analysis. Dom L is exhibit A

His takes are observably trash.

Tell me you don't read The Athletic without telling me you don't read The Athletic.

 

[mas0764]

So a reasonable person could assume excellent goaltending an improvement 5v5 because the kids mature.

A reasonable person probably wouldn’t expect the goaltending to be quite so all time high again hence the model predicting regression.

Yes a reasonable person would also expect the kids to improve but the model admits it can’t project that because it’s statistics based. So if the kids do improve we will be better than the model predicts; if they don’t we probably regress. It says that up front so I’m not sure what the issue is.

 

[DanielBrassard]

So a reasonable person could assume excellent goaltending an improvement 5v5 because the kids mature.

Excellent goaltending would be a quite substantial downgrade compared to the season Igor had last season. According to Goals Saved Above Expected he was more than twice as good as Vasilevskiy who was 4th, and BTW Igor played 700 less minutes. It's a cumulative stat. I'm not sure you or others comprehend just how spectacular Igor's season was last year. One of the better years, maybe the best we have seen on a per minute basis in the analytics era. So it's not as simple as some improvement from the kids or whatever.

Perhaps Igor has another historic season and the Rangers completely outperform what is expected or the kids explode. But if he is just merely great the Rangers are going to have to improve drastically at 5v5 to be anywhere close to where they were last season. I don't think it's unreasonable to project it won't happen when the Rangers track record with this same core group over the last few years has consistently been below average at 5v5. Hopefully that changes.

 

[TominNC]

Excellent goaltending would be a quite substantial downgrade compared to the season Igor had last season. According to Goals Saved Above Expected he was more than twice as good as Vasilevskiy who was 4th, and BTW Igor played 700 less minutes. It's a cumulative stat. I'm not sure you or others comprehend just how spectacular Igor's season was last year. One of the better years, maybe the best we have seen on a per minute basis in the analytics era. So it's not as simple as some improvement from the kids or whatever.

Perhaps Igor has another historic season and the Rangers completely outperform what is expected or the kids explode. But if he is just merely great the Rangers are going to have to improve drastically at 5v5 to be anywhere close to where they were last season. I don't think it's unreasonable to project it won't happen when the Rangers track record with this same core group over the last few years has consistently been below average at 5v5. Hopefully that changes.

Please don’t casually dismiss other fans knowledge of the game and the significance of Shesty’s season. He can be quite excellent this year without being that good. On the other side I think we can reasonably expect a decent jump in performance from multiple young players.

 

[Kaapo Cabana]

Yeah, he wrote a whole f***ing book on the Devils, lol.


Tell me you don't read The Athletic without telling me you don't read The Athletic.

Yeah, he wrote a whole f***ing book on the Devils, lol.


Tell me you don't read The Athletic without telling me you don't read The Athletic.

I certainly don’t anymore. Not interested in galaxy brain think pieces that fly in the face of reality

 

[NickyFotiu]

This is model nonsense but I wish NYR fans would stop assuming we will just pick up where we left off last season. Hopefully we will but there are always teams that take steps forward and backwards. The cap leads to parity. Nothing is guaranteed for us. We can beat almost anyone but almost anyone can beat us as well.