The Athletic's Predictions (Rangers 14th)

[chosen]

2022-23 NHL team previews: Projections and ranking of every NHL team

Between Sept. 19 and Oct. 4, we’re previewing all 32 NHL teams in order of predicted regular-season finish, according to Dom’s model.

theathletic.com

Click to read more...

 

[Kane One]

Oh look, another self-proclaimed expert with their own fool-proof model.

 

[EpicDing]

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.

Oh look, another self-proclaimed expert with their own fool-proof model.

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

 

[tlk]

These need a mandatory "[MODEL]" in the title

 

[Kaapo Cabana]

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.

Then it’s not an article worth reading

 

[GENESISPuck94]

I'm not expecting any real drop off for Shesterkin. He's entering his prime, I think we can count on him being elite for at least a few years.

Is Halak dependable? I'm not sure.

 

[irishlaxburger2]

Dom's model consistently downgrades the ability for goaltenders to win games. So while he's probably right there's some level of regression that may happen with Shesterkin, his model still always undervalues quality goaltending.

It's basically just a model looking at 5v5 possession and shot metrics, and acts like that's all that matters. We know we're not elite in that regard.

Now, do I think his model is actually qualified in predicting a Stanley Cup winner? Yeah, I honestly do think it has some validity given the importance of 5v5 play throughout the entire playoffs. But the model loses out on predicting literally anything else, including how to rank the teams that likely aren't going to win the Cup.

 

[OriginalLimbo]

I'm not expecting any real drop off for Shesterkin. He's entering his prime, I think we can count on him being elite for at least a few years.

Is Halak dependable? I'm not sure.

I get why statistically, one could expect a revert to the mean. But, maybe excellence is his mean now? I would prefer to think this is who he is until his play proves it wrong.

 

[EdJovanovski]

Analytics are for losers
f*** the Athletic, pathetic rag that begs for a spare dollar like a pan handler on a subway.

 

[Kane One]

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.

That’s not said in the snippet of the article given to people who don’t have a subscription. So go off, though.

 

[KirkAlbuquerque]

Dom's a little bitch.

 

[irishlaxburger2]

It should come as no surprise that Dom's favorite team coincidentally tops his model.

Funny how that seems to happen.

 

[Fitzy]

But what do the 8 ball, pachinko machine, and the paper origami fortune teller say?

 

[chosen]

It should come as no surprise that Dom's favorite team coincidentally tops his model.

Funny how that seems to happen.

Do you think he fudged the numbers or that he created a model he knew would ensure his team came in first?

 

[Shesterkybomb]

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.

Why would Shesterkin regresss, i can see Kreider dropping from his high goal total but why can't an elite goalie be elite for more than one year? If we're talking goalie pairings i think that could regress but i dont see any reason the best goalie in the league continues to be especially with a d core that is a year older and without Nemeth or Braun.

 

[Bob Richards]

I'm not expecting any real drop off for Shesterkin. He's entering his prime, I think we can count on him being elite for at least a few years.

Is Halak dependable? I'm not sure.

I don’t think the regression here means Igor wouldn’t be elite. I think everyone knows he will be.

It’s whether or not you can count on him repeating last season which was one of the greatest regular seasons by a goalie in NHL history.

I think it’s a fair question to ask.

 

[irishlaxburger2]

Do you think he fudged the numbers or that he created a model he knew would ensure his team came in first?

Based on his commentary - via both Athletic articles and the stuff he spews on Twitter - it wouldn't shock me if he allowed his own personal biases to affect the inputs in his model.

I don't think he's flat out making up numbers. But what he thinks successful hockey looks like can absolutely affect how he weights certain metrics within the methodology he uses.

The fact his model basically hardly counts goaltending at all is curious given Toronto's historic goaltending woes. Among other things.

 

[IDvsEGO]

Based on his commentary - via both Athletic articles and the stuff he spews on Twitter - it wouldn't shock me if he allowed his own personal biases to affect the inputs in his model.

I don't think he's flat out making up numbers. But what he thinks successful hockey looks like can absolutely affect how he weights certain metrics within the methodology he uses.

The fact his model basically hardly counts goaltending at all is curious given Toronto's historic goaltending woes. Among other things.

I think he’s basically tweaked the model to reduce goalie impact because goalies fluctuate far more than players, so it’s hard to use data modeling to get an accurate read. That said yeah it’s bad. Goalie is the singularly most important position.

 

[Raspewtin]

The reactions to anytime Dom L says anything about the Rangers are exhausting

Just ignore the article and move on if you think it's so inaccurate. The Rangers have very real flaws that the vast majority of modern era cup winners do not overcome. Jesus christ just get over it.

Dom's model consistently downgrades the ability for goaltenders to win games. So while he's probably right there's some level of regression that may happen with Shesterkin, his model still always undervalues quality goaltending.

It's basically just a model looking at 5v5 possession and shot metrics, and acts like that's all that matters. We know we're not elite in that regard.

Now, do I think his model is actually qualified in predicting a Stanley Cup winner? Yeah, I honestly do think it has some validity given the importance of 5v5 play throughout the entire playoffs. But the model loses out on predicting literally anything else, including how to rank the teams that likely aren't going to win the Cup.

well, do you consistently expect Shesterkin to have historical goaltending in the modern era seasons?

possession and expected goals matters and most cup winners are elite at them. the Rangers aren't any different.

I think he’s basically tweaked the model to reduce goalie impact because goalies fluctuate far more than players, so it’s hard to use data modeling to get an accurate read. That said yeah it’s bad. Goalie is the singularly most important position.

Goalie is probably the least important position actually. 80% of the goalies in this league have no special impact and are completely interchangable (and it's probably more). It feels more important than it really is because it's a strong independent variable that can outweigh 5 v 5 play in small samples (along with special teams).

I'm not expecting any real drop off for Shesterkin. He's entering his prime, I think we can count on him being elite for at least a few years.

Is Halak dependable? I'm not sure.

he just had the best season any NHL goalie has had in at least 11 years. you expect that consistently?

 

[broadwayblue]

Lot's of hate. Personally, I thought the article was fine. He states multiple times that the model could be wrong, and that the team could win it all. But he also makes a number of arguments that could show how the team could regress. Do we really think we are a slam dunk to make it back to the ECF?

 

[nyr2k2]

Haven't read the article. I will say--Dom's model, from what I can recall, is generally very good at predicting the top handful of teams in the league. However, you and I and anyone who follows the game closely and isn't a homer, can also do that pretty reliably. Ask me who will be good: Florida, Toronto, Tampa, maybe Calgary, obviously Colorado, Carolina... I bet they're at the top of the list, right?

The not-obviously-elite teams are where it gets harder. And here, I also don't know if his model is any better than your average informed hockey fan.

The model is what it is. Take it or leave it. It's just one way of projecting things. If we ranked the teams, the lists would probably be close with the exception of things like, "Toronto always chokes so they drop five spots" and "The Rangers have clearly arrived and all the players will be better so they're a lock for the top five" which is stuff the model isn't going to capture, because it's based on quantifiable measures.

 

[HockeyBasedNYC]

Pffft

 

[mas0764]

Oh look, another self-proclaimed expert with their own fool-proof model.

Don't think even the authors are saying it's fool proof.

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

Yeah, look, another self proclaimed expert who thinks he's trendy by thumbing his nose at analystics in favor of his own complete layperson "eye test."

 

[mas0764]

I don’t think the regression here means Igor wouldn’t be elite. I think everyone knows he will be.

It’s whether or not you can count on him repeating last season which was one of the greatest regular seasons by a goalie in NHL history.

I think it’s a fair question to ask.

Yeah, I mean, even among elite, Hall of Fame goalies, there is regression from career peaks. Did we just see Igor's? I mean, his save percentage was like third all time wasn't it? Behind like, the elitist of the elite Dominick Hasek?

I think he’s basically tweaked the model to reduce goalie impact because goalies fluctuate far more than players, so it’s hard to use data modeling to get an accurate read. That said yeah it’s bad. Goalie is the singularly most important position.

I actually think goaltending has proven to be overrated in building your team.

The reactions to anytime Dom L says anything about the Rangers are exhausting

Just ignore the article and move on if you think it's so inaccurate. The Rangers have very real flaws that the vast majority of modern era cup winners do not overcome. Jesus christ just get over it.

Exactly.

Does this mean the Rangers can't be successful or even win a Cup? Of course not. Of course they could win a Cup.

But anything that points out their very real flaws and doesn't kiss their ass, people get out the pitchforks.

 

[mas0764]

Goalie is probably the least important position actually. 80% of the goalies in this league have no special impact and are completely interchangable (and it's probably more). It feels more important than it really is because it's a strong independent variable that can outweigh 5 v 5 play in small samples (along with special teams).

Preach it.

If Igor doesn't put up one of the best seasons all time again, we might feel a big sting, if we don't improve drastically at 5v5.