NHL Thread

[Frag]

Analytics guys on Twitter seem to hate him so much. I mean I don't think he is a top 7 player but hes still a pretty good prospect. Marner seems to be pretty underrated though and I think he could be the steal of the draft.

 

He ranked 24th among CHL forwards after adjusting for height. There's definitely reason for concern when it comes to drafting Crouse in the top 10.

[burlingtonbandit]

He ranked 24th among CHL forwards after adjusting for height. There's definitely reason for concern when it comes to drafting Crouse in the top 10.

 

If its 10th overall thats not so bad. I think those rankings give a decent idea but like the author said take it with an extreme grain of salt. McDavid is 6th after all.

[Frag]

If its 10th overall thats not so bad. I think those rankings give a decent idea but like the author said take it with an extreme grain of salt. McDavid is 6th after all.

 

It is true, but unlike McDavid and Marner who are limited by the upper limit of the study, Crouse's PPG is within the limits (10% chance of playing 200+ NHL GP). That's concerning.

 

Also, and this is unrelated but funny anyways, draft year PPG was more highly correlated (R = 0.41) with NHL GP than CSS ranking (-0.19!). That's nuts! Before you say that it's The Hockey Writers, the author is Josh Weissbock, the same guy who did this.

[burlingtonbandit]

So one of the co-founders or co-owners of the side war-on-ice was hired as a consultant by the Oakland As. Lol the As now are hiring smart hockey guys.

[Frag]

So one of the co-founders or co-owners of the side war-on-ice was hired as a consultant by the Oakland As. Lol the As now are hiring smart hockey guys.

 

Doesn't shock me. Andrew Thomas has done tremendous work in building War-On-Ice after Extra Skater went down as well as working on a replacement for CapGeek.

[Nox]

This shows why you don't draft Lawson Crouse in the top 10.

 

I don't agree with this assertion at all.

 

If someone looks at his 7 assists this year and doesn't have the immediate first thought "that's probably really unlucky and that will most likely regress heavily to the mean", they need to try again.

 

Secondary assist rate is hugely random at the NHL level. My prior would be that it's entirely noise in Jr.

[BTS]

Ya I have G/60 at 370 minutes, A1/60 at 660 minutes, and A2/60 at 3360 minutes.

 

Wow

[GD]

I never did any math on that myself but I kinda had a hunch that was the case so I've always looked at G+A1 rather than points.

[Nox]

Wow

 

For Dmen at even strength, it's literally entirely random.

[Nox]

Ya I have G/60 at 370 minutes, A1/60 at 660 minutes, and A2/60 at 3360 minutes.

 

Without looking my stuff up, that seems to be in range for the aggregate population.

[Frag]

I don't agree with this assertion at all.

 

If someone looks at his 7 assists this year and doesn't have the immediate first thought "that's probably really unlucky and that will most likely regress heavily to the mean", they need to try again.

 

Secondary assist rate is hugely random at the NHL level. My prior would be that it's entirely noise in Jr.

 

The issue is that we don't have the numbers to look at Crouse's play better, like teammate WOWY and whatnot (this is where I miss Extra Skater). It could be he's unlucky or it could be he's not the best passer (or both). However, going by the information we currently have, his point totals have not matched up to what is expected out of a supposed top 5 pick in the upcoming draft and that's worrisome.

 

Edit: FWIW, there is one site that has estimated P/60. Crouse is currently sitting at around 1.71 eP/60.

Edited January 23, 2015 by Frag

[Frag]

I never did any math on that myself but I kinda had a hunch that was the case so I've always looked at G+A1 rather than points.

 

I do, too. Secondary assists are very low in value unless you're Tyler Bozak.

[BTS]

Wow, guys like Mitch Marner and Dylan Strome are putting up numbers that would have them 1st overall in a lot of years.

 

Edit; I guess Strome plays on a line with McDavid, so there's that. Everything I'm reading on Marner has him sounding pretty legit though.

[King]

Edit; I guess Strome plays on a line with McDavid

 

McDavid 1st line center, Strome 2nd line center

[burlingtonbandit]

Wow, guys like Mitch Marner and Dylan Strome are putting up numbers that would have them 1st overall in a lot of years.

 

Edit; I guess Strome plays on a line with McDavid, so there's that. Everything I'm reading on Marner has him sounding pretty legit though.

 

He's looked really good in the limited time I have seen him. He seems to have great hockey sense which is one of the things I like to look for and his hands are fantastic. So much talent in the draft.

 

Another guy who looked really good in the Prospects game is Konecny from the 67s.

[Nox]

However, going by the information we currently have, his point totals have not matched up to what is expected out of a supposed top 5 pick in the upcoming draft and that's worrisome.

 

The current information we have points to him being underrated by raw point totals.

[GD]

Would be nice to repeat the study with goals + A1's rather than points.

[Frag]

The current information we have points to him being underrated by raw point totals.

 

Did you see that link I posted? Anyway, even if his assist totals regressed (in the past, he mostly had more goals than assists in seasons, but not in the level it currently is; plus, scouting reports mostly rave about his goal scoring, not his passing), would you say he's worth picking in the top 10 of this year's draft? I wouldn't.

 

Also, Crouse is not a dman.

 

Edit: So, I counted how many primary/secondary assists Crouse had in each game this season by boxscore stats (thankfully, OHL posts who got first and second assists). So far, he has 3 primary assists and 4 second assists this year. Last year, he had 8 primary assists and 4 secondary assists.

Edited January 23, 2015 by Frag

[Nox]

Did you see that link I posted? Anyway, even if his assist totals regressed (in the past, he mostly had more goals than assists in seasons, but not in the level it currently is; plus, scouting reports mostly rave about his goal scoring, not his passing), would you say he's worth picking in the top 10 of this year's draft? I wouldn't.

 

Also, Crouse is not a dman.

 

I feel like you don't understand something. You're saying Crouse shouldn't be drafted because of his low assist totals (you said points but given that his goal rate is fine for his age, a problem with his perceived assist rate is the simple deduction). I'm saying that's a really noisy measure and that our best guess of his future NHL assists would use very little from his measured past Junior assist totals.

 

I don't even know if he's worthy of a top 10 pick. I just think to say that he's not a top 10 based on the simple fact his assist rate is lower than you expect from a top prospect is not the right way to think about it.

 

And where did I say that he's a dman? If he was, he'd obviously NOT be in the conversation for a top 5 pick (over drafting of dmen is well documented and anecdotally obvious).

[Nox]

Edit: So, I counted how many primary/secondary assists Crouse had in each game this season by boxscore stats (thankfully, OHL posts who got first and second assists). So far, he has 3 primary assists and 4 second assists this year. Last year, he had 8 primary assists and 4 secondary assists.

 

And what amount of predictive power does that have in terms of his future JR assist rate? His NHL assist rate?

 

Other holes in the study:

 

1) Using career GP is dumb. A team should only be focused on the years in which the player is under control after the draft.

2) Using GP as a measure of performance is inherently shaky. All sorts of biases are introduced.

[Frag]

I feel like you don't understand something. You're saying Crouse shouldn't be drafted because of his low assist totals (you said points but given that his goal rate is fine for his age, a problem with his perceived assist rate is the simple deduction). I'm saying that's a really noisy measure and that our best guess of his future NHL assists would use very little from his measured past Junior assist totals.

 

I feel like we misunderstood each other here. I didn't say Crouse shouldn't be drafted, as I think he's still 1st round material. I just don't think he's worthy of a top 10 pick.

 

I don't even know if he's worthy of a top 10 pick. I just think to say that he's not a top 10 based on the simple fact his assist rate is lower than you expect from a top prospect is not the right way to think about it.

 

What is the right way to think about it with regards to CHL players in your opinion? All we have are scouting reports and a few statistics, some of which need to be inferred (eg. eP/60). Where scouting is concerned (at least, scouting from the CSS), the CSS in the past has done a far worse job of predicting whether prospects will play at least 200 NHL GP than simple draft year points per game.

 

As well, read this recent article from the same author about Crouse. He is given a lot of ice time (mind you, he kills penalties, which shoots up his TOI) and plays with good teammates, yet still isn't producing the point totals you expect from a top 10 pick. That's concerning to me, even if he's been unlucky with the assists. That quote about size in the article sums up why scouts love Crouse so much, and it has less to do with his skill set and more how well he hits and kills penalties. Even though I think Crouse will be better, that gives me Tyler Biggs flashbacks.

[Nox]

GP obviously has a size bias, which is what they are trying to test for. Pretty much meaningless.

 

Then there's the survivorship bias stuff too. Guys who are luckier than normal (unrealistically high shooting %s, assist rates etc) will get more games than they should. Really unlucky guys will be labeled as guys who can't "bear down on chances" or something dumb and go play in Europe.

[Frag]

And what amount of predictive power does that have in terms of his future JR assist rate? His NHL assist rate?

 

Other holes in the study:

 

1) Using career GP is dumb. A team should only be focused on the years in which the player is under control after the draft.

2) Using GP as a measure of performance is inherently shaky. All sorts of biases are introduced.

 

Testing for GP is not dumb. If you want to test whether the player you pick will even play NHL games, then GPs is a fine way to look at it. If you want to look at quality of those picks, then you are right.

[Frag]

GP obviously has a size bias, which is what they are trying to test for. Pretty much meaningless.

 

I think this is more of an issue with the study than Nox pointed out (which is more random variation than bias) and is something I did ask the author about.

[Nox]

I feel like we misunderstood each other here. I didn't say Crouse shouldn't be drafted, as I think he's still 1st round material. I just don't think he's worthy of a top 10 pick.

 

Yea, honestly he might not be a top 10 guy, especially in this draft. I haven't looked close enough to have a true evaluation of his asset value.

 

 

 

What is the right way to think about it with regards to CHL players in your opinion? All we have are scouting reports and a few statistics, some of which need to be inferred (eg. eP/60). Where scouting is concerned (at least, scouting from the CSS), the CSS in the past has done a far worse job of predicting whether prospects will play at least 200 NHL GP than simple draft year points per game.

 

As well, read this recent article from the same author about Crouse. He is given a lot of ice time (mind you, he kills penalties, which shoots up his TOI) and plays with good teammates, yet still isn't producing the point totals you expect from a top 10 pick. That's concerning to me, even if he's been unlucky with the assists.

 

Any analysis HAS to at the very least split his playing time into strengths.

[Frag]

Any analysis HAS to at the very least split his playing time into strengths.

 

This is a fair point. I really miss Extra Skater's CHL statistics, since it included even strength points/G and P/60.

[Frag]

But aren't you the person (GM) who decides if your pick gets NHL time or not? It's not like baseball where they have to progress through multiple levels. Size and some scoring ability gets you in.

 

I'm not sure I'm getting you. 1st round picks tend to play more games than picks in later rounds for the reasons you specified (plus the GM giving 1st round picks more benefit of the doubt relative to later picks). However, the reason why such players play more games could also be due to the fact that they are actually talented players (Forwards picked in the 1st round have a higher career NHL PPG than forwards picked in later rounds). If you want to look at whether such players actually play NHL games in the long run, then GP is the measure you look at (until hockey comes up with its own WAR-like stat). There are players who don't make the NHL despite being picked high by GMs, too.

 

Also, I'm not sure baseball is a good example of what you were getting at. The GM of the baseball team could technically rush prospects to the major league level. They just don't do so most of the time because it could ruin the prospect.

[Spanky99]

ITT: ROBOTS

 

 

 

 

 

 

cf

[burlingtonbandit]

NHL allstar game has to be one of the worst sporting events. Its getting worse than the ProBowl. MLB and NBA all star games are such big things for the leagues but NHL treats it as crap.

[Sammy225]

NHL allstar game has to be one of the worst sporting events. Its getting worse than the ProBowl. MLB and NBA all star games are such big things for the leagues but NHL treats it as crap.

 

2 things that always bugged me about the NHL. The All star game is a joke and it runs too long into summer. I think they should just skip it so it cuts the league down a week lol