Minnesota Gophers Season Thread: 2014-2015

[captain beefheart]

Re: Minnesota Gophers Season Thread: 2014-2015

For those of you not into stats read the following at your own risk!

I meant a negative, not weak correlation and exactly what you thought I didn't mean. I should preface this by saying my statement about a negative performance correlation between the USHL and D1 needs more analysis. Out of curiosity, I did do an informal study of USHL 20+ goal scorers compared to a USHL baseline of 7 or less goal scorers. The sample population of interest was approximately 40 players consisting of leading USHL goal scorers, Gopher players and others who played in the USHL and D1 from 2004-2013.

With the same amount of games played, the findings indicated USHL players who scored 20+ goals showed a 25-35% (in several cases > 35%) reduction in goal scoring in D1 college hockey. On the other hand, USHL players who scored 7 goals or less showed only a 15-25% drop in scoring in D1 college hockey. This suggests that an increase in goals scored in the USHL may have a weak negative correlation with the number of goals scored in D1 college hockey.

This is an informal study. I would need to run ANOVA, Pearson correlation and ad hoc procedures (p = .05) with a larger randomized sample size to prove my hypothesis which I'd like to do someday when I have time. But given the increase in the parity of college hockey talent over the past several years, I would expect to find the sampling variance between the two frequency distributions to be significant, meaning the coefficient of determination is fairly high (65-70%). The closer the effect is to -1.00 the greater negative correlation (I would expect a weak -.10 to -.30) and predictability indicating the drop in scoring may be due to a weak negative correlational relationship in goal scoring between USHL and D1. But this is only stating the possible presence of a relationship, but keep in mind this cannot indicate anything in regards to its source, confounds or causal inference.

There were a few Gopher outliers in the past like Ryan Potulny, Kyle Okposo, Erik Haula to name a few. But there seemed to be a lots of confirmation data with Gopher players like Howe, Cepis, Hansen, Bostrom, Ambroz, Warning, Condon, Kloos, Cammarata, Lettieri, etc. who were all USHL 20+ goal scorers who significantly declined in goal scoring since playing D1 with the Gophers. This was the case with other USHL players on other D1 teams as well.

Uhhh...that means we won, right?

 

[SJHovey]

Re: Minnesota Gophers Season Thread: 2014-2015

Uhhh...that means we won, right?

I've never seen a guy who tries harder on this board to convince people he's super smart, but seems to lack basic common sense.

What he just told you is this:
He selected 40 former USHL players.
Some had scored 20+ goals in the USHL, some had scored 0-7 goals in the USHL.
Production across the board decreased when they got to college (who'd thunk that).
The decrease in goals scored by those players who put up large numbers in the USHL was greater than the decrease observed in those players who didn't score hardly at all in the USHL (another huge surprise).

But, of course, this is the same guy who said that even if the gophers won their last 16 regular season games they still wouldn't be in an at-large position in the pwr, even though as some pointed out those 16 wins would have put the gophers at a win total equal to how they finished the entire season last year, making Harley's theory obviously nonsense.

I'm just glad to see in the MN-Mich thread he has backed off on some of the crazier aspects of his gopher pwr theory.

 

[5mn_Major]

Re: Minnesota Gophers Season Thread: 2014-2015

For those of you not into stats read the following at your own risk!

I meant a negative, not weak correlation and exactly what you thought I didn't mean. I should preface this by saying my statement about a negative performance correlation between the USHL and D1 needs more analysis.

Not sure I caught your full position. But isn't this just an example of a larger and more normal trend...that of a decreased scoring output as all players move up in competition? On average, a kid will likely score more in MN high school than in the USHL. And more in the USHL than in D1...and more in D1 than in the pros. Of course, there are exceptions...but normally this is true. So I think you could do a study and find that players score much more in MN high school than in D1 also. So assuming as you are that the driver of much higher scoring is an easier league (which may not be the case)...scoring in the USHL doesn't need to keep up with D1 scoring for it to be a great league; it needs to stay closer to D1 than does high school or other competing leagues.

 

[Bonin21]

Re: Minnesota Gophers Season Thread: 2014-2015

So the players that scored more in the USHL scored more than their counterparts again when they made to the NCAA?

 

[HarleyMC]

Re: Minnesota Gophers Season Thread: 2014-2015

Not sure I caught your full position. But isn't this just an example of a larger and more normal trend...that of a decreased scoring output as all players move up in competition? On average, a kid will likely score more in MN high school than in the USHL. And more in the USHL than in D1...and more in D1 than in the pros. Of course, there are exceptions...but normally this is true. So I think you could do a study and find that players score much more in MN high school than in D1 also. So assuming as you are that the driver of much higher scoring is an easier league (which may not be the case)...scoring in the USHL doesn't need to keep up with D1 scoring for it to be a great league; it needs to stay closer to D1 than does high school or other competing leagues.

The point was a poster asked about what I meant by "negative" performance correlation as opposed to "weak". That means at a certain ceiling, the more goals scored in the USHL, the less goals will be scored in D1 based upon a baseline measure.

So the players that scored more in the USHL scored more than their counterparts again when they made to the NCAA?

If by counterpart you mean the baseline, then yes which means the predictors explain the variance.

 

[5mn_Major]

Re: Minnesota Gophers Season Thread: 2014-2015

Gotcha

 

[Bonin21]

Re: Minnesota Gophers Season Thread: 2014-2015

So the "What does this mean to me?" is that getting higher scorers in the USHL is better for your team. Well that felt like a lot of reading.

 

[HarleyMC]

Re: Minnesota Gophers Season Thread: 2014-2015

So the "What does this mean to me?" is that getting higher scorers in the USHL is better for your team. Well that felt like a lot of reading.

Maybe, but not necessarily because at a point in the distribution it becomes an inverse (negative, not weak) relationship and the return on investment may diminish drastically. That translates into failed expectations for a D1 team that is seeking to structure their offense around a perceived USHL top scorer.

 

[5mn_Major]

Re: Minnesota Gophers Season Thread: 2014-2015

Look for the Gophers to always play Michigan hard. Our play style of integrated 5 player puck domination...will typically take care of their blitzkrieg of individual efforts. In this case, our 'rock' play style will give us an advantage to their 'scissors' play style more times than not.

 

[Koho]

Re: Minnesota Gophers Season Thread: 2014-2015

So MN is back in the Pohl. Good to see. But why is it that Eichel U. doesn't get a single first place vote?

 

[William Blake]

Re: Minnesota Gophers Season Thread: 2014-2015

So MN is back in the Pohl. Good to see. But why is it that Eichel U. doesn't get a single first place vote?

Just lost to well under .500 UNH?

 

[Slap Shot]

Re: Minnesota Gophers Season Thread: 2014-2015

Harley, based upon your own math a previous +20 goal scorer is still more likely to score more goals at the next level compared to a previous <7 goal scorer despite the probability of.a greater % drop in production. Having said that there's no perfect predictor and even hockey teams need ditch diggers, but you seemed to go a long ways there for nothing.

 

[HarleyMC]

Re: Minnesota Gophers Season Thread: 2014-2015

Harley, based upon your own math a previous +20 goal scorer is still more likely to score more goals at the next level compared to a previous <7 goal scorer despite the probability of.a greater % drop in production. Having said that there's no perfect predictor and even hockey teams need ditch diggers, but you seemed to go a long ways there for nothing.

I was asked a question by another poster on negative performance correlation. I don't mind other posters responding to it to add to the broader discussion of the informal results if the original point of the hypothesis is understood. What you described may be true, but that's not the point of my initial comment and response to the other poster. As I have said a few times a threshold rotation of the regression line in the distribution may result in a significantly lower ROI because the correlation is weak negative NOT weak positive. If I run several inferential stat procedures I can determine at what point the main effect is significant with at least a 95% confidence level of predictability. To my knowledge this has never been done before. The initial indication is it may have some interesting applications in recruiting strategies.

I disagree that I did this for nothing, as I sufficiently answered the original posters inquiry that was addressed to me. Nor am I trying to impress anyone here as I assumed the poster understood something about stat procedures based upon his response to my comment, so I responded in kind.

I've never seen a guy who tries harder on this board to convince people he's super smart, but seems to lack basic common sense. INCORRECT

What he just told you is this:
He selected 40 former USHL players. INCORRECT
Some had scored 20+ goals in the USHL, some had scored 0-7 goals in the USHL. INCORRECT
Production across the board decreased when they got to college (who'd thunk that You forgot to write a verb Caveman...haha).
The decrease in goals scored by those players who put up large numbers in the USHL was greater than the decrease observed in those players who didn't score hardly at all in the USHL (another huge surprise). INCORRECT

But, of course, this is the same guy who said that even if the gophers won their last 16 regular season games they still wouldn't be in an at-large position in the pwr, even though as some pointed out those 16 wins would have put the gophers at a win total equal to how they finished the entire season last year, making Harley's theory obviously nonsense. INCORRECT

I'm just glad to see in the MN-Mich thread he has backed off on some of the crazier aspects of his gopher pwr theory. INCORRECT

The Whioux fan sounds a little bitter. For a confused old stalker who claims to have a lot of common sense you missed the mark...again. I noted several errors in your post and I can provide evidence to support all of it. However, I won't take the time to do that because you're so "super smart" you should be able to figure it out.

 

[goldy_331]

Re: Minnesota Gophers Season Thread: 2014-2015

So the "What does this mean to me?" is that getting higher scorers in the USHL is better for your team. Well that felt like a lot of reading.

Sounds like a typical Harley post. One of the reasons I have him on ignore.

 

[HarleyMC]

Re: Minnesota Gophers Season Thread: 2014-2015

Sounds like a typical Harley post. One of the reasons I have him on ignore.

This is too funny. Says the guy who still thinks I'm a Whioux fan. And I suppose you will go to the grave thinking that since you can't read my post.

 

[D2D]

Re: Minnesota Gophers Season Thread: 2014-2015

Good article on Leon Bristedt, talking about the difficulties of adopting to life and hockey here. The kid has really picked up his game lately.

Gophers freshman Bristedt learning on the fly
http://www.startribune.com/sports/292831511.html

 

[5mn_Major]

Re: Minnesota Gophers Season Thread: 2014-2015

Good article on Leon Bristedt, talking about the difficulties of adopting to life and hockey here. The kid has really picked up his game lately.

Gophers freshman Bristedt learning on the fly
http://www.startribune.com/sports/292831511.html

Absolutely. Based on his prior resume, I expected him to have a frosh year much like Kloos. Its taken him a bit longer than I expected, but he's actually turning into a Kloos/Cammy hybrid. Definitely will be a critical player going forward.

 

[D2D]

Re: Minnesota Gophers Season Thread: 2014-2015

Speaking of Kloos, what a huge goal yesterday. He's scored on that same sweet move across the middle a couple of times before. Faked that Penn State D right out of his breezers!

 

[Koho]

Re: Minnesota Gophers Season Thread: 2014-2015

Absolutely. Based on his prior resume, I expected him to have a frosh year much like Kloos. Its taken him a bit longer than I expected, but he's actually turning into a Kloos/Cammy hybrid. Definitely will be a critical player going forward.

I like the way Bristedt has been playing lately too, but am still not sure he is quite living up to the pre-season hype of some. As good as he has looked in the last month, I wouldn't put him at the same level as Kloos last year. Kloos was single-handedly taking over shifts at times and scoring. Bridstedt is keeping plays alive and helping his linemates look better at times, and making some impressive stickhandling in limited space, but still hasn't had the take-over play and score opportunity. He may turn into a great player and he is playing very well, just splitting hairs a little.

 

[Slap Shot]

Re: Minnesota Gophers Season Thread: 2014-2015

A lot Id say. What more from that article and recent play do you need to reconcile the results vs expectations?