How Would the Flames Handle a 2016 Top 3 Pick Next Season?

[Henry Fool]

Originally Posted by Enoch Root

For more clarity, Tkachuk's NHLe is a much more valid and useful stat. There are plenty of data points from the CHL to make his numbers a useful source of information.

But not the kids in Europe.

We already know what good CHL numbers are for that age group without taking any extra steps. It's just pointless number manipulation.

Or maybe someone should alert the scouts who are ranking Puljujärvi at 2-3 that when you turn his stats into other stats, they're still not that great.

[Enoch Root]

Originally Posted by Henry Fool

There's Barkov. Different leagues bring different types of players to the NHL, through different paths, at different ages. Finland for example simply hasn't produced scorers in many years.

Stats are of course relevant, but NHLe doesn't add to our understanding of them in any way as far as I can see. You can just look at stats and have enough of an idea without adding extra layers of ambiguous methodology.

Waste of energy as far as I have been able to tell.

Disagree. NHLe can be very useful. But it has to be used in context.

A couple recent examples with respect to the Flames:

Gaudreau: NHLe 65 (or was it 67 - something like that) actual pts 64

Monahan: NHLe 33, actual points 34

Bennett: NHLe 55 (IIRC), on pace for 40 pts.

But you need large amounts of data in order to build meaningful distributions.

[Enoch Root]

Originally Posted by Henry Fool

We already know what good CHL numbers are for that age group without taking any extra steps. It's just pointless number manipulation.

Or maybe someone should alert the scouts who are ranking Puljujärvi at 2-3 that when you turn his stats into other stats, they're still not that great.

It's not pointless.

Yes we know what the numbers in the CHL mean.

But NHLe helps to translate those numbers into comparisons with other leagues.

[Enoch Root]

Originally Posted by GranteedEV

Barkov. He had 61% more points per game than Puljujarvi in Liiga, and as an NHL rookie put up alright, but unimpressive stats. He's now as a 20 year old he's a 60+ point forward, which took quite a while for him to arrive at. So at the very least, this data tells us Puljujarvi is less productive than a young player who didn't really find his NHL offense until he was 20.

It's information that's obviously a lot more useful to the scouts than it is to us. I hope the scouts are looking at his stats and trying to understand the "why" behind them, rather than ignoring the stats outright.

It doesn't tell us that.

To be more accurate, it might tell us that, but when comparing two individuals, there are so many variables to consider.

Maybe Barkov was more physically mature at 17. Maybe Barkov had an injury in his post draft year that held him back.

Etc. etc. etc.

[Henry Fool]

Originally Posted by Enoch Root

It's not pointless.

Yes we know what the numbers in the CHL mean.

But NHLe helps to translate those numbers into comparisons with other leagues.

I totally fail to see it. There's a wealth of knowledge and experience with respect to how a certain type of player with certain stats projects. You don't need to calculate a theoretical model to predict his exact numbers in his rookie year. It simply abstracts away from the player's specific profile and situation.

[Henry Fool]

It actually does tell us that Barkov was more productive than Puljujärvi. Of course a five-second look at their basic stats tells that as well.

Then, like you say, we get into the specifics of individual players' player profiles, development, and playing situations in both leagues.

Where does this leave NHLe? It's not relevant at any point.

Edit: a response to the post above. Having trouble with the edit function.

[Enoch Root]

Originally Posted by Henry Fool

It actually does tell us that Barkov was more productive than Puljujärvi. Of course a five-second look at their basic stats tells that as well.

Then, like you say, we get into the specifics of individual players' player profiles, development, and playing situations in both leagues.

Where does this leave NHLe? It's not relevant at any point.

Again, NHLe is a macro-type stat that gives a backdrop for any player against the statistical distribution of all the players that came before them (well the past 15 or 20 years anyway).

That is a completely different type of analysis than comparing two players head to head.

But they both have a purpose.

[Jay Random]

Originally Posted by Henry Fool

I totally fail to see it. There's a wealth of knowledge and experience with respect to how a certain type of player with certain stats projects. You don't need to calculate a theoretical model to predict his exact numbers in his rookie year.

This is why you fail to see it.

It isn't about predicting a player's numbers in his rookie year. It's about making stats comparable between feeder leagues.

Is an 80-point season in the OHL better or worse than 40 points on an NCAA Div 1 team? Is a 22-year-old with 40 points in the AHL more or less likely to be NHL-ready than a 22-year-old with 30 points in Sweden? These are the kinds of questions NHLe is supposed to help you answer.

Incidentally, there is no theoretical model. It's just a matter of adding up all the point production of all the players who graduated from league X to the NHL, both their last year in league X and their first year in the NHL, and calculating the ratio between them. Nothing sophisticated about it, but it helps you gauge roughly how many firkins equal one hogshead.

[Henry Fool]

Originally Posted by Jay Random

This is why you fail to see it.

It isn't about predicting a player's numbers in his rookie year. It's about making stats comparable between feeder leagues.

Is an 80-point season in the OHL better or worse than 40 points on an NCAA Div 1 team? Is a 22-year-old with 40 points in the AHL more or less likely to be NHL-ready than a 22-year-old with 30 points in Sweden? These are the kinds of questions NHLe is supposed to help you answer.

Incidentally, there is no theoretical model. It's just a matter of adding up all the point production of all the players who graduated from league X to the NHL, both their last year in league X and their first year in the NHL, and calculating the ratio between them. Nothing sophisticated about it, but it helps you gauge roughly how many firkins equal one hogshead.

No, it's just that all of that is redundant and if taken too seriously actually obfuscates one's view of the players since we already have so much information about any individual player. I know the methodology. I can call it a model, you can call it something else.

To me it's doesn't add any new information and there are dangers in that because it adds extra steps of methodology that have to be interpreted as well.

[dammage79]

Gaudreau's NHLe was pretty accurate, but what of Barkov's NHLe? Or Mikael Granlunds at the time. Or Adrian Kempe's? Elias Lindholm's? Do you negate those draftees who did not play their first year in the NHL? or do you go with anyone? I'd be curious to know what other ranking for other Euro players looked at in high regard measure up. How accurate can NHLe be for the Euro leagues where they play against men?

For Euro players, it kind of feels like a game of lawn darts, it appears to me from a half assed glance that NHLe is more accurate to north american skaters.

[Henry Fool]

Originally Posted by Enoch Root

Again, NHLe is a macro-type stat that gives a backdrop for any player against the statistical distribution of all the players that came before them (well the past 15 or 20 years anyway).

That is a completely different type of analysis than comparing two players head to head.

But they both have a purpose.

I know what it is. It has a purpose, certainly. But if your aim is to predict a player's development, it is too abstract to tell us anything new about any individual player.

[Henry Fool]

Originally Posted by dammage79

Gaudreau's NHLe was pretty accurate, but what of Barkov's NHLe? Or Mikael Granlunds at the time. Or Adrian Kempe's? Elias Lindholm's? Do you negate those draftees who did not play their first year in the NHL? or do you go with anyone? I'd be curious to know what other ranking for other Euro players looked at in high regard measure up. How accurate can NHLe be for the Euro leagues where they play against men?

For Euro players, it kind of feels like a game of lawn darts, it appears to me from a half assed glance that NHLe is more accurate to north american skaters.

On Barkov's NHLe:
http://edmontonjournal.com/sports/ho...peter-forsberg

[Enoch Root]

Originally Posted by Henry Fool

I know what it is. It has a purpose, certainly. But if your aim is to predict a player's development, it is too abstract to tell us anything new about any individual player.

It isn't geared to predict a player's development.

It is geared to illustrate what might be expected from the player NEXT YEAR.

It has a purpose. It is not supposed to be the be all, end all.

[Enoch Root]

Originally Posted by Henry Fool

On Barkov's NHLe:
http://edmontonjournal.com/sports/ho...peter-forsberg

That shows that Barkov had a great pre-draft year.

But it is a ####show with respect to NHLe because most of those players didn't play in the NHL in the next year.

And the author is comparing points in the SM-Liga with points in the SEL (directly).

[Henry Fool]

Originally Posted by Enoch Root

It isn't geared to predict a player's development.

It is geared to illustrate what might be expected from the player NEXT YEAR.

It has a purpose. It is not supposed to be the be all, end all.

The ultimate goal is to predict a player's development and predicting next year's stats is intended to serve that purpose. But I think this difference in scope might partly explain our differences here.

One additional thing, separate from the above, to consider is that European leagues have changed many times over the time period. So for example saying that Barkov's NHLe is such and such compared to Näslund's is problematic from this perspective as well. Comparisons even within a single European league would be tough within such a timeline.

[Enoch Root]

Originally Posted by Henry Fool

The ultimate goal is to predict a player's development and predicting next year's stats is intended to serve that purpose. But I think this difference in scope might partly explain our differences here.

One additional thing, separate from the above, to consider is that European leagues have changed many times over the time period. So for example saying that Barkov's NHLe is such and such compared to Näslund's is problematic from this perspective as well. Comparisons even within a single European league would be tough within such a timeline.

Yes exactly, good point.

[Henry Fool]

Originally Posted by Enoch Root

Yes exactly, good point.

Also, in terms of player material, different countries produce different types of players, and during different periods. These are small countries that go up and down. Barkov for example was a first-line talent in a a country that hadn't produced one since the 90's.

[Enoch Root]

Originally Posted by Henry Fool

Also, in terms of player material, different countries produce different types of players, and during different periods. These are small countries that go up and down. Barkov for example was a first-line talent in a a country that hadn't produced one since the 90's.

With enough data though, NHLe would capture that.

Your first point was a good one though, because it speaks to the robustness of that data. If the characteristics of the subject data are constantly evolving and changing, then past data isn't much good for analyzing current subjects.

[Calgary4LIfe]

I agree that the article pushing the NHLe is a bit of a junk article. Good posts above explaining why.

Also, I would like to add this to some of the reasons why.

Doesn't Puljujarvi play on one of the better teams (as Laine does, but on another?). It is difficult for me to imagine they were simply given the reigns with top-line or top-six icetime, PP time, etc. I would bet that they started off pretty slowly and were given more ice-time as they earned it. Their first quarter/half/three-quarter season numbers might have been much lower than when they finally pushed themselves up on the top line (or just plain got the coach's trust) and have shown who they really are as players.

Given that they are so young, and the sample size so small, and the fact that they apparently both play for top clubs makes me not put too much stock in the numbers there.

Edit: Another explanation - though more of a reach - would be Stamkos' rookie season. Stated off with a coach that didn't like him in Tampa and gave him low minutes. Coach got canned, and Stamkos proceeded to explode. Without looking at the numbers more carefully, without having a larger sample size, and without really following the Finnish league, I find it difficult that the author could draw even modestly reliable data (no offence if he posts here). His example (which he quickly debunked) about basing the numbers off the World Junior Championships was arguably a better one. Much smaller sample size due to games played, but at least there is a much larger sample size of players that went through that tournament and into the NHL at a younger age. (And no, I don't think we can derive any conclusions out of the world championships - just merely pointing out sample sizes).

[Henry Fool]

You always end up discussing the individual situations which doesn't actually leave you anywhere with the NHLe numbers. Puljujärvi hasn't produced a lot of offense over the entire season - we've discussed it and at one point in the season I said that in my view he should have been falling in rankings because of that - but he has been an offensive contributor since the U20.

So scouts do what they always do, they figure out the hows and whys, and so far they are still saying that Puljujärvi is a top prospect. Let's not insult them by entertaining the idea that they're not interpreting a player's stats. To me it's odd to throw NHLe at them and proclaim that you've found something startling: 28 points in 50 games isn't that impressive!