Gwyddbwyll's Ultimate Draft Ranking - all 30 teams

[KL*]

Had some time recently so I compiled a spreadsheet to analyse and rank the drafting records of all 30 teams. Since it's usually recognised a draft cant be fully judged until 5 years later, I started with the 2001 draft and went back 15 years to 1987 which covered 3893 draft picks in total. I defined a NHL player as those who played 100+ games. This may seem low to some but look at it this way. What is a NHLer? I'd define it as someone who played in the NHL for at least a full season's worth of games? That player has acheived more than over 80% of all draft picks.

I also added in two more indicators of draft success - rankings based on draft record not including first rounders and the ability of teams to find the highest quality players. These arent included.. they are just for additional reference. My final rankings are determined on one score alone - the number of NHLers a team finds, divided by the number of picks it had. I wanted it to be a purely statistical ranking rather than bring in more and more factors. Its complicated enough!

We all know how our team drafts and now we can compare to other teams accurately and on an equal basis. We can also use this data to see if our team's drafting is improving - predict how many of your players from the drafts in 2002-6 will make at least 100 games in the NHL and work out the percentage score yourself.

I will be doing this in time honored HF fashion.. the bottom 10 teams will follow shortly. Monday - the middle 10, Tuesday the top 10 teams. There are quite a few surprises in each one! I will answer any questions you have, but will be away from a computer over the holiday period. Cheers

PS I do have the full breakdown of how every team scored in every round but not sure how to put it up here in a neat fashion. If you want to know your team's score in each round, post on here and I'll reply to it later.

This has been done a million times before. I actually did something similar as my Masters thesis.

Your calculations are completely irrelevant, as they do not take into consideration the fact that nobody really has the same GM for 15 years, and there isn't a GM in the league who was the GM for his team 15 years ago.

That, compounded by the incredible rates of turnover in scouting staffs over the years means that your numbers aren't the least bit relevant to today's NHL.

Then you have to take into account draft position, philosophy (some teams build/t through the draft, while others use/d the picks in trades), free agency (teams that signed more UFAs had fewer spots to be won), minor league affiliates, injuries (what if Mario Lemieux broke his neck during his first game? does that make the Lemieux pick a bust? It's Mario Freaking Lemieux...) and the Iron Curtain. Oh, and there's always the fact that 100 games played for a goalie is a lot more than 100 games played for a forward. There are starting goalies in the NHL who take more than two seasons to accrue 100 games played.

Any one of these factors would have rendered your calculations useless. All of them combined have made your work completely pointless.

It is an interesting thing to look at though. You don't have complete and accurate numbers though. Unless there's a new source that came up since I did mine (did it in 2003), you don't have a complete set of stats for all picks, and hockeydb has stats for some players transposed for others.

 

[cheg1349]

I think some of you guys need to calm down. He's clearly just having some fun with this, and put some hard work into it. Instead of trying to find holes and picking apart what he did, why not just enjoy it and have fun with it?

 

[Gwyddbwyll]

Your calculations are completely irrelevant, as they do not take into consideration the fact that nobody really has the same GM for 15 years, and there isn't a GM in the league who was the GM for his team 15 years ago.

That, compounded by the incredible rates of turnover in scouting staffs over the years means that your numbers aren't the least bit relevant to today's NHL.

As said already, it's simply a ranking of achievement. There are many ways you can use the data for comparision to today's NHL.

Then you have to take into account draft position, philosophy (some teams build/t through the draft, while others use/d the picks in trades), free agency (teams that signed more UFAs had fewer spots to be won), minor league affiliates, injuries (what if Mario Lemieux broke his neck during his first game? does that make the Lemieux pick a bust? It's Mario Freaking Lemieux...) and the Iron Curtain. Oh, and there's always the fact that 100 games played for a goalie is a lot more than 100 games played for a forward. There are starting goalies in the NHL who take more than two seasons to accrue 100 games played.

I did take draft position into account since I have each team's score in every round. As far as injuries go, you just have to let luck even itself out. It could be argued that drafting high-risk players is bad scouting as well. Other teams had the chance to draft Bure and Fedorov. And teams that had fewer spots available can trade guys who are good enough to be NHLers.. many players make it at another team.

Any one of these factors would have rendered your calculations useless. All of them combined have made your work completely pointless.

Obviously I disagree. These are factors but from what I've seen, their effect is considerably lessened the larger the sample data becomes. If you dont like your team's ranking then feel free to bring up all the injured Marios you had or slightly lower picks. It might bump you up a spot or two but its really not going to make a big difference. It's worth introducing ways of improving the rankings but that is something for the future.

It is an interesting thing to look at though. You don't have complete and accurate numbers though. Unless there's a new source that came up since I did mine (did it in 2003), you don't have a complete set of stats for all picks, and hockeydb has stats for some players transposed for others.

I noticed some errors already. However I'm not going to go through 4000 draft picks and check the history of every single player! It is accurate enough. This is just a fun project, not nanotechnology. Thanks for your input. Your Masters project sounds interesting.. do you plan to post any of your findings on here?

 

[Gwyddbwyll]

The CBJ have Klesla, Johnson, and Leclaire (off the top of my head) logging major TOI, from those two classes.

Johnson and Leclaire are very late developers so they couldnt be included yet - you have to judge all teams by the same standards (Koivu also didnt meet the criteria for Minnesota).

 

[Kritter471]

Hah. Watch me lose. I didn't check Koivu because I was sure he'd have that by now, and I just assumed on the CBJ players. Shows me.

 

[KL*]

I think some of you guys need to calm down. He's clearly just having some fun with this, and put some hard work into it. Instead of trying to find holes and picking apart what he did, why not just enjoy it and have fun with it?

Maybe because some of us actually spent more than 20 minutes compiling our data, writing our analysis, seeking funding, etc., and we know the pitfalls of a half-hearted attempt at a statistical analysis.

 

[KL*]

As said already, it's simply a ranking of achievement. There are many ways you can use the data for comparision to today's NHL.

Not really. Your "analysis" took into consideration drafts that were pre- and post-Soviet bloc, mid- late- and post-firewagon eras, pre- mid- and post-trap eras, not to mention the fact that every team had at least one goon who easily racked up 100 NHL games, while that would not happen in today's NHL.

That's ignoring the fact that you are attempting to turn something analytical into something predictive.

If you had framed your "work" (put in quotes because you simply copied the work that others had done before you - someone on these boards does what you did at least once every six or eight months and comes here thumping their chest like you did, as if they just invented the wheel) as simply an analysis of an era in NHL drafting as opposed to some sort of analysis of how a current NHL team has drafted (which implies prediction of future draft acumen), then perhaps your "work" might have some validity. It doesn't, thus it doesn't.

I did take draft position into account since I have each team's score in every round.

I see. So the #31 overall and the #59 overall are equal, and should result in an equal number of "hits", all other things being equal? Had you run the numbers, you would know that this isn't the case.

That disparity becomes even more obvious in the first round. The #1 and the #29 are both first rounders. Some teams have a bunch of picks in the low 20s because they are good teams, while others (Quebec anyone?) have a bunch of picks in the top 5. By doing it by round, you put Quebec on equal ground with Detroit. You don't see the flaw in that logic?

That ignores the fact that you ran these numbers over the course of the expansion era. The #22 overall was a high 2nd rounder in 1987, and a 1st rounder in 2000. Doing it by round, you ignore that distinction.

As far as injuries go, you just have to let luck even itself out.

How convenient.

It could be argued that drafting high-risk players is bad scouting as well.

Not with any validity, you couldn't.

Drafting a high-risk player isn't bad scouting. It can be bad personnel management, but drafting is a decision made based upon scouting. Drafting isn't scouting.

Other teams had the chance to draft Bure and Fedorov. And teams that had fewer spots available can trade guys who are good enough to be NHLers..

Your numbers don't account for any of that.

Obviously I disagree. These are factors but from what I've seen, their effect is considerably lessened the larger the sample data becomes.

You're going to have to prove it. You say you've "seen" that. I want to see you do the work to prove it.

Show me specifically what, in the numbers, has "shown" you that these factors have been mitigated by sample size.

If you dont like your team's ranking then feel free to bring up all the injured Marios you had or slightly lower picks.

My team? You think my problem with your analysis is that you ranked my team too low? Who says my team was even ranked in your analysis? Who says I even looked at your rankings (I didn't...I knew by your stated research methods that your results were flawed beyond usefulness)?

It might bump you up a spot or two but its really not going to make a big difference. It's worth introducing ways of improving the rankings but that is something for the future.

You might want to work on making the rankings even the slightest bit useful in a present-tense sense before you act as if an improvement is something that you would merely consider.

I noticed some errors already. However I'm not going to go through 4000 draft picks and check the history of every single player!

You wouldn't know where to check in the first place. I'll give you a hint - without financial backing and a considerable time commitment, your numbers will never be accurate.

But it is quite telling that you think so little of the accuracy of your project that you would make excuses like that.

It is accurate enough.

Not for what you are portraying it to be.

This is just a fun project, not nanotechnology. Your Masters project sounds interesting.. do you plan to post any of your findings on here?

I am in negotiations with a publisher as we speak. If everything goes well, we're looking at a publication date sometime next fall.

 

[cstu]

Wow SCThrash, what is your problem? You may not think much of this project, but there are obviously a lot of other people interested. I fail to see why you need to take it so personally. Chill out.

 

[NashisCash]

Seriously, chill out Thrash, I hope some one comes on here and pisses all over your work.

The NHL draft as a Masters' thesis? Wow, sounds like you're doing some important work for humanity there

 

[Dominator13]

oh man, it sucks he stopped at 2001 cause the the Habs drafting sucked TILL 2001!

 

[Flannelman]

Love to see the rest!!

 

[wedge]

SCThrash, why don't you just shut up and avoid this topic?

Or if you're so smart, why don't you just give us access to your master thesis? The guy is doing that just for fun. He doesn't have 200 hours to spend on a project like that like you that. So just go away and let him have fun. His project is interesting. It's far from perfect but there is absolutely no need to keep bashing him.

good job Gwyddbwyll. I'm looking forward for the rest

 

[eddy]

This has been done a million times before. I actually did something similar as my Masters thesis.

Your calculations are completely irrelevant, as they do not take into consideration the fact that nobody really has the same GM for 15 years, and there isn't a GM in the league who was the GM for his team 15 years ago.

That, compounded by the incredible rates of turnover in scouting staffs over the years means that your numbers aren't the least bit relevant to today's NHL.

Then you have to take into account draft position, philosophy (some teams build/t through the draft, while others use/d the picks in trades), free agency (teams that signed more UFAs had fewer spots to be won), minor league affiliates, injuries (what if Mario Lemieux broke his neck during his first game? does that make the Lemieux pick a bust? It's Mario Freaking Lemieux...) and the Iron Curtain. Oh, and there's always the fact that 100 games played for a goalie is a lot more than 100 games played for a forward. There are starting goalies in the NHL who take more than two seasons to accrue 100 games played.

Any one of these factors would have rendered your calculations useless. All of them combined have made your work completely pointless.

It is an interesting thing to look at though. You don't have complete and accurate numbers though. Unless there's a new source that came up since I did mine (did it in 2003), you don't have a complete set of stats for all picks, and hockeydb has stats for some players transposed for others.

Rather then trash talking the work that this guy has done, why don't you show us yours if it's that much better.

 

[Michigan Wolverine]

His tone aside, I think SCThrash is just pointing out things that make this an imperfect analysis of the data. I don't think it's wrong to debate the methodology here. It wouldn't be worth debating had Gwyddbwyll presented data without an attempt to make a meaningful interpretation of it. But since he and others are going to argue that this rough data is meaningful, then it is worth discussing how the data should be analyzed.

 

[Gwyddbwyll]

SCThrash, firstly there are a few wrong assumptions you made with your criticisms. In general it seems you are going out of your way to slam me rather than ask. You are somewhat hostile and sarcastic and it doesnt seem as if you are genuinely interested in discussing this. This is a message board for enthusiasts and I am sharing my findings with anyone interested.

One thing I will state is that I did not copy anyone. I fully acknowledge hockeydb.com as the source of my data. If there have been millions of 15-year studies on here, then I'm unaware of them. I'd appreciate some links if there are so many of them? I'm also "trumpeting it" in a light hearted way as my Sensational Verdicts are supposed to imply - obviously that sentiment didnt register with you.

I wish you best of luck with your publication.

 

[jacketracket*]

Johnson and Leclaire are very late developers so they couldnt be included yet - you have to judge all teams by the same standards (Koivu also didnt meet the criteria for Minnesota).

"Late developers"? I'm not trying to bust your chops here, but it would seem as though they're either playing in the NHL or they're not.

What is the cut-off for a player being in the appropriate draft class, but not "developed" enough? Just sounds rather arbitrary and subjective.

 

[The Nemesis]

His tone aside, I think SCThrash is just pointing out things that make this an imperfect analysis of the data. I don't think it's wrong to debate the methodology here. It wouldn't be worth debating had Gwyddbwyll presented data without an attempt to make a meaningful interpretation of it. But since he and others are going to argue that this rough data is meaningful, then it is worth discussing how the data should be analyzed.

The whole problem is the tone though. There's no need for him to be so smug and arrogant about it. I don't think anyone will say that Gwyddbwyll's work would carry the same amount of depth as someone who put far more hours and dollars into their work. However, presented purely in the context that it is (this is a message board, after all) and with the somewhat tongue-in-cheek manner of the analysis, there's no need for SCThrash to thump his chest and say "my work was important, yours is worthless, therefore it should be ignored."

If such comments carried weight, then there'd be no need for HFboards at all, would there?

 

[shortcat1]

Had some time recently so I compiled a spreadsheet to analyse and rank the drafting records of all 30 teams. Since it's usually recognised a draft cant be fully judged until 5 years later, I started with the 2001 draft and went back 15 years to 1987 which covered 3893 draft picks in total. I defined a NHL player as those who played 100+ games. This may seem low to some but look at it this way. What is a NHLer? I'd define it as someone who played in the NHL for at least a full season's worth of games? That player has acheived more than over 80% of all draft picks.

I also added in two more indicators of draft success - rankings based on draft record not including first rounders and the ability of teams to find the highest quality players. These arent included.. they are just for additional reference. My final rankings are determined on one score alone - the number of NHLers a team finds, divided by the number of picks it had. I wanted it to be a purely statistical ranking rather than bring in more and more factors. Its complicated enough!

We all know how our team drafts and now we can compare to other teams accurately and on an equal basis. We can also use this data to see if our team's drafting is improving - predict how many of your players from the drafts in 2002-6 will make at least 100 games in the NHL and work out the percentage score yourself.

I will be doing this in time honored HF fashion.. the bottom 10 teams will follow shortly. Monday - the middle 10, Tuesday the top 10 teams. There are quite a few surprises in each one! I will answer any questions you have, but will be away from a computer over the holiday period. Cheers

PS I do have the full breakdown of how every team scored in every round but not sure how to put it up here in a neat fashion. If you want to know your team's score in each round, post on here and I'll reply to it later.

Sure, I'm game...

The Montreal Canadiens... I don't expect them to do that very well, their drafts in the 80's (late, especially) & 90's were not abominable by any means but they were often exercises in near futility.

I'm expecting them to be somewhere between 12 - 18. Tha's wide, I know but the 12 - 15 is probably a rose-coloured glasses generous estimate on my part (hoping for the best, that is...)

I'd appreciate it if you sent the reply to [email protected] also.

 

[Gwyddbwyll]

"Late developers"? I'm not trying to bust your chops here, but it would seem as though they're either playing in the NHL or they're not.

What is the cut-off for a player being in the appropriate draft class, but not "developed" enough? Just sounds rather arbitrary and subjective.

It is an arbitrary mark. I defined NHLers as having played a minimum of 100 games. I also took 2001 as the most recent draft to have "matured". Five years has been commonly accepted on here as a good benchmark by when you can start judging drafts properly. There are still a few players emerging from the draft.. all drafts continue to produce players after 5 years. One team in particular has a lot of them. You just have to set the mark somewhere to start the study in the first place. As it goes on, these two guys will be included next year and Columbus will improve noticably (something I did note in the summary). They dont get to benefit from late developers like teams with a full draft history in the study. I originally planned to leave out expansion teams entirely but decided to include them for interest. I'm guessing you're more interested than if I had left them out

 

[KL*]

Wow SCThrash, what is your problem? You may not think much of this project, but there are obviously a lot of other people interested. I fail to see why you need to take it so personally. Chill out.

Who is taking it personally? All I am doing is poking holes in his analysis because nobody else seems to be able to.

And as for other people being interested in the project, well, OF COURSE other people are interested in it. Otherwise, it wouldn't have been done half a dozen times over.

 

[KL*]

Seriously, chill out Thrash, I hope some one comes on here and pisses all over your work.

I am as "chill" as can be. And if you think you have the noodle to piss all over my work, well, find it and piss on it. The kind of person who would have the brainpower to piss on it would know where and how to find it.

The NHL draft as a Masters' thesis? Wow, sounds like you're doing some important work for humanity there

When you've done a Master's, get back to me.

 

[timlap]

22nd place - Toronto Maple Leafs - 18.5%

Inept in the first round, their successes are as mediocre as they come with Yanic Perreault as the pick of the bunch - although Boyes may well be better. The second round is a disaster zone. Toronto's drafting begins to redeem itself in rounds 3-6 finding 13 players. Quantity helped it up over quality. The Leaf Nation may not like their ranking but in truth, it could be lower.

Sensational Verdict: Who cares about quality!

I don't disagree that the Leafs did not draft especially well during those 15 years, but a few thoughts:

- Yanic Perrault was not a first rounder- you need to research more carefully.

- Our best first rounder during that phase, imo, was Kenny Jonsson (even if you counted YP). Luke Richardson, Drake Berehowsky, Scott Thornton, Grant Marshall, Nik Antropov, Brad Boyes all had some success. Overall, I agree it's not a strong group of first rounders.

- You hit on a bad period for Leafs first rounders. I'll take the four we've had since (Colaiacovo, Steen, Rask, Tlusty) as a higher quality group, though time will tell.

- The five years before '87 we had more success (while drafting higher, take note): Damphousse, Clark, Iafrate, Courtnall, Nylund.

- Did you only go up to round six? That leaves out some pretty good NHLers from each team.

- You hit on a bad period for Leafs first rounders. I'll take the four we've had since (Colaiacovo, Steen, Rask, Tlusty) as a higher quality group, though time will tell.

- The five years before 87 we had more success (while drafting higher, take note): Damphousse, Clark, Iafrate, Courtnall, Nylund.

- One problem with these attempts to summarise drafting is that they don't usually take into account what was received for picks that were traded. The point is to use your picks as assets to improve the team, whether by drafting or through trades.

 

[KL*]

SCThrash, why don't you just shut up and avoid this topic?

Because I don't want to. Why don't you take your own advice and stop whining about me?

Or if you're so smart, why don't you just give us access to your master thesis?

I never said I was smart. But nobody gives access to their theses pre-publication. That would be the dumbest thing I could do. If you had an advanced education, you wouldn't even ask.

The guy is doing that just for fun. He doesn't have 200 hours to spend on a project like that like you that.

I am sure he doesn't. And that's one reason why his work is so flawed.

So just go away and let him have fun. His project is interesting. It's far from perfect but there is absolutely no need to keep bashing him.

Nobody is bashing him.

 

[KL*]

SCThrash, firstly there are a few wrong assumptions you made with your criticisms. In general it seems you are going out of your way to slam me rather than ask.

1. I haven't slammed you. Not once. I slammed your work. If you cannot take criticism of your work, which you obviously cannot, then you have no business attempting any sort of academic pursuit. Because I will tell you from experience that the things I said were tamer than anything you will hear from an advisor or committee.

2. There's nothing to ask. I know what you did and how you did it. Thus, I know why and how what you did is flawed.

You are somewhat hostile and sarcastic and it doesnt seem as if you are genuinely interested in discussing this.

I am not hostile at all. You posted your little analysis as the "ultimate draft ranking." What, you wanted to use ridiculously flawed methodology and expected not to be called out for it when you refer to your work as the "ultimate"?

This is a message board for enthusiasts and I am sharing my findings with anyone interested.

That's fine, but you didn't find anything. You ran a quick analysis and you don't understand what you found, how it is valid and what its application is.

One thing I will state is that I did not copy anyone.

Besides hockeydb. But I never said you copied anyone.

I fully acknowledge hockeydb.com as the source of my data. If there have been millions of 15-year studies on here, then I'm unaware of them. I'd appreciate some links if there are so many of them?

You act as if a 15-year study took you 15 years. It's a simple Excel spreadsheet. Took you 20mins to do. And yes, similar projects have been posted here numerous times in the past. I don't keep a desktop rolodex of the links, but if you are interested, I assume you know how to search.

 

[King'sPawn]

The problem, SCT, is that you're treating him like he IS proposing a thesis, which, as you already said, would be a dumb thing to do pre-publication.

It's a rudimentary compilation of the team's drafts between 1987 and 2001, with limited resources. He explained his criteria. I didn't see where he was using it as a predictor. The only criticism I have is, instead of the editorial after each ranking, something more objective be put in that was factored into the ranking.

You're right that there are so many factors that would affect the ranking. Any methodology to rank these teams, much like the scouting and drafting of these teams, yields an imperfect system.

I suggest, instead of bashing his methodology, that you provide your own input since you've done this yourself. We already know it's not a master's thesis, so how about you lay off a little bit and not treat it as such.

And Gwyddbwyll, keep at it. Your effort is certainly appreciated, and it will be interesting to see how all the teams rank through your criteria.