[CP Story] Draft History Analysis - The "Hossa Rule"

[Bingo]

Part one of either two or three this week.

Examining my old "Hossa Rule" across a ten year data set (2004-2013), Calgary doesn't do well.

Thoughts and comments welcome as always

http://www.calgarypuck.com/2015/06/n...he-hossa-rule/

[UKflames]

Interesting theory and a really good read, thanks.

[Badgers Nose]

That is very cool. I like pretty charts.

[madmike]

Great article. 04, 05 and 06 were really depressing drafts.

[Red]

2004 Sutter traded down when Zajac was still on the board. Took Chucko instead ugh....

Trading down did not work out other years either, I believe Pelech was also a result of one.

[theslymonkey]

The trade down for Jankowski actually could be looked at losing Tereveinen. He had fallen to the Flames and I was pumped and then we pass...

Hopefully Janko turns out to be the diamond that Feaster promised...

[Bingo]

I was really surprised at how quantity beat draft acumen almost 100% of the time.

Clearly there are particulars that defy the numbers though. Take the Kings in 2005. They draft 2 guys that play and have 7 others that bust completely. The end up with a -475 in games played suggesting failure in crude terms. They miss out on Marc Staal, Mason Raymond, Kris Letang so not a good draft right?

The two players that play? Kopitar and Quick.

[ricardodw]

Originally Posted by Bingo

I was really surprised at how quantity beat draft acumen almost 100% of the time.

Clearly there are particulars that defy the numbers though. Take the Kings in 2005. They draft 2 guys that play and have 7 others that bust completely. The end up with a -475 in games played suggesting failure in crude terms. They miss out on Marc Staal, Mason Raymond, Kris Letang so not a good draft right?

The two players that play? Kopitar and Quick.

The fact that quantity beats draft acumen has been what Trevling and the Flames have been preaching with their talk about ledges. They have sort of admitted that when it comes to a particular teams pick there might be 5 guys evaluated to be equivalent prospects.

The Levels are smaller in the first round. The top group of 4 last year to a group of 5-6 and then a group of 10.

I would think that if you limited your analysis to the first 2 rounds, because the bigger breaks in levels draft acumen wins out over quantity....

also your statistics do not do justice to drafting a D-man for the last 4 years.. A forward who does not make it to the NHL by 22 is almost done. A defense-man who does make it by 22 is a great pick. Forward will have a lot more games than a d-man.

[edslunch]

If I understand your methodology you're comparing games played by the Flames' pick to the games played of BOTH the next two players added together? Wouldn't it make more sense to compare to average it the greater of the two?

[Hockey Fan #751]

One thing I think that skews the data is terrible teams with a high "ranking" as when you're consistently drafting in the top 5 like Edmonton or Buffalo the players you get are almost certain to play a large number of games. And the two players following are generally not going to be as good since the rankings are much more defined right at the top.

[Hockey Fan #751]

Oh, another thing is Columbus got roasted in 2004. If you did 2005-2013 instead they would be right at the top instead of 22nd.

[Bingo]

Originally Posted by edslunch

If I understand your methodology you're comparing games played by the Flames' pick to the games played of BOTH the next two players added together? Wouldn't it make more sense to compare to average it the greater of the two?

Not quite

I compare the drafted player to the player taken in the next two picks who has the most games played.

So if the Flames take Joe Smith and he plays 100 games, but the next player taken was Kevin Brown who played 200 and then next up went Steve Green who played 250 the Flames would get a -150 (Smith vs Green), not -250 Smith versus both guys.

[Bingo]

Originally Posted by Hockey Fan #751

One thing I think that skews the data is terrible teams with a high "ranking" as when you're consistently drafting in the top 5 like Edmonton or Buffalo the players you get are almost certain to play a large number of games. And the two players following are generally not going to be as good since the rankings are much more defined right at the top.

Exactly, and that's up tomorrow or Wednesday

[edslunch]

Ok, I misunderstood. Makes sense

[Bingo]

Originally Posted by ricardodw

I would think that if you limited your analysis to the first 2 rounds, because the bigger breaks in levels draft acumen wins out over quantity....

also your statistics do not do justice to drafting a D-man for the last 4 years.. A forward who does not make it to the NHL by 22 is almost done. A defense-man who does make it by 22 is a great pick. Forward will have a lot more games than a d-man.

Agreed on your first point, had that same thought. I broke the draft by first 60 picks to avoid the chaos of a moving 2nd round with compensation picks but the same idea

I think you have a point on dmen maturing slower, so that would be a factor for 2009-2013 drafts for sure. Less of an impact over the course of a career though.

[mikephoen]

Interesting article and I really appreciate people finding new ways to look at things like this.

That said... something doesn't feel right about how this turned out. I think there must be too much luck or variance even with the fairly large sample size. I say this because cup winners LA and Chicago are right down at the bottom and horrid teams like Buffalo and Edmonton are right at the top.

Later round picks probably really screw this up. One guy who plays 800 games who came from the 6th round for example hammers down 2 teams score, but really all 30 teams missed this guy for multiple rounds but don't 'suffer' in the rankings for it. Anything after the 3rd round is such a crap shoot I don't know if it's fair to use that data.

Looking forward to the rest of the analysis though!

[Bingo]

Originally Posted by mikephoen

Interesting article and I really appreciate people finding new ways to look at things like this.

That said... something doesn't feel right about how this turned out. I think there must be too much luck or variance even with the fairly large sample size. I say this because cup winners LA and Chicago are right down at the bottom and horrid teams like Buffalo and Edmonton are right at the top.

Later round picks probably really screw this up. One guy who plays 800 games who came from the 6th round for example hammers down 2 teams score, but really all 30 teams missed this guy for multiple rounds but don't 'suffer' in the rankings for it. Anything after the 3rd round is such a crap shoot I don't know if it's fair to use that data.

Looking forward to the rest of the analysis though!

Agreed

Two weeks ago I asked people to name "good drafting teams" as a I thought I'd run the numbers for what people think are the best and see how they compare.

Then I got addicted to it and did all 30 teams for 10 years to get a more comprehensive study.

But my marker was always Edmonton, and you're right about having high picks and being bad teams getting skewed. I cover that next.

[Red Menace]

I started doing something similar to this where i tried to compare each pick to the next 9 selections to see how the pick ranked among those 10 picks...but it got to be so time consuming I gave it up.
I would say that this type of analysis weighs heavily towards the earlier years of the date range (as they obviously have more games available), so the Flames poor records in 2003-2008 are sure to overshadow any good results we've had over the past 5 years.... perhaps that is obvious.

[Oil Stain]

Originally Posted by mikephoen

Interesting article and I really appreciate people finding new ways to look at things like this.

That said... something doesn't feel right about how this turned out. I think there must be too much luck or variance even with the fairly large sample size. I say this because cup winners LA and Chicago are right down at the bottom and horrid teams like Buffalo and Edmonton are right at the top.

I think luck plays a bigger part in drafting than most people care to admit.

One thing to keep in mind about Chicago is that they have had loads and loads of picks over the years. They had 17 selections in 2004. In 2010 they had 4 second rounders and they all busted.

It's easy to overlook stuff like that when the NHL team is stacked with talent though.

[CSharp]

God, why was Sutter ever the Flames GM in the first place and lasted that long in that position really puzzles me. After his dismissal, everything started to look good! Still can't believe he drafted a guy named Chucko! He must've had a great chucko over that!