HFBoards

Go Back   HFBoards > General Hockey Discussion > By The Numbers
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
By The Numbers Hockey Analytics... the Final Frontier. Explore strange new worlds, to seek out new algorithms, to boldly go where no one has gone before.

First round pick historical performance

Reply
 
Thread Tools
Old
08-16-2012, 03:05 PM
  #1
Sureves
Registered User
 
Sureves's Avatar
 
Join Date: Sep 2008
Location: Ottawa
Posts: 7,131
vCash: 500
First round pick historical performance

Hey guys, this is from draft weekend, thought I'd post it in the new forum:

Being that the draft is on Friday, I thought it would be a good time for me to post this.

Basically I've calculated the cumulative goals, assists, and point totals for all forward draft picks between 2000 and 2007. I use a (perhaps bold) assumption that forwards drafted in the first round are meant to put up points.

The tables below document the results, I considered any player who played over 160 games in the NHL to be an "NHL player". As such, when you look at the average goals, assists, and points, keep in mind that those numbers are only looking at "NHL players" (ie. it does not include the busts that never even played in the NHL, and for that reason those numbers are inflated).

For the first 5 picks I separated them individually because there is allegedly a pretty big difference between picking 5th and picking 4th, or picking 1st instead of 2nd, etc. It should be noted that doing it that way reduces our sample size substantially, so once we got to pick 6,7,8 I began to pool them together assuming (perhaps boldly) that the picks are equivalent in terms of player talent. Once we got to pick 12, the pools got larger again to include 5 picks.

To begin, I first calculated how many players scored less than 0.50 points per game for all of the draft positions, and then calculated similar numbers for different point ranges.



In other words, if we look at #1 overall for example, 5 players were drafted, all 5 of which played over 160 games. 0 of them scored below 0.50 points per game, all 5 of them scored above 0.50 points per game, all 5 above 0.60 points per game, all 5 above 0.70 points per game, all 5 above 0.80 points per game, 4 of them scored over 0.90 points per game, and 3 of them scored above 1.00 points per game.

That's how you should read the chart above.

The table below shows a clearer breakdown of the actual distribution of players who scored between a range of points per game. I personally think this table better illustrates what's going on. It also calculates the average pace of a player drafted over an 82 game season - again keep in mind that these numbers are inflated since it excludes players who never made the NHL in the first place



Lastly, a percentage breakdown. This means this takes the numbers from the table above and divided by the number of NHL players (ie. divides by the number of players who played over 160 games).



It should be noted that those with larger pools of draft picks (rather than just looking at 1 specific draft position for instance) are much more statistically significant and reliable. Caution should be exercised in looking at small sample sizes.

I'll leave the bulk of the analysis to yourselves and you can interpret as you wish, but one thing I will say is that it makes a lot of sense why those who pick #1,2 are generally very cautious about trading down, because it looks like the #1 is almost a surefire to be a superstar, and #2 a solid first line player, whereas #3 and below has a lot more risk involved in those picks.

Once you get outside the top 5, there's only about a 70% chance that your pick is going to make the NHL at all, and from there, the point totals aren't exactly mesmerizing.

Just my two cents. Have at it.

Sureves is offline   Reply With Quote
Old
08-16-2012, 03:24 PM
  #2
Czech Your Math
Registered User
 
Czech Your Math's Avatar
 
Join Date: Jan 2006
Location: bohemia
Country: Czech_ Republic
Posts: 3,665
vCash: 500
Good work. I think your sample size needs to be much larger to draw many conclusions. Having 15-20 sample size for each of the top 5, instead of 5 will increase the reliability of the data by a large amount.

One thing I wonder, from looking at what data you do have, is if teams in the middle are "reaching" with their first picks. IOW, they've already lost out on top 5 type of talent, and know they have another pick in the middle of second round. I wonder if they are taking more chances on high risk prospects that don't pan out quite as often, but when they do perform a little better. If the pick is a dud, they probably get a player not that much worse in the second round. It's something I thought of while looking at the data, but probably just my imagination.

Czech Your Math is offline   Reply With Quote
Old
08-16-2012, 06:52 PM
  #3
Sureves
Registered User
 
Sureves's Avatar
 
Join Date: Sep 2008
Location: Ottawa
Posts: 7,131
vCash: 500
Quote:
Originally Posted by Czech Your Math View Post
Good work. I think your sample size needs to be much larger to draw many conclusions. Having 15-20 sample size for each of the top 5, instead of 5 will increase the reliability of the data by a large amount.

One thing I wonder, from looking at what data you do have, is if teams in the middle are "reaching" with their first picks. IOW, they've already lost out on top 5 type of talent, and know they have another pick in the middle of second round. I wonder if they are taking more chances on high risk prospects that don't pan out quite as often, but when they do perform a little better. If the pick is a dud, they probably get a player not that much worse in the second round. It's something I thought of while looking at the data, but probably just my imagination.
Certainly a plausible theory, and yeah ideally I could increase the sample size, the problem is I feel that drafting has changed so much in the last decade or so, so I'm not sure it's as relevant.

Sureves is offline   Reply With Quote
Old
08-17-2012, 11:16 PM
  #4
wgknestrick
Registered User
 
Join Date: Aug 2012
Posts: 2,146
vCash: 500
http://myslu.stlawu.edu/~msch/sports..._NHL_Draft.pdf

http://www.etwsports.com/p/nhl-draft-value-chart.html

wgknestrick is offline   Reply With Quote
Old
08-18-2012, 07:48 AM
  #5
Fourier
Registered User
 
Fourier's Avatar
 
Join Date: Dec 2006
Location: Waterloo Ontario
Country: Canada
Posts: 8,785
vCash: 500
Interesting stuff!!!

Quote:
Originally Posted by Czech Your Math View Post
Good work. I think your sample size needs to be much larger to draw many conclusions. Having 15-20 sample size for each of the top 5, instead of 5 will increase the reliability of the data by a large amount.

One thing I wonder, from looking at what data you do have, is if teams in the middle are "reaching" with their first picks. IOW, they've already lost out on top 5 type of talent, and know they have another pick in the middle of second round. I wonder if they are taking more chances on high risk prospects that don't pan out quite as often, but when they do perform a little better. If the pick is a dud, they probably get a player not that much worse in the second round. It's something I thought of while looking at the data, but probably just my imagination.
I agree it would be better to have a larger sample size, but since there is only one draft per year we get what we get I guess. But as you say it would be interesting to continue to track these numbers. I would resist going back in time very much since it seems clear that the effort teams are putting into scouting these days is much greater than it had been in the past.

I am not as convinced that teams do much reaching in the middle rounds. I think it really does come down to something as simple as the fact that the step from junior to the NHL is so large that the vast majority of player simply fade. And with only a handful of elite players expected to come out of any given year it is far more likely that this small group would be populated by those kids who significantly separated themselves from the pack. Once you get outside of the top 5 the gap between players is much smaller making projections for the future much more difficult.

I also thought this study was interesting:

http://www.etwsports.com/p/nhl-draft-value-chart.html

(Thanks to wgknestrick for posting it).

It seems to confirm, again on a small sample though, something that I have suspected for quite sometime and that is that once you get to the second round it is more the number of picks a team has rather than where they are that will determine your team's success with the draft. Performance does seem to be fairly random.

Now again, I think that with the emphasis on scouting it is possible that we see the cutoff where randomization begins to be the norm move a little further out in the draft. It would not surpirse me if today the first 10 picks in the 2nd round are not a little more distinquished than they were in the past. But probably not by much.

I know that games played is a pretty good measure of a drafted players sucess. But I do like that these two studies break down the impact a player has beyond just games dressed. I also wonder if in measuring games played alone if some of the numbers are skewed by bias for draft position in terms of the opportunities a player gets. Do teams really invest more time in a late 1st than they might in a 3rd?


Last edited by Fourier: 08-18-2012 at 08:08 AM.
Fourier is offline   Reply With Quote
Reply

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 08:54 PM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2014 All Rights Reserved.