Pengwin7's 2010-2011 Defensemen Rankings (Draft Version)

[Pengwin7]

Mathematical Model based on H2H league with 6 categories: G, A, +/-, PIM, PPP, SOG
Scores are based on actual 2010-2011 NHL Statistics.

Defensemen are given a score (10/10 = typical maximum**) in each category.
Scoring:
*G. 50G = 10pts. 25G = 5pts. 0G = 0pts. [AVG Top 120 Fs = 24G, Top 60D = 8G]
*A. 70A = 10pts. 35A = 5pts. 0A = 0pts. [AVG Top 120 Fs = 33A, Top 60D = 29A]
*+/-. +30 = 10pts. EVEN = 5pts. -30 = 0pts. [AVG Top 120 Fs & Top 60D = +6]
*PIM. 120PIM = 10pts. 60PIM = 5pts. 0PIM = 0pts. [AVG Top 120 Fs & Top 60D = 55 PIM]
*PPP. 40PPP = 10pts. 20PPP = 5pts. 0PPP = 0pts. [AVG Top 120 Fs = 17 PPP, Top 60D = 15PPP]
*SOG. 400SOG = 10pts. 200SOG = 5pts. 0 = 0pts. [AVG Top 120 Fs = 200 SOG, Top 60D = 150 SOG]

Notes:
*Scoring is the same as for forwards. A defensemen's score can be compared directly to forward's scores (see Forwards threads).
*In most categories, max score = 10, min score = 0.
*In PIM, the max score = 16.5 (for Theo Peckham's 198 PIM)**.

Discussion:
*This is a work in progress - open to discussion.
*It is a little slow at work these days.


Hope you enjoy.
Comments welcome!

[Pengwin7]

re: PIM.
I expect that some people will find the PIM scores to be inflated.

Is it really accurate for Theo Peckham to get 16.5pts for his 198PIM?
It is debatable, but I think that if a player gets 3.3x the average PIM output (60PIM), then he should get a score of 3.3x the average score. (3.3x5 = 16.5).

FWIW, I believe this is real value in this H2H format!

[Shoeless]

Interesting and intriguing - without putting too much thought into what you have done here I think your rankings are more applicable to roto scoring than they are to H2H. If I was going to use your numbers for H2H I think I might throw out the high and the low category for each player and use the remaining 4 to derive a rating - I think I would have more confidence in the resulting rankings.

[Hey Robbie]

re: PIM.
I expect that some people will find the PIM scores to be inflated.

Is it really accurate for Theo Peckham to get 16.5pts for his 198PIM?
It is debatable, but I think that if a player gets 3.3x the average PIM output (60PIM), then he should get a score of 3.3x the average score. (3.3x5 = 16.5).

FWIW, I believe this is real value in this H2H format!

I think you are right on that players that are way, way above average in a stat should be able to get "ranking points" or whatever you want to call them outside of the normal 1-10. In H2H, if a player can almost single-handedly win you a cat many weeks, his value is really that high, regardless if it's goals, PIMs, or air kisses.

[tgraveline]

I think this list is projected better than the forwards one. I am curious to see if Buff can live up to it though. He's good, but I don't know if he can sustain that kind of dominance. Eventually teams will know what to do and it still depends on who's around him. Having Enstrom is mighty helpful.

[Go Puck Yourself]

I like it and it is a reasonable way to score players. Thanks for sharing!

[Pengwin7]

Interesting and intriguing - without putting too much thought into what you have done here I think your rankings are more applicable to roto scoring than they are to H2H. If I was going to use your numbers for H2H I think I might throw out the high and the low category for each player and use the remaining 4 to derive a rating - I think I would have more confidence in the resulting rankings.

There are lots of things one could do with these numbers.

The +/- statistic is very volatile (like a goalie's SO stat).
Ex#1. Kovalchuk is ranked #100.
This is because he finished -26. -26 gives him a +/- score = 0.7.
However, say he finished at +10. +10 gives him a +/- score = 6.7. (Eriksson)

The extra +6pts would add 1.0 to his "Base10" score. (+6pts/6cats)
Instead of having a score of 4.12 (100th best forward)
he would've had a score 0f 5.12 (42nd best forward).

The PIM could also be re-ratioed and based on 240 PIM = 10pts.
Everybody's PIM value would be cut in half.
Ex. #2. Konopka is ranked #39.
If we halved PIM values, Konopka would go from a PIM10 score of 25.6 to a PIM10 score of 12.8.
This would give him an overall "Base10" score of only 3.05.
He would fall out of the top 150 forwards... likely.


There is a lot of mathematical interpretation required to extract value from this model. Use whichever categories you like and feel free to ratio players for a full healthy season!

[Koolaid Man]

I'm always trying to put together statistic as a way to rank my players and players that I want to draft. There are always more equations that can be added to make the results as accurate as possible.

If I could make a suggest. This suggest will undoubtedly add a lot of time during the compilation period.

You said that your values were based on 2010-2011 statistic. Those stats that you are using to compare players is a little biased because some players have played more games then others (injuries). If you could somehow incorporate a per game value and then compare that number to a new base number as you have done.

All of a sudden you have to start thinking about a minimum number of games played, and some way to subtract value from players final ranking value (or add) based on their durability.

Just an idea. Great work though!!