Go Back   HFBoards > General Hockey Discussion > By The Numbers
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie

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.

Weighted Standard Deviations

Thread Tools
07-25-2013, 05:20 PM
Registered User
AD1066's Avatar
Join Date: Sep 2011
Posts: 5,453
vCash: 500
Weighted Standard Deviations

Hi everyone,
I'm working on a little project out of curiosity and would greatly appreciate some input. I've put together a spreadsheet of stats from the 2012-2013 season that contains the following data:

Player, Age, Team, Position, Games Played, Goals, Assists, Points, Goals Created, Plus/Minus, PIM, Shots, S%, Hits, Blocked Shots, Missed Shots, Giveaways, Takeaways, TOI, ATOI, EV, PP, SH, Game Winners, FOW, FOL, and FO% - a pretty exhaustive list.

What I'm trying to do is to create an aggregate measure of a player's performance based on a weighted average of their standard deviation in every category, and see how well that list approximates their perceived performance last season. Ex:

Player A's "performance" = (A's STDV of games played * weight1) + (A's STDV of goals * weight2) + (A's STDV of assists * weight3) ... You get the idea.

What I'm looking for help with is assigning weights to the different categories. I'm sorry if this is vague, but I'm not really sure how to approach this.

Should goals weigh more than assists?
Do we give the same value to ES and PP points?
Should PIM or missed shots carry a negative weight?
Should game winners be taken into consideration? Should faceoffs?
Etc, etc.

Here is a link to the forwards spreadsheet. Thanks in advance for any and all suggestions.

AD1066 is offline   Reply With Quote
07-25-2013, 06:42 PM
Registered User
torero's Avatar
Join Date: Oct 2007
Location: Border of lake Leman
Country: Switzerland
Posts: 4,059
vCash: 500
why not ? it is interesting to know how consistent some players are relative to a field.

dividing the standard deviation by the average ... you get a measure of dispersion. which is always nice ... you can go ahead by integrating shorter periods to see where the player stand in his evolution / regression.

could be cool.
it is a top down approach ... you can choose a player according to criterias ...
like this you build the team of the future !!

too often the bottom up approach is used ... probably a mistake.

keep up with you great ideas.

torero is offline   Reply With Quote

Forum Jump


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 08:40 AM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"

vBulletin Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. @2017 All Rights Reserved.