View Single Post
  #13   Spotlight this post!  
Unread 18-04-2011, 23:52
Ed Law's Avatar
Ed Law Ed Law is offline
Registered User
no team (formerly with 2834)
 
Join Date: Apr 2008
Rookie Year: 2009
Location: Foster City, CA, USA
Posts: 752
Ed Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond reputeEd Law has a reputation beyond repute
Re: paper: New Scouting Database from Team 2834

Quote:
Originally Posted by Travis Hoffman View Post
Thanks, Ed!

Can you please explain in more detail the calculation on the WM Results tab, column G, "Wt. Avg. CCWM", when calculated in "default" mode? How is that number obtained? How is the weighting done?
The weighted average is calculated this way. This year there are 6 weeks of competition. By default, the weighted average is projected to the week after which is 7.
The OPR from each event is scaled by multiplying the reciprocal of the number of weeks between the event and week 7. For example if the event was in week 3, then the OPR (x) will multiply by 1/4. Let's say this team also attended a Week 5 event, that OPR (y) will multiply by 1/2. The weighted average will be (x/4+y/2)/(1/4+1/2). In this case, the week 5 event will have double the weight than the week 3 event since x is divided by 4 and y is only divided by 2.
I hope this makes sense. There is nothing special about this way of assigning weight. There can be many other ways to do it. The idea is the events closer to week 7 should have a higher weight than events that are earlier. I did it this way out of simplicity and convenience. You can use a quadratic function or exponential function if you wish.
__________________
Please don't call me Mr. Ed, I am not a talking horse.
Reply With Quote