View Single Post
Old 18th November 2014, 02:58 PM
Yorkshiretea's Avatar
Yorkshiretea Yorkshiretea is offline
B Grade Bandit
Join Date: Aug 2013
Member of: Anston FTC
Location: Sheffield
Posts: 2,165

Originally Posted by maestro View Post
You're absolutely right. If we'd like to know somebody's real position then the top 5% and bottom 5% of the stastics should be removed from the averaging.

Removing also the second worst score makes sense, though, because everybody can have extremely good and extremely bad days, but a gun failure can happen, too. So, rolling 20 but removed best 1 and worst 2 would work fine.
Not the problem and simply doesn't work, if you took top and lowest score from an AA shooter here's what it would do:

So let's take Our Kes's old scores, he's on 93.36% to start with, so remove one 100% = 93.01. Remove lowest 80 = 93.76%. Nice increase Andy

Now lets do that to a C Grader, me, 57.45% to start with, let's remove the 80% = 56.27 and let's remove that 18 = 58.39%. Nice increase Martin

It doesn't address the problem at all when dealing with a data set of 20 or the problem with using Mean as an average or Outliers. In my example, thanks for the increase tho I can tell you it's still not accurate and you'll only be punishing the lower grades, which is what the Outliers in Mean are already doing so you don't need to double lads Also what do you do if they have a bad start and then stay constant so they get 12, 18, 15 to start with and then hit 21 and 20 for the rest on the season? It would make .35 difference!!!

Last edited by Yorkshiretea; 18th November 2014 at 03:16 PM.
Reply With Quote