[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

Suggested Idea for Performance Grading

Posted By: Rick Janowski
Date: Friday, 3 October 2014, at 5:13 p.m.

In Response To: Suggested Idea for Performance Grading (Ian Dunstan)

The discussion on match win rate criteria acknowledged that there is an obvious alternative to measuring performance by error rate, which is the traditional one which is fundamentally based on win rate. There are two problems with this. Firstly, the average strength of the players participating will vary from tournament to tournament, but this can be normalised to consider a standard opponent strength. Secondly, unfortunately because win-rate variance is considerably higher than error-rate variance for any given sample size, a much larger number of matches need to be played to obtain the same level of confidence.

I think removing outliers is problematic for 2 reasons. Firstly, the statistical distribution for pr tends to be significantly skewed. For example, a pr 4 player may play one game at pr 12 but cannot play a symmetrical error of minus 4 pr. Removing outliers will tend to falsely lower pr level. Secondly, worst result outlier in most cases isn't an outlier at all just a very bad game.

I am not sure on how logical or practical these suggestions are. The standard measure of pr seems adequate without employing some seeming arbitrary adjustments.

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.