| |
BGonline.org Forums
Estimating winning percentage from MWC errors
Posted By: sandokan In Response To: Estimating winning percentage from MWC errors (Daniel Murphy)
Date: Monday, 20 December 2010, at 8:46 a.m.
It was a bad statement of mine that you know which matches you win and which matches you lose. You know only that you win 500 matches and you lose 500 matches. Imagine a graph having number of moves on the horizontal axis and MWC on the vertical axis. You have 1000 lines (they are not continuous of course) starting on the point (0,50%) and they oscillate until 500 of them reach 100% MWC and 500 reach 0% MWC. That if you play perfectly.
But you will not play perfectly and you will make an error of 10% in each match. You do your best (with that error exception) and try to win in all matches but in 500 of them (you don't know which) you will lose even playing perfectly. Because of your error of 10% MWC you will lose 50 matches more. So you are wrong when you say that you can make a 10% error in your winning matches knowing that you'll win them anyway.
Again it is a confusion between MWC and winning percentage which is tracked from the start of the match. In your position you lose indeed 10% MWC in this match. The winning percentage was 50% if you played perfectly until now and become 45% after your error because you lose 10% of your winning matches (your previous winning percentage) more.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.