| |
BGonline.org Forums
Cube error rates
Posted By: playBunny In Response To: Cube error rates (Joe Russell)
Date: Wednesday, 22 July 2009, at 5:12 p.m.
Q. In an absurd extension of this, suppose each side danced a 100 times. Do you think that you would have cost yourself 90% MWC
A. Why not?
An error measures equity loss in a given position. Between two positions there's a change in equity. Here, if you fail to cube and dance and your opponent dances then there's equity lost by you not having cubed, equity lost by not having come in and equity gained by your opponent not cubing and not coming in. Those changes in equity should sum to zero because you are now in the starting position. If 90% MWC gets lost through the repeated cube error then 90% must have got pumped back in.
If you make the same mistake 100 or 1000 times the total error (and the total restoration) will be huge but the average error will only be the size of one of them. It's the average error that is used for comparing players. If you only take the largest of those errors then you must also omit all those cube decisions from the move count else they will incorrectly reduce the error when computing the average.
That 90% MWC is 0.9% MWC per decision from 100 turns. Would you prefer to take the largest, 0.9%, and declare that the other 99 turns don't count?
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.