| |
BGonline.org Forums
Recube - skewed score
Posted By: Matt Cohn-Geier In Response To: Recube - skewed score (Frank_N_Stein)
Date: Thursday, 23 April 2009, at 9:08 p.m.
I do not understand match equity, doubling windows, how to do the math, or how to apply the information. I have Kits book on the subject, but it could be written in another language for all I understand.
I take a simple approach to these positions. I know if I win the score will be 6-0 crawford to 7. That is not easy to overcome by anyone.
It is largely a question of "how bad?" Blue's position is overwhelming by money standards but the score matters.
Suppose Blue's chances from this position are 90%. If it's Crawford 5-away instead of 7-away the deficit is still not easy to overcome but is now a huge pass. If it's instead Crawford 8-away it's a trivial take.
METs basically tell you your probability of winning the match from any given score. How the first METs were constructed is largely irrelevant, although it is somewhat interesting how accurate people were able to be in the prebot days. g11, the current standard MET, was constructed using GNU 0-ply rollouts.
Big cubes like this are the easiest since they are for the match. Since it's for the match, you only have to consider your odds of winning the match if you pass (which is whatever the entry is in the MET), vs. your odds of winning the match if you take (which is whatever your odds are of winning the game). Since the MET gives 91% or so for Crawford, 7-away, you should take if your chances are better than 9% and pass if they are worse than 9%.
From there you can derive the general formula TP = risk/(risk+gain) and some other stuff like doubling windows, gammon values, but most of that doesn't matter IMHO. If you know the takepoints you can generally do well enough.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.