|Subject:||[Bug-gnubg] Replicating XG Roller setting in gnubg a question|
|Date:||Thu, 3 Sep 2009 19:42:25 +0000|
I set up gnubg to mimic XG Roller settings|
first 2 moves 1-ply cube 0-ply checker
No Variance Reduction
XG Roller uses 3600 trials.
Instead of using number of trials I decided to roll them out until JSD reached 2.58 minimum trials 144.
If you do the same with VR ON, it takes MUCH less time by a huge factor to reach the same JSD.
I'm not a statistician but isn't JSD important if you only need to rank plays rather than need absolute equity?
Why not do it with VR ON ?
I'm sure I must have overlooked something obvious...
View your other email accounts from your Hotmail inbox. Add them now.
|[Prev in Thread]||Current Thread||[Next in Thread]|