Re: [Bug-gnubg] Bug in analysis - position enclosed - more details

[Joern Thyssen]

On Mon, Oct 28, 2002 at 11:26:07PM -0800, Ned Cross wrote
> > > I'm wondering if there are two separate issues here:
> > > 
> > > 1) That GNU doesn't understand this position.
> > > 
> > > 2) That the reported 2-ply win %  for the candidate move ("after the 
> > > roll")
> > > does NOT equal the reported 2-ply win %  for the position resulting from 
> > > the
> > > candidate move ("before the next roll"). Right or wrong, shouldn't those 
> > > two
> > > numbers be the same?
> > > 
> > 
> > Should be. Perhaps one is cubelss and the other cubefull? Are you sure the
> > parameters are the same?
> 
> (NC) The output in my original post was generated from a GNU 2-ply
> analysis (cubeful checker play) with 2-ply cube. The analysis
> parameters were the same.  After digging deeper, I have found the
> cause of the discrepancy.  The value for 2-ply analysis AFTER the move
> is the same as the value for 1-ply analysis BEFORE the next move.
> (Similarly, a 3-ply move analysis of the 6/1* 5/1 move predicts the
> same 82.2% wins as the 2-ply cube analysis before the next roll).
> This does not seem correct to me.  Shouldn't 2-ply = 2-ply if there is
> no move in between?

There is actually a move in between: O fans!!! If check the cube
decision for O as he fans you see that he has exactly 14.2% chance of
winning.

The 2-ply chequer play decision will work as follows:

X moves 6/1 5/1. Evaluation the resulting position at 2-ply:

equity = 0
loop over all rolls for O
   find best move for O
   equity += evaluate resulting position at 1-ply   [a]

final equity = equity/36

However, since there are no legal moves for O all positions evaluated
at [a] are the same, thus "equity" is the same, hence the resulting "final
equity" is identical to the 1-ply equity (which is what you've
observed).

I hope this makes sence!

Jørn

-- 
Joern Thyssen, PhD
Vendsysselgade 3, 3., DK-9000 Aalborg, Denmark
+45 9813 2791 (private) / +45 2818 0183 (mobile) / +45 9633 7036 (work)
Note: new mobile number!