Re: [Bug-gnubg] change in calculation of cube decision ratings

[Joern Thyssen]

On Sun, Aug 04, 2002 at 11:54:06PM +0000, Øystein Johansen wrote
> On Sun, Aug 04, 2002 at 05:53:45PM +0000, Joern Thyssen wrote:
> > Hi,
> > 
> > It's also possible to go one step further and divide all chequer plays
> > and cube decisions into several categories (forced, very easy, easy,
> > medium, etc. etc.).
> 
> I like the idea, but the problem is that what the computer think is easy,
> can be very difficult for us human beeings. (Else we would not have ever
> blundered!)
> 
> I'm sending you this position:
> 
>     +24-23-22-21-20-19------18-17-16-15-14-13-+  O: gnubg
>     | X     O  O  O  O |   | O        O     X |  0 points
>     |       O  O  O  O |   | O              X |  
>     |                O |   |                  |  
>     |                  |   |                  |  
>     |                  |   |                  |  
>     |                  |BAR|                  |v (Cube: 1)
>     |                  |   |                  |  
>     |                  |   |    X             |  
>     |                X |   |    X             |  
>     | O        X  X  X |   |    X             |  Rolled 16
>     | O  O  X  X  X  X |   |    X             |  0 points
>     +-1--2--3--4--5--6-------7--8--9-10-11-12-+  X: root
> 
> Don't put it into your gnubg, but make up your opinion whenever this is a
> very easy, easy, tough, difficult or very difficult decision. (Or is it
> forced?) At least there is three strong candidates 24/17, 13/7 8/7 and
> 8/2* 3/2. 

Well, I'm not sure what I would call this position. I would instinctly
move 13/7 8/7, so I find it an easy decision. Of course 8/2 3/2 and
24/17 would be considered, but quickly discarded, as they
counter-strategic.

Anyway, you're right that just looking on equity differences  will not
always yield the correct difficulty of a move.

> (What would you play...? I must admit I got this right, but I really had
> think for 3-5 minutes before I was sure I had the right move.)
> 
> The problem will be getting gnubg to understand that this really is a
> difficult decision for a human beeing. How do we solve that? Can we?

An alternative strategy would be to call a move difficult if 0-ply
differed from 2-ply, or 2-ply from 3-ply etc etc, combined with an
equity difference.

Jørn

-- 
Joern Thyssen, PhD
Vendsysselgade 3, 3., DK-9000 Aalborg, Denmark
+45 9813 2791 (private) / +45 2077 2689 (mobile) / +45 9633 7036 (work)