Re: computer-go: Tsume-Go database.

[Thomas Wolf <T.Wolf@xxxxxxxxx>]

Jan Ramon wrote:

> Therefore, there is a need for publicly available high quality benchmarks
> in the tsume-go domain. 
> 1) A well-known dataset is the database generated by
> gotools (by T. Wolf).  However, this dataset also has disadvantages:
> - Gotools has produced it.  It can be expected that gotools
> performs better on problems it has generated itself than on others.  

This is only true insofar as the problems are enclosed. Otherwise
their generation involves as much as possible randomness. Their
shape,... has nothing to do whether GoTools is suited for them or not.

> This
> makes comparison with gotools difficult using gotools problems.
> - The problems are somehow artificial.  

True.

> They contain only bounded
> positions.  

Yes.

>They are of basic level (i.e. easy to solve by dan players).

Yes, at least most. Still there are many also hard ones for dan players.
I picked a few for the 64-problem set in

"Forward pruning and other heuristic search techniques
 in tsume go", Information Sciences {\bf 122}
          (No.\ 1) (2000) pp.\ 55--76.

and as the test showed, 4th and 5th Dan players had a hard time.

Thomas