Stockfish tablebases?

General discussion about computer chess...
Post Reply
Paddington
Posts: 2
Joined: Thu Jun 10, 2010 1:16 pm

Stockfish tablebases?

Post by Paddington » Fri Jul 09, 2010 2:11 pm

Just wondering...
Why does Stockfish not support tablebases? Hard to implement?

Edmund
Posts: 20
Joined: Thu Jun 10, 2010 2:07 pm

Re: Stockfish tablebases?

Post by Edmund » Fri Jul 09, 2010 3:24 pm

Nalimov doesn't fit with their open source licence, as one needs the permission of Nalimov (Probing Code) and Kadatch (Compressing Code) to be allowed using the code.

IIRC the team is however considering implementing the Gaviota Tablebases.

Tord
Posts: 35
Joined: Mon Jun 14, 2010 9:08 pm
Real Name: Tord Romstad

Re: Stockfish tablebases?

Post by Tord » Mon Jul 12, 2010 12:25 pm

Paddington wrote:Just wondering...
Why does Stockfish not support tablebases? Hard to implement?
That depends on your definition of "implement". Implementing the ability to probe tablebases is easy. Finding a way to use them to actually improve strength is hard. We strongly dislike adding complexity to the program unless it increases the strength significantly. It's possible that support for Gaviota tablebases will be added somewhere down the road, but we're not yet sure.

Michel Van den Bergh
Posts: 24
Joined: Thu Jun 10, 2010 4:30 pm

Re: Stockfish tablebases?

Post by Michel Van den Bergh » Tue Jul 13, 2010 10:45 am

We strongly dislike adding complexity to the program unless it increases the strength significantly.
That's reasonable. But one of the major uses of engines is analysis. In that case tablebases are very useful.

The added complexity seems quite minor.

Post Reply