Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Level of complexity

Author: Tord Romstad

Date: 02:36:20 01/08/04

Go up one level in this thread


On January 07, 2004 at 20:40:49, Rick Rice wrote:

>How would you rate the level of complexity for the following games, in the
>programming sense?
>
>Shogi
>Go
>Chinese Chess
>Orthodox Chess
>Gothic Chess
>Fisher Random Chess
>Hexagonal Chess
>Xiangki
>
>Feel free to add any of your own.

Being one of the less than 1% readers who knows the rules of all games, I'll
try to answer.  :-)

First of all, of course, it depends on your definition of "level of
complexity".  If your goal is to make the strongest engine in the world
for a specific game, orthodox chess is almost certainly by far the
hardest, while Gothic chess and hexagonal chess are the easiest.  This
has nothing to do with the games in themselves, the reason is simply
that much less effort have been made on making strong engines in
all the other games than in orthodox chess.

A more interesting question is how the different games compare regarding
the difficulty of creating a challenging opponent for human players.  In
this sense Go is incomparably much more difficult than all the other
games, with shogi as a distant number two.  The remaining games should
be roughly comparable.  Gothic chess and hexagonal chess are perhaps
somewhat more complicated than orthodox chess, FRC and xiangqi (note
that the correct spelling is "xiangqi", not "xiangki"), but the
differences between those games are not big.

Unlike what most people think, the big branching factor in go and shogi
is not the most important reason why it is more difficult to write strong
computer programs for these games.  The real problem is that it is very
difficult to write good evaluation functions.  In orthodox chess, FRC,
xiangqi, Gothic chess and hexagonal chess, a material evaluation with some
simple positional stuff added is enough to create a reasonably strong
engine (at least strong enough to beat the average human player).  In Go,
there is no equivalent of material, and also no other similarly dominant
component of the evaluation function.  Shogi occupies a middle position:
It is possible to count and evaluate the material balance, but material
has much less importance than in chess.  For instance, the best human
players easily beat the strongest computers with a rook handicap.

Tord



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.