Thread: bitboards
View Single Post
  #6  
Old November 1st 03, 01:38 PM
Frank Andreas de Groot
external usenet poster
 
Posts: n/a
Default Xiangqi/Shogi engine strength Was:( bitboards)

"Ed Seid" wrote

On a related note, what is the current status of Shogi engines? Shogi
presents some interesting computing problems.


Nothing comes close in complexity of any 2-person, zero-sum, total-information
game, as the game of "Go" (Weiqi, Igo, Baduk). For a period of 12 years, there
was a 1.5 million USD prize for the first program that could beat a little girl
(Japan has plenty of pre-teens who have low-amateur status in Go).

Experts in neural networks, top chess programmers etc. have worked decades on Go
programs, with rather pathetic results. I made a gallery of about 70 of those
guys:
http://www.zenhacker.com/pioneers.htm

Many are very strong Go players, engineers, CPU designers etc. None has achieved
any success that can remotely be compared to what has been achieved in Chess. Go
is very, very much more complex than chess, also for humans. It takes about a
year before most people *understand* the beyond-basics of the game and the
rating scale is "deeper". The branching factor is staggering, with hundreds of
legal moves per ply, many moves having effects 100 moves deeper in the search
tree, or more.

Still, Go has much simpler rules than chess..



Ads
 

Guitar Lessons - Myspace Comments - Car Loan - Krakow hotels - vBulletin