Author: Vincent Diepeveen
Date: 10:57:45 12/11/02
Go up one level in this thread
On December 11, 2002 at 13:10:25, Edward Seid wrote: >On December 11, 2002 at 12:54:47, Richard Pijl wrote: > >>This is hard to answer as the question is not clear. >>Do you mean: >>1) the amount of programming required to produce a playing program? >>2) the amount of programming required to produce the best playing program? >>3) the amount of programming required to produce a (human) world-champion level >>playing program >>4) The amount of research done in programming a playing program (which reduces >>complexity for the programmer as ideas can be copied) > >You're correct... the original question was poorly phrased. I think what I had >in mind was the computational complexity to 'solve' each game. Which is loosely >related to #3 above. There is a difference in beating a world champion and by solving a game. I hope you realize that in GO when 'brute force' searches will be capable of searching around 15 ply (excluding of course extensions for ladders and such) that a world champion GO must come from a real good family to beat the thing, because not a single tactical line will ever win for him and also all possibilities to cut groups and distract the program will get very hard. For GO the basic problem right now is something we figured out back in 1980 for chess already. Also the hashtable knowledge level in the go world is from around 1989 as it was in chess. In fact most even claim that a 32 bits hashing in GO is perfectly working for them. The beginners :) >As I look at your list, I see that I'm not interested in #1. >Vincent made a compelling argument that #2 is driven by commercial interests in >each game. >Besides the fact that western chess is computing's 'Drosophila melanogaster', >I'm not sure about #4.
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.