Author: Dann Corbit
Date: 21:39:25 01/29/99
Go up one level in this thread
On January 29, 1999 at 23:58:21, KarinsDad wrote: >On January 29, 1999 at 23:51:30, Dann Corbit wrote: > >>On January 29, 1999 at 23:36:40, KarinsDad wrote: >>[snip] >>>Are you analyzing only leaf nodes, or at you going through positions from the >>>start of games? >>> >>>Please explain your numbers. 10^47 appears to be about 16 to 18 moves down for >>>each side down considering all legal moves. Since about >>>9.999999999999999999999999999999999999999999 x 10^46 of those moves are >>>inferior, it seems likely that you are more or less covering the bases as it is. >>>What am I missing? If the opponent makes an inferior move, then let a program's >>>tactical engine figure it out. If the opponent makes a theoretical novelty, let >>>your learning routines record it and analyze it later. >>KK made an interesting observation, and clearly he is right. You don't have to >>compute every game, only every unique possible board position. I might be off >>by a 6 orders of magnitude or so, but that would not matter much. >>[snip] > >You still did not explain your numbers. Not all of us have been following the >C.A.P. results. This is interesting. Please provide more detail. One of the famous HAKMEM problems is problem 95, solve chess. From this URL: http://www.lysator.liu.se/~mc/hackerdom/hakmem/proposed.html PROBLEM 95: Solve chess. There are about 10^40 possible positions; in most of them, one side is hopelessly lost. If you can enumerate the positions, you can solve the game. While, on the one hand, there are 20 possible 1st moves, and 20 possible 2nd moves, leading to 400 possible moves for the first choice by white or black, players don't move at random. If you get a big database of millions of games you will find that certain moves are favored from every position. The reason is that in playing millions of games, people have figured out what tends to win and what tends to lose. Lots of bonehead moves are eliminated off the bat. I won't shove my queen in front of your pawn unprotected. But the average moves are harder to locate and the very good moves can be very hard to find. Project Heartwood is taking the most frequently played chess positions and analyzing them at 12 minutes of PII 300Mhz time per position. We are starting with moves that are played at least 200 times (about 60,000 in our database). Then we will analyze positions played at least 100 times (about 100,000) as we drop down to 50 the number rapidly escalates. But by analyzing the most frequent moves first, we will have a database containing moves with a high probability of occurrence. I have sent a message to distributed.net, and if they get involved we might well be able to analyze tens of millions of positions. With an opening book such as that, you are much less likely to fall into a canyon. There are lots of other things we can do with the data. Why not read our FAQ? It will only take you five or ten minutes and it will give you more sex appeal, whiter teeth, fresher breath, and the confidence to make a real fool out of yourself if you care to. HEY! HAVE YOU HEAR WHAT IS INSIDE THE C.A.P. FAQ?! ftp://38.168.214.175/pub/Chess%20Analysis%20Project%20FAQ.htm Yes, and for $50 I'll keep my mouth shut. But it may be too late. The word is getting out.
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.