Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How much further to go in Man-Machine?

Author: Jorge Pichard

Date: 18:10:27 07/15/00

Go up one level in this thread


On July 15, 2000 at 20:47:41, Steve wrote:

>On July 15, 2000 at 20:27:42, Pete R. wrote:
>
>>On July 15, 2000 at 18:56:41, pete wrote:
>>
>>>To add another opinion :-)
>>>
>>>Current results IMO prove that under certain circumstances "computer entities"
>>>like the Junior one in Dortmund can reach GM-like results.
>>
>>I don't see any need to tiptoe are the "are they GM strength?" question, it
>>seems pretty clear to me.  If you define GM strength as entering tournaments and
>>achieving a GM norm, then I think computers can do this, in fact we just saw it
>>done twice.  If you mean "can they assess a position as well as a human GM" then
>>the answer is generally no, because they are not good in certain types of
>>position due to lack of positional understanding.  We can speculate that
>>opponents might generally adopt such anti-comp strategies and so a machine
>>wouldn't be able to achieve a GM norm against such a field, but this is neither
>>here nor there.  Some players play anti-comp, some don't, and some refuse to
>>play at all.
>>
>>The important thing is how well they play generally.  No offense to the DJ team
>>in any way, but this tournament has given me zero incentive to buy Junior in
>>addition to my other software, because I see that it plays similarly aimless
>>moves.  I already have this in Fritz and Hiarcs. ;) If it played good positional
>>chess and never made a stupid or aimless move, even without winning a single
>>game, I would pull out my credit card in an instant.  How it plays is vastly
>>more important than whether it wins.

Just because DJ Lost to Piket who has been practicing ever since he had a draw
against Deep Junior last time they met in G/60, doesn't mean that Deep Junior
is not one of the best positional program in the market against Human. I
honestly don't think that none of us can match the anti computer knowledges that
Mr Piket has gained ever since he decided to practice against Deep Junior.

Pichard.
>
>I feel the same way, but what you're asking computers to do is operate with
>human-like intelligence, and the direction that computer chess has taken
>indicates that the people who really understand this stuff (programmers) don't
>feel this goal is attainable.  Frankly, I'm happy for the GMs (not to mention
>IMs and NMs), because those who are willing to teach ordinary players about
>chess -- through lessons, books, game annotations, etc. -- should be able to
>make a decent living from the game.  Computers may beat them in tournaments (or
>not), but they can't match their teaching skills.



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.