Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The future of computer chess...

Author: Georg Langrath

Date: 23:33:44 08/04/98

Go up one level in this thread


On August 04, 1998 at 23:19:08, Ed Schröder wrote:

>Hi Fernando,
>
>I think you are right. I already pointed this out in other postings. The
>emphasis for future versions is to extend the chess engine with more
>useful functions. If you have a look at the Rebel10 Diary you will see
>that this is already the case and more is to come over there.
>
>Changing the subject. It's now about 2 weeks after the match. In the
>meantime I have thought a lot about what happened on Ischia. Maybe too
>early for my final thoughts and personal conclusions. I wonder where do we
>stand in computer chess?
>
>It looks like playing against the world-top (even against 2800 players!) on
>blitz and semi-blitz makes no sense anymore. Rebel played 4 blitz games
>5:00 plus 5 Fischer seconds increment for each move which makes the
>4 games actually 10:00 games. Then 2 games on 15:00. Result 4.5-1.5
>and believe me I never ever expected this to happen. Apparently you have
>to play first to find out.


I think that other features will be more impoertant in future! Above all
handicap levels. Average player has about 1200-1300 in strength! A good handicap
level give you an even match without feeling that computer make bad moves on
purpose!

Georg>
>The same counts for the 2 tournament games. Before the 2 games my
>main fear was that Rebel would have been slaughtered, first positional
>out-played, then slowly strangled, all in the well known Seirawan style at
>Aegon but then even more effective :)
>
>The opposite happened. Rebel in both tournament games took the initiative
>(a complete new experience for me!) had good winning chances in the
>first game and some even say also in the second game, which I tend
>to disagree on.
>
>What on earth is happening?
>
>I know, I know, just 2 games!
>
>In this respect I am really looking forward to the 6 Ferret games! It will tell
>us more about the current state of computer chess. Please Fernando make sure
>this new match will take place!
>
>One of my worries and also coming to my point, say Ferret will do well,
>say a  victory of 4-2. Next a 6 game match on 40/2:00 versus comp_X
>and Kasparov or Anand ending in 3-3 or 4-2 in favor of comp_X.
>
>All speculation of course. But if it happens my worry is that it is a killer
>for computer chess! My assumption is that people will lose interest as
>the race on playing strength is finally decided.
>
>My prediction, if it happens some of us have to look for a new job :)
>
>- Ed -
>
>
>>Hi Ed:
>>Knowing that some people is asking you to shoot on your feet, then I dare to
>>ask
>>something for Rebel 11 around 2 years in advance. Besides, after seeing the
>>post
>>written by the presumed  mummy of Sean Evans, everything is possible.
>>Well, the core of my wish list is not strength but coach functions. I am
>>already
>>beaten 95% of the times by rebel 9 and the rest of Top programs and looking at
>>an old poll, it seems I am not the only one here. So, strength is not anymore
>>the issue. I don’t say you should weaken your engine or not to devote some
>>time
>>to improvement, OK, but I would like to see a greater effort in the coach
>>functions in order, some day, to improve that 95% loses to, let us say, only
>>90%.
>>What I want, specifically?
>>a) a coach function like that of Fritz, BUT not delivered in the setting of a
>>weakened level. Coach should be present in any level, at will.
>>b) after the end of the game, if on, the coach should appear and show which
>>moves were the cause of defeat, explain why and show which were the good
>
>>moves,
>>step by step
>>c) after that, the coach should push the student to examine a lot of exercises
>>similar to the position pattern where the mistake was committed. Of course
>>this
>>assumes Rebel 11 has a huge database with a thoroughly detailed number of key
>>positions
>>d) rebel 11 should keep in store the mistakes committed and the results of
>>exercises so if in another game if a similar positions arises, the coach
>>should
>>warn that the player must be careful as much in a similar position he ruined a
>>game, remember your exercisses, etc.
>>In other words, Ed, I see a coach function operating together with normal
>>game,
>>not as something different. Besides, it is more rewarding and motivating to do
>>some study of positions when you still are in the mood of playing chess. There
>>are very good programs for learning chess, but they are somewhat boring and
>>ask
>>from the student an special moment to go through. That is somewhat demanding
>>for
>>an old fart like me, above 40 years old and in fact in the verge of 50.
>>Give it a thought. Learning functions are the future. I don’t see other reason
>>to purchase now a new program. Sure many guys feel like me.
>>Regards
>>fernando



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.