Programs designed for especially good results against humans

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

Marc MP

Re: Programs designed for especially good results against hu

Post by Marc MP »

JensBNielsen wrote:Thanks for your replies.

I assume that noone so far has made a program that replies instantly.

When I wrote the preparation material (80 pages) for GM Bent Larsen and other danish players before their meeting with Baby Deep Blue in Copenhagen 1993, I really feared the machine had such a feature that might totally confuse and stress the players.

Jens
Hi Jens,

What was the result of the match? Are the games still available somewhere? I vaguely remember reading something about it in the now long defunc Inside Chess. They were rapid games weren't they?

If I remember correctly Larsen played a four knigth games with white, right? (and an English against Deep Thought couple of years before?).
JensBNielsen

Re: Programs designed for especially good results against hu

Post by JensBNielsen »

Marc MP wrote:
JensBNielsen wrote:Thanks for your replies.

I assume that noone so far has made a program that replies instantly.

When I wrote the preparation material (80 pages) for GM Bent Larsen and other danish players before their meeting with Baby Deep Blue in Copenhagen 1993, I really feared the machine had such a feature that might totally confuse and stress the players.

Jens
Hi Jens,

What was the result of the match? Are the games still available somewhere? I vaguely remember reading something about it in the now long defunc Inside Chess. They were rapid games weren't they?

If I remember correctly Larsen played a four knigth games with white, right? (and an English against Deep Thought couple of years before?).
Perhaps I should upload the article from the danish chessmagazine Skakbladet...

I just found this link, where you can see the results: http://www.chessgames.com/perl/chessplayer?pid=29912
(You must be a member to download the games)

You are right; Larsens first game was a four knigt opening which he won.
After that he drew 3 games (there must be an error in the list. Larsen also drew with the scandinavian opening).
All Larsens games were at 40/2.

Also most games against the other danish top players were at 40/2.

But there were also some rapid games, and they were very interesting.
Deep Blue did very poor in these games although they were faster.

In Deep Blue - Bjarke Kristensen the computer made a big error in the first move after the opening book and lost in 24 moves!

In Jens Kristiansen - Deep Blue the game was closed and the computer did a lot of silly moves without any plan and lost.

The game Deep Blue - Svend Hamann can be seen here:
[url]http://www.chessgames.com/perl/che ... id=1472183[/url]

Because of the horizon effect (a lot of white checks) white allowed black a passed pawn on the a-file and lost. The Deep Blue team (Hsu, Murray and Hoane) told me, that Deep Blue only had needed a few seconds more to spot that the played move was a big error.

Jens
JensBNielsen

Re: Programs designed for especially good results against hu

Post by JensBNielsen »

bob wrote: I did something similar in Cray Blitz when playing blitz games against opponents, and you are correct. With the program moving _very_ quickly, the humans (GM level) tended to fall into the same tempo, where they were more likely to make tactical errors even though they had plenty of time on their clock...
But what do you think about it at 40/2?

Jens

PS. Long time ago I read much of your stuff in ICCA JOURNAL etc!
("It was twenty years ago today" - Beatles, Sgt. Pepper)
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Programs designed for especially good results against hu

Post by bob »

JensBNielsen wrote:
bob wrote: I did something similar in Cray Blitz when playing blitz games against opponents, and you are correct. With the program moving _very_ quickly, the humans (GM level) tended to fall into the same tempo, where they were more likely to make tactical errors even though they had plenty of time on their clock...
But what do you think about it at 40/2?

Jens

PS. Long time ago I read much of your stuff in ICCA JOURNAL etc!
("It was twenty years ago today" - Beatles, Sgt. Pepper)
I am not sure it will work there. There is an almost exponential degradation of tactical accuracy as time is shortened significantly. But at 40/2, the human could _always_ use 30 seconds and that culls a great many tactical mistakes that he would make if only using one second or less...

Probably worth testing, but something tells me it won't work at longer games... At 5 minute games, it is easy to lull your opponent into a faster tempo since he knows time is critical. But at 40/2hr, I am not sure that will happen.
JensBNielsen

Re: Programs designed for especially good results against hu

Post by JensBNielsen »

My aim was not to lull the opponent into a faster tempo - but to prevent him from pondering or taking a pause, while the computer is calculating - and to make him tired and stress him.

Any pause he takes gives him less time on the clock.

Another effect is, that he doesn't know if the computer is in the book, or he played the move that the computer expected (which usually gives an instant or fast reply).

Jens