engines and deep analysis

   		Free Internet Chess Games Server

Install FICGS apps
play chess online


Game result  (chess)


I. Cirulis, 2094
F. de la Foret, 1960

1/2-1/2

See game 107678




 Hot news
 Discussions
 Files search
 Social network



European Go Database

                                          
Forum


Back to forum


Normajean Yates    (2009-01-15)
engines and deep analysis...

A year ago, Thibault posted links to record of GMs v engines -- but almost all games were blitz.

raising questions:

1. At standard [2.5 hours / 40 moves] time controls, and at correspondence time controls [like here - or even 1 day/move] - humans are still better than engines. true?

If true, how long are they expected to be better?

[I mean in competitive chess, not in specifically design positions which are at present very difficult for engines]

2. which engine is best for standard and correpondence time controls (as I defined in 1)?


Thibault de Vassal    (2009-01-15 12:34:27)
Engines vs. Human

Hi Normajean,

1. At standard time controls, I think Rybka 3 can beat the best GM, there is just no time enough for a human to avoid a single error. But grandmasters probably still have a better vision of the game at least in calm positions. At correspondence time controls, no one can say it but I feel a good GM could rivalize yet with the best engines. This is unlikely to change before a while IMO.

2. The best anti-human engine at any time control is probably Rybka 3, but there will be concurrence soon.


Normajean Yates    (2009-01-15 14:18:37)
thanks for the info, thibault!

its all in the title, like my other short posts :)


Wayne Lowrance    (2009-01-16 05:22:27)
Engines & deep analysis

I agree with Thibault, in blitz or standard time controls Rybka and 2 or 3 others have no pier with human top gm's. But in correspondence time, I think the top GM's will certainly hold their own, or more Wayne


Normajean Yates    (2009-01-16 05:49:31)
thanks thib & wayne - I thought so 2...

..except at standard time controls. At standard I thought top GMs would be better even now :(

IMO, two of the reasons why correspondence is still an exception:

1. engines still understand positional aspects in a clumsy way (mainly through eval function even now I think..)[a]

2. top engines are commercial - so they have to 'show off' to compete in the market - 'showing off time' at corrspondence is too long for the software market.. so top engines are tuned towards faster play...


[a] I wish that after copyright etc. expires, commercial chess engine vendors must be legally forced to make public their algorithms.

(Ideally, I wish - no copyright, only moral right of actual authors! - but that needs a diiferent economic system than capitalism)








 		 

Message



Chat


 

February 18, 2019

 FICGS is also a social network including seo forums, a hot news & buzz blog, a free web directory and discussion forums to meet people from all over the world. Discuss the last events, improve your search engines optimization, submit your website, share your interests...


Feel free to link to FICGS chess server, register & win Epoints :






FICGS Go server, weiqi baduk banner facebook      
Correspondence chess

World championship

Play chess games

Go (weiqi, baduk)

Advanced chess

Play big chess

Chess trainer apk

Rated tournaments

Poker texas hold'em

Fischer random chess

       FICGS correspondence chess banner facebook