How do you see computer chess in 100 years?

[Kempelen]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

[Matthias Gemuh]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

100 years ?
This world, as we know it, will not survive that long.

Prophet Matthias.

[Kirill Kryukov]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

Chess will be practically solved by that time, like checkers today. Which means that games between top programs will always (or almost always) end in a draw. Programmers will move on to more challenging games, like Shogi.

[S.Taylor]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

Chess will be practically solved by that time, like checkers today. Which means that games between top programs will always (or almost always) end in a draw. Programmers will move on to more challenging games, like Shogi.

For SURE, chess programming will be long ended. In fact, as soon as Rybka reaches version 6, I don' know if here will be much more, except for some almost futile efforts at trying to get better results with planning etc. This may continue for a further 10 years. So 15 years morer, and programming will be history. ( will it? I assume so!).

Solved? No, PRACTICALLY solved, yes. But a long way from solved. We already see it in the air that chess is "practically solved" in that it is unlikely that any chess playing human or machine will win a match vs Rybka, as of now.

[swami]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

* Programmers will run out of ideas.

* The Elo limit will be 3600 - 3800 on computer chess rating scale.

* And the strongest engine will be one of the Fruit/Toga derivatives on the faster hardware. Free Source code.

* Chess engine development will be sporadical. At the estimated average rate of zero to one genuine engine per year for the following ten years on.

* programmers then will mostly try to improve and modify the existing strongest free derivative.

* Testers will stop running tournaments on genuine (weaker) engines, very few people will try running the tournaments between modified Toga's (and such) only to find 95% draws.

* Weaker opening lines will be eliminated. The opening book size will be diminished, as the result to only include lines that are equal for both the sides.

* There will be no more official computer chess championship.

* Very few chess programmers will move on to other complex board games.

* Some other board game (that's equivalent to chess) will be commonplace. Perhaps something that's difficult for computers to come near to getting solved. 10 x 10 board or something.

[George Tsavdaris]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

Chess will be practically solved by that time, like checkers today.

Checkers has been weakly solved by using from 100 to 200 computer for 18 consecutive years to search 10^14 positions(out of 10^20 possible of Checkers).

Even if we assume that we will have 10000 times better computers than these in the period of 1989-2007 that had weakly solved Checkers then:

The computers that solved Checkers were searching:
10^14 / 18 = 5·10^12 positions per year.

Our computers(and this even is an overestimation) will search 10000 times faster so they will search:
5·10^16 positions per year.
But Chess has around 10^50 positions so let's say like Checkers we would have to search in order to weakly solve Chess, only 10^44 positions.
That with the aforementioned speed will take around:
10^44 / (5·10^16) = 2·10^27 years.

Checkers approach can't work with Chess.

Which means that games between top programs will always (or almost always) end in a draw.

How does this is implied? (the DRAW i mean)

[pedrox]

In year 2011, we will get the 1000 engines.

In 2013 someone will make a chess engine generator, any can choose which advanced searching techniques and knowledge of evaluation include his engine, after this generator will create a unique executable, the number of engines will are 4000 at the year 2014.

In year 2016, Rybka continues to dominate the other programs, although now by a narrow margin. But there is an unexpected fact, the program Rybka seeks to exploit the resources of the computer at 200% and many users see their computers burned, the company of Vas is demanded and she is going to ruin.

In year 2017, Dr. Yuri Osipov reveals the secrets of Rybka 11, some programmers apply some of this knowledge to their engines and some manage to surpass Rybka.

From the year 2018 becomes more difficult to raise the ELO, is considered a success when a new version win the previous 5 points.

In 2050, a well-known personage business called Vegan said he could prove that chess is resolved through the opening books and his tables of the endgames of 16 pieces.

[Zach Wegner]

In 2050, a well-known personage business called Vegan said he could prove that chess is resolved through the opening books and his tables of the endgames of 16 pieces.

[Kirill Kryukov]

Checkers has been weakly solved by using from 100 to 200 computer for 18 consecutive years to search 10^14 positions(out of 10^20 possible of Checkers).

Even if we assume that we will have 10000 times better computers than these in the period of 1989-2007 that had weakly solved Checkers then:

The computers that solved Checkers were searching:
10^14 / 18 = 5·10^12 positions per year.

Our computers(and this even is an overestimation) will search 10000 times faster so they will search:
5·10^16 positions per year.
But Chess has around 10^50 positions so let's say like Checkers we would have to search in order to weakly solve Chess, only 10^44 positions.
That with the aforementioned speed will take around:
10^44 / (5·10^16) = 2·10^27 years.

Checkers approach can't work with Chess.

That's why I said "practically solved", as defined by results of games between top programs. I also don't think that anyone will take effort necessary to weakly solve chess.

Which means that games between top programs will always (or almost always) end in a draw.

How does this is implied? (the DRAW i mean)

Current evidence does not suggest that game of chess is not a draw with the best play. Proportion of draws is higher in games of strong programs, than in games of weak programs. You can see this, for example, in "Proportion of draws" tables on CCRL sites.

[Sylwy]

How do you see computer chess in one hundred years? Do you think there will not be people programming due to all will be invented? or do you think there will be programmers who will fun programing engines?

100 years ?
This world, as we know it, will not survive that long.

Prophet Matthias.

Agreed !