Ray Johnstone wrote:

I think we can never know the answer to the question of the
solvability of chess:

well i've seen your site, but i 'm still convinced chess is a draw
(since a few weeks only);
only way to win is exploiting book mistakes of your opponents.
If someone like eg. J.Noomen would continue his work,
after a while top chess engines playing with his book should
get a draw result, that's 99.9999999999 % certain for me;
(with conventional chess rules of course)
and otherwise i could help Noomen i guess..

best regards,
jef
http://superchess.blogspot.com


http://members.iinet.net.au/~ray/Chessgames.htm

On Mon, 04 Feb 2008 19:07:24 +0100, jefk wrote:

Ray Johnstone wrote:

I think we can never know the answer to the question of the
solvability of chess:

well i've seen your site, but i 'm still convinced chess is a draw
(since a few weeks only);
only way to win is exploiting book mistakes of your opponents.
If someone like eg. J.Noomen would continue his work,
after a while top chess engines playing with his book should
get a draw result, that's 99.9999999999 % certain for me;
(with conventional chess rules of course)
and otherwise i could help Noomen i guess..

best regards,
jef
http://superchess.blogspot.com


http://members.iinet.net.au/~ray/Chessgames.htm

You may of course be right.
For me there remain 2 problems. One is that the chess tree branches
so rapidly that no computer can ever examine all positions.
The other is that even experts can be deceived by a position. A
classic example is
Byrne-Fischer 1963:
On page 297 of Fischer's "My 60 Memorable Games," Byrne is
quoted as saying "The culminating combination is of such depth that,
even at the very moment at which I resigned, both grandmasters who
were
commenting on the play for the spectators in a separate room believed
that I had a won game."White resigned on his 22nd move.
So if you decided to cut the tree short and do a computer-human
analysis by
printing the positions after
21 moves and have them evaluated by GMs there would likely be serious
errors.



www.iinet.com.au/~ray

On Feb 5, 8:56*am, Ray Johnstone wrote:
On Mon, 04 Feb 2008 19:07:24 +0100, jefk wrote:
Ray Johnstone wrote:

I think we can never know the answer to the question of the
solvability ofchess:

well i've seen your site, but i 'm still convincedchessis a draw
(since a few weeks only);
only way to win is exploiting book mistakes of your opponents.
If someone like eg. J.Noomen would continue his work,
after a while topchessengines playing with his book should
get a draw result, that's 99.9999999999 % certain for me;
(with conventionalchessrules of course)
and otherwise i could help Noomen *i guess..

best regards,
jef
http://superchess.blogspot.com

http://members.iinet.net.au/~ray/Chessgames.htm

You may of course be right.
For me there remain 2 problems. One is that thechesstree branches
so rapidly that no computer can ever examine all positions.
The other is that even experts can be deceived by a position. A
classic example is
*Byrne-Fischer 1963:
On page 297 of Fischer's "My 60 Memorable Games," Byrne is
quoted as saying "The culminating combination is of such depth that,
even at the very moment at which I resigned, both grandmasters who
were
commenting on the play for the spectators in a separate room believed
that I had a won game."White resigned on his 22nd move.
So if you decided to cut the tree short and do a computer-human
analysis by
printing the positions after
21 moves and have them evaluated by GMs there would likely be serious
errors.

/~ray- Hide quoted text -

- Show quoted text -

In Chess after every move new 30-40 moves are generated.

So For 1 depth : 30
2 depth : 900
3 depth: 27000
4 depth: 810000
5 depth: 24300000
6 depth :729000000
7 depth: 21870000000
8 depth: 656100000000
9 depth: 19683000000000
10 depth: 590490000000000

So just in 10 depth we reached 590 Trillion Moves.

Say Each Move Generation needs 1000 Calculations.

So 10 depth search needs 590,000 Trillion Calculations.

Fast Desktop works on 10 GFlops. So It will take 59,000,000 seconds on
fast desktop to perform exhausted 10 depth search. It will take 2
years to do 10 depth Checking of all moves.

10 depth means just 5 Moves. So for 5 Moves We need our desktops to
think for 2 Years.

In case we use Fastest Super Computer That work PetaFlop

PetaFlop SuperComputer will do 10 depth search in 1/2 second.

But if we go to say 15 depth Even Petaflop will take.

10 depth: 1/2 seconds
11 depth: 15 seconds
12 depth: 450 seconds
13 depth: 13,500 seconds
14 depth: 405,000 seconds
15 depth: 12,150,000 seconds

So 15 depth on a PetaFlop Computer which fastest supercomputer in the
world will take: 12,150,000 seconds

Which is 3375 hours. or 140 days or ~ 5 Months.

So to evaluate all moves upto 15 depth even on Fastest Supercomputer
that works at speed of Petaflop will take 5 Months.

To Solve Chess we need to search till atleast 100 depth. That will
take even more than Billions of Billions of Billions of years on a
Petaflop Super Computer.

Such a Computer can never be made atleast in next 100,000 years of Man
Kind even him Moores Law is Followed.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html

Here I said "I think" Which means what I think about something. It
does not mean I am going to bet for it just because I think of it.

Here, I am proposing something to have more information so that I can
judge whether I was Correct.

* Sanny, you already have oodles of information showing that your
program plays nowhere near perfectly on any of its levels. It is poor-
to-mediocre at best, atrocious at worst. You are quite clearly NOT
correct.

* So, you are either a liar or a fool. Which is it?

I think you have a bit Problem in understanding things. I am neither
liar nor a fool. You are doing unnecesary arguments on a thing that do
not matter much.

If you had beaten Master Level then it does not mean it cannot improve
further . It is improving every week. Earlier everyone used to play
with Master level and Win. Now most of them are playing with Beginner
& Easy Level. Why? The time taken by Master has dropped down from
earlier 1 hour to 10 Minutes now. Still people are not willing to play
with Master level as they are afraid of loosing the game.

Now I find many games where Beginner Level has Won. And I find that
everyones rating is going down.

If the improvement continues as before the day is not far when It will
play as good as Crafty/ Fritz/ Rybka and other Strong Chess Programs.

GetClub Game is made just to win against Human Players, I have no
motive of having it play with other Commercial programs. As it is
made just for fun and not for competing with other Commercial Chess
Programs.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html

Ray Johnstone wrote
For me there remain 2 problems. One is that the chess tree branches
so rapidly that no computer can ever examine all positions.

correct, but i don't think we have to examine all positions,
see my next reply to 'Sanny'

The other is that even experts can be deceived by a position. A
classic example is Byrne-Fischer 1963:
On page 297 of Fischer's "My 60 Memorable Games,"
found it, 18-12-1963

both grandmasters who were commenting on the play for the spectators
in a separate room believed that I had a won game."White resigned
on his 22nd move.
well that's humbug, or those grandmasters were not looking very well,
Byrnes last move Kf1 was a clear mistake, even an engine as Crafty
very quickly evaluates the result as minus 3, which usually will become
a loss.
Kf2 would have been better, although he already had an inferior position.
he went the wrong track in the opening already at move 12 with Qd2!?,
in Poldauf-Heinig (both about 2300) in 1988 12. Rac1 was played ,
a much better move.

So if you decided to cut the tree short and do a computer-human
analysis by printing the positions after 21 moves and have them
evaluated by GMs there would likely be serious errors.


that's not the way how i've done my analysis; please note i'm involved in
computer chess for about 15 years, and had quite some discussions
with other expert, like eg. Hyatt. In the positional database which i
use for my opening research most positions (more than 15 million)
were only loaded in if they had been played by GM's two times or more.
The end evaluations have been done mostly by Crafty, at a depth
deep enough to recognize mistakes like Byrne made, but i also
have made a lot of corrections, mostly automaticly by Crafty
at the early stage of the opening (resulting in many sidelines
and transpositions), and done a lot of manual finetuning/analysis,
last year with Rybka, with interesting results.

You could say ofcourse that a computer eval at ply 12 (6 moves) means
nothing
and that with a much longer combination of say 10 moves or longer even
evals as low (or high) as -3 can be compensated by later moves,
but this is extremely rare in practical/real games.

But let's use a scientific approach: the conjecture i've put forward is:
chess is a draw; (some prelimanry 'lemma's which confirm this
theory/ strong hypothesis/confecture are on my weblog,
http://superchess.blogspot.com

Now feel free to falsify this (you or others) and try to find
a game won by white where black didnt make much mistakes;
then i will show you the mistake black made, and i will
give you the correct move necessary for black to maintain
a drawish position. good luck..


jef

jefk wrote:
only way to win is exploiting book mistakes of your opponents.

That just isn't true. Sorry.


Dave.

--
David Richerby Poetic Perforated Postman (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a man who delivers the mail but
it's full of holes and in verse!

In article ,
Sanny wrote:
In Chess after every move new 30-40 moves are generated.

As Guy has pointed out, this is not the case for the
initial position; but it's a reasonable approximation further
down the tree.

[...]
So just in 10 depth we reached 590 Trillion Moves.
Say Each Move Generation needs 1000 Calculations.
So 10 depth search needs 590,000 Trillion Calculations.

Only if you feel the need to do a complete calculation
for *every* move sequence. (a) If you have a sufficiently large
transposition table, then positions at depth 10 are typically
reached by approximately (5!)^2 == 14400 different routes, meaning
you could save "something like" that factor in speed. (b) If you
are interested only in finding the evaluation [and best move] at
the root, then alpha-beta pruning roughly square-roots the number
of lines you need to look at, giving you an improvement by around
30^5 == 24300000. (c) Sadly, these effects are not entirely
independent, otherwise you could speed up your search by a factor
of around 300 000 000 000 [cutting your time from 2y to 0.2ms].
(d) But they are good enough that Fritz [etc] can do complete
searches to depth 12 and more [plus lots of extensions down the
"interesting" lines, often to depth 25+] in typical middle-game
positions in a few seconds. Go figure ....

To Solve Chess we need to search till atleast 100 depth. [...]

If you insist on searching every possible game, you are
short by a factor of around 120. Once we enter these realms of
fantasy, you are better off searching *positions* rather than
games, as there are *many* fewer of them. [But there are plenty
of legal positions that cannot be reached within 100 ply.]

Such a Computer can never be made atleast in next 100,000 years of Man
Kind even him Moores Law is Followed.

Actually *if* Moore's Law is followed, then we gain a
move every 3 to 7.5 years or so, depending on assumptions, and
1000 years of progress will work wonders. But long before that
we run up against some pretty fundamental laws of physics, so
at least we can say that we won't get 1000 years of Moore's Law
applying to conventional computers of the sort we have today.
After that, all bets are off. As Guy says, quantum computing
may come riding to the rescue. Or perhaps not.

[More interesting, to me at least, is that while Moore's
Law operates, we gain a piece in the endgame tablebases every
decade or so. So 10 or 11 piece tablebases are not implausible,
even with conventional computers. It seems hard to imagine the
Law operating *much* beyond that; but then, it always did.]

--
Andy Walker
Nottingham

Sanny wrote:
In Chess after every move new 30-40 moves are generated.



well Sanny, that might be correct, but many of these moves are tactical
mistakes, or give away material immediately. and with perfect play
the game in such a case is subsequently lost (excluding some deep
combinations, but these are recognized in 99.99999 % of the
cases by an engine as Rybka imho);

so in reality/practice, there are about 3-12 moves possible
for each side, in sharp play sometimes there's only one
correct move, in positional play/endgames there might
even be sometimes more moves possible than 12;
in any case its much less than the 3-40 you mention

So just in 10 depth we reached 590 Trillion Moves.


well it's less, see above, but it's not a smart way of generating
a database; if you generate a positional database (thats the
way modern opening books are made), because of
the many *transpositions*, the number of positions will
be mcu& *much* less.
In fact, if we exclude tactical mistakes, i believe it would
be possible to generate a database at 25 ply (12.5 moves)
deep with all possible correct/non losing moves.

Say Each Move Generation needs 1000 Calculations.

all this calculations can be done offline, and on other computers,
or by GM's; by loading in slow GM games from history i
already have used lots of knowledge/ human calcuation power..
\
And when using computers/engines, you forget a very important
element in speeding up the calculation, namely the socalled
*alfa/beta* algorithm. from what i vaguely remember
this speeds up the calculation process -to find the right move-
with about 50 times.
Not everybody seems to be aware of such methods, eg.
GM Kasparov in his book 'why chess looks like life',
claims that there are more positions in chess than
atoms in the univers. humbug, but Gary isn't an
astrophysicist, you know; he's a politician..


So to evaluate all moves upto 15 depth even on Fastest Supercomputer
that works at speed of Petaflop will take 5 Months.

i've been busy with computerchess about 15 years, and gradually
have been building up my database; while its'not perfect, i'm
certain that at least until about ply 25 no advantage for white
can be made, wheter you play Italian, 1.d4, or whatever.
In a later stage i thought that a minimal advantage with
the Ruy Lopez could be achieved, but with extensive
analysis with Rybka i've found some novelties and
now am convinced that black also can maintain a draw.

To Solve Chess we need to search till atleast 100 depth. That will
take even more than Billions of Billions of Billions of years
we don't have to fully analyze the game i think.
for every winning plan for white, i believe there
exists a proper defence system for black.
(there exist of course some simple games
where all possibilities *can* be calculated , and
it's proven with 100 % certainty they are a draw,
but indeed such a method does'nt seem
possible for chess, at least for some millennia)

example, although i'm not an expert in the
game of draughts (or checkers), from what i've
vaguely read/seen in some articles, the scientific/
computer games community seems to be
convinced these games are a draw, even although
they haven't fully calculated all possibilities;

well, i now am convinced the same hold for chess..

but if you disagree, well, than i challenge you just
like Ray Johnston -or others- to show me
a winning line for white; and then i guarantee
that i will find improvement(s) for black
which ensure the draw. Good luck..

best regards,
jef

David Richerby wrote:
jefk wrote:

only way to win is exploiting book mistakes of your opponents.


That just isn't true. Sorry.


well it's the main way nowadays , and as i wrote, when the
programs are getting stronger, eg. Rybka on a fast mp, against
eg. Hydra, the chance that the stronger comp (probably Hydra)
will find a winning plan (or combination, an even more unlikely
event) is very small, as the better the comps are
getting, the higher the draw percentage.

But anyway, whether a book mistake, or a winning plan
which cannot be defended by black, i als challenge you,
mr Richerby, to show me a winning game for white,
and i promise i will show you a better defence for black;
after which you can think for some years to find
another winning plan again.. (a futile effort,
and i know it, cause i tried it myself for years..)

So i believe there will be a 'weak solution' for chess, just as in
English draughts (checkers), see:

http://en.wikipedia.org/wiki/Solved_game

correcting my previous statement on Dutch draughts,
maybe there is no weak solution yet, but i'm
strongly convinced it's a draw, just like checkers.

as for chess, i forgot to mention the CAP project
organized by Dann Corbit, well known by experienced
computer chess experts; the program Chess Assistant
claims to have a CAP database with about 30 million
positions, but i don't believe all these positions have
been analyzed by top engines, and certainly not Rybka.
Maybe we should repeat this CAP exercise with
many Rybka engines ? ah well, a futile effort,
a the result is known by me: chess is a draw..

Although not proven as strongly as the program
Chinook did with checkers, with Rybka, further
anallysis and using some other chess experience,
for me i can add chess to the games which have
been solved with the 'weak' method.
Chances that it's a win for white is indeed,
smaller than one divided by the nr of atoms
in the universe.(haven't counted them though..)

best regards,
jef

jefk wrote:
Sanny wrote:
In Chess after every move new 30-40 moves are generated.

well Sanny, that might be correct, but many of these moves are
tactical mistakes, or give away material immediately. and with
perfect play the game in such a case is subsequently lost (excluding
some deep combinations, but these are recognized in 99.99999 % of
the cases by an engine as Rybka imho);

so in reality/practice, there are about 3-12 moves possible for each
side, in sharp play sometimes there's only one correct move, in
positional play/endgames there might even be sometimes more moves
possible than 12; in any case its much less than the 3-40 you
mention

You're begging the question. You can't know that all of those moves
are tactical mistakes until you've analyzed them fully. Now, if all
you're interested in is playing *good* chess, it's acceptable to have
a quick look at these moves and throw them away if they don't come
good within a few ply. However, if you want to play *perfect* chess,
you have to look at absolutely everything. It might just be that the
winning strategy for White involves a queen sacrifice on move five
that doesn't pay off until move 100. Staggeringly unlikely but still
possible.

To Solve Chess we need to search till atleast 100 depth. That will
take even more than Billions of Billions of Billions of years

we don't have to fully analyze the game i think.

You can't *solve* chess without fully anazlyzing it.

for every winning plan for white, i believe there exists a proper
defence system for black.

For a *winning* plan for white there is, by definition, no proper
defence. If there were a proper defence, the plan wouldn't be
winning!

(there exist of course some simple games where all possibilities
*can* be calculated , and it's proven with 100 % certainty they are
a draw, but indeed such a method does'nt seem possible for chess, at
least for some millennia)

Agreed.

example, although i'm not an expert in the game of draughts (or
checkers), from what i've vaguely read/seen in some articles, the
scientific/ computer games community seems to be convinced these
games are a draw, even although they haven't fully calculated all
possibilities;

Schaeffer et al have fully solved checkers.


Dave.

--
David Richerby Confusing Cheese Cat (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a cat that's made of cheese but you
can't understand it!