| |
Main
Date: 22 Aug 2006 13:43:39
From: voil
Subject: as a chess player,what do you think of eastern game Go?
|
it's question to only those who play chess only while
knowing Go but not plays it..
|
|
| |
Date: 27 Sep 2006 10:08:42
From:
Subject: Re: as a chess player,what do you think of eastern game Go?
|
I admire chess and go as two of the classiest games in the world.
As a player, I suck at both games.
I've noticed that go has more "snob appeal" these days. Fans point out
that it's much subtler, with far more strategic options than chess.
There may be some truth to that, but sometimes such comments are made
by people who are better at criticizing than at actually playing a
game.
Here's a great little article comparing chess, go, and checkers:
http://tinyurl.com/l5a47
Personally, I keep coming back to chess. It just has the right "feel"
to me, and since I play exclusively against the computer AI, chess may
be the perfect game for me: there's plenty of excellent chess software
around, and there's also a wealth of literature to help me along.
Also, my greatest weakness is tactics, and tactics are especially
important in chess; so playing the game is a form of self-improvement.
But go is a superb game too. So is checkers.
--Patrick
|
| |
Date: 22 Sep 2006 13:25:43
From: [email protected]
Subject: Re: as a chess player,what do you think of eastern game Go?
|
> Unless I'm mistaken, you can easily store a checkers position in 80
> bits (a 32-bit field giving the occupied squares and two 24-bit fields
> indicating which pieces are white and which are kings).
That would work; it is more storage efficient, but perhaps slightly
less computationally efficient, than three 32 bit fields.
> Since chess
> seems to require about 160 bits to represent a position, this implies
> that the number of chess positions exceeds the number of checkers
> positions by a factor of about 2^80, which is about twenty-four
> decimal orders of magniture. That's roughly the difference between a
> microsecond and thirty billion years (about twice the current age of
> the universe) or, if you prefer, between the size of a small virus and
> the distance to the nearest star.
You certainly make a good point about chess having unimaginably more
possible positions than checkers. Even if you reduce this to possible,
or even practical, positions, your comparison would still stand, and if
it is 2**75 instead of 2**80 your point would be just as valid. So I
concede that you are right; solving chess might take longer than 10
years :) That raises the theoretical question of just how long, if
ever, it might take.
Let's say that computing power increases according to the misquoted
version of Moore's law, doubling every 18 months or so, and let's say
checkers will be solved very soon. (Approximate numbers are more than
good enough for this sort of thing.) So in ten years computing power
goes up about 15 doubling cycles or 2**15. If we say chess is 2**75
harder than checkers then we need about 50 years to have enough
computing power just to tackle the job. Actual computation time is
hard to say; it will probably be substantial.
So a better estimate may be that chess will be solvable in around 50
years, and perhaps actually solved in 75. Maybe. Because the number
of chess positions turns out to be pretty close to estimates of the
number of atoms in the known universe, so storage of those positions
could be an issue :)
|
| | |
Date: 25 Sep 2006 10:59:27
From: David Richerby
Subject: Re: as a chess player,what do you think of eastern game Go?
|
[email protected] <[email protected] > wrote:
> David Richerby wrote:
>> Unless I'm mistaken, you can easily store a checkers position in 80
>> bits (a 32-bit field giving the occupied squares and two 24-bit
>> fields indicating which pieces are white and which are kings).
>
> That would work; it is more storage efficient, but perhaps slightly
> less computationally efficient, than three 32 bit fields.
Oh, it's much less computationally efficient that three 32-bit fields
-- working out the colour and status of a piece requires counting the
number of set bits in BOARD before that piece's square and then
looking at the relevant bit of WHITE and KINGS. Much more efficient
to do it in 32-bit fields where you can just take BOARD&WHITE to find
the white pieces, BOARD&BLACK&KINGS to find the black kings and so on.
But I was interested in the representation from an information-
theoretic, rather than computational, point of view.
> Let's say that computing power increases according to the misquoted
> version of Moore's law, doubling every 18 months or so, and let's
> say checkers will be solved very soon. (Approximate numbers are
> more than good enough for this sort of thing.) So in ten years
> computing power goes up about 15 doubling cycles or 2**15. If we
> say chess is 2**75 harder than checkers then we need about 50 years
> to have enough computing power just to tackle the job. Actual
> computation time is hard to say; it will probably be substantial.
I don't have a cite but I read a wonderful result a little while ago.
If you have a computational task that's going to take more than
thirteen months of CPU time, it's better to invest the money you were
going to spend, for half of that time. During this time, you'll earn
a little interest on your deposit, computers will get significantly
cheaper and slightly faster (more than slightly if we're talking, say,
a couple of years, rather than six months). At the half-way point of
your original task, buy a much more powerful computer than you would
have been able to afford on day one and that will finish the job
earlier.
This definitely applies to solving chess! :-)
Dave.
--
David Richerby Radioactive Soap (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ personal hygiene product but it'll
make you glow in the dark!
|
| |
Date: 20 Sep 2006 12:40:32
From: [email protected]
Subject: Re: as a chess player,what do you think of eastern game Go?
|
David Richerby wrote:
> [email protected] <[email protected]> wrote:
> > This is true as well of Chess and Shogi. It will take much longer to
> > "solve" chess but I would say ten years is certainly long enough.
>
> Do you have any evidence for that? I'm not aware of any techniques
> which make the solving of chess remotely close.
I have no evidence whatsoever. It is just a guess based on the rapid
rate at which technology advances.
Innovations such as quantum computing will make some extraordinary
things possible and ten years in the computer business is a very long
time.
But, to be a little more specific. Checkers is being solved by
developing larger and larger endgame databases. The largest full
database is 10 pieces (from a starting complement of 24 pieces on the
board) and there are partial databases for 11 pieces. As I understand
it, the University of Alberta team works exhaustive searches until it
can be shown that a position leads to a definitively solved position or
positions in the database. The original position is then considered
solved (which indeed it is) if those "forced" conclusions are
consistent. (Anyone out there please correct this if I haven't
expressed it correctly.) They have thus far shown that numerous
positions of interest to checker players are draws, some of them quite
unbalanced in over the board play but nevertheless containing often
surprising drawing lines of play.
I would expect that "solving" chess would be done in a similar manner;
if a position arrives at an inevitable tablebase result, it is then
solved. Of course the amount of solving to do is many, many orders of
magnitude greater.
Will it happen? While I don't have any real evidence for a timescale,
it seems inevitable. What was fiction just ten years ago is reality
today.
|
| | |
Date: 21 Sep 2006 13:42:35
From: David Richerby
Subject: Re: as a chess player,what do you think of eastern game Go?
|
[email protected] <[email protected] > wrote:
> David Richerby wrote:
>> [email protected] <[email protected]> wrote:
>>> This is true as well of Chess and Shogi. It will take much longer
>>> to "solve" chess but I would say ten years is certainly long
>>> enough.
>>
>> Do you have any evidence for that? I'm not aware of any techniques
>> which make the solving of chess remotely close.
>
> I have no evidence whatsoever. It is just a guess based on the rapid
> rate at which technology advances.
OK. I don't think technology's advancing fast enough for that.
> Innovations such as quantum computing will make some extraordinary
> things possible and ten years in the computer business is a very
> long time.
Firstly, not all problems have good quantum algorithms. Secondly,
quantum computers are still moving very slowly -- look how long it
took transistor-based computers to get fast enough to play decent
chess!
> But, to be a little more specific. Checkers is being solved by
> developing larger and larger endgame databases. The largest full
> database is 10 pieces (from a starting complement of 24 pieces on
> the board) and there are partial databases for 11 pieces.
OK but checkers is a *much* smaller game than chess. There are only
two types of piece (vs six), only 24 pieces (vs 32) and only 32
squares on the board (vs 64). Additionally, checkers is a much more
forcing game than chess because captures are obligatory.
> I would expect that "solving" chess would be done in a similar
> manner; if a position arrives at an inevitable tablebase result, it
> is then solved. Of course the amount of solving to do is many, many
> orders of magnitude greater.
Unless I'm mistaken, you can easily store a checkers position in 80
bits (a 32-bit field giving the occupied squares and two 24-bit fields
indicating which pieces are white and which are kings). Since chess
seems to require about 160 bits to represent a position, this implies
that the number of chess positions exceeds the number of checkers
positions by a factor of about 2^80, which is about twenty-four
decimal orders of magniture. That's roughly the difference between a
microsecond and thirty billion years (about twice the current age of
the universe) or, if you prefer, between the size of a small virus and
the distance to the nearest star.
Dave.
--
David Richerby Impossible Radioactive.com (TM):
www.chiark.greenend.org.uk/~davidr/ it's like an E-commerce portal but
it'll make you glow in the dark and
it can't exist!
|
| |
Date: 12 Sep 2006 14:38:39
From: [email protected]
Subject: Re: as a chess player,what do you think of eastern game Go?
|
> Also, go is very abstract game. Chess and shogi on the other side are
> games with strict rules, you can not implement them in real life and
> they most likely to be "solved" (that means computers will become too
> good, and games will become too boring) in the next half century.
I disagree with the contention that games that become computer-solved
necessarily become boring for humans. Checkers, for instance, is
likely to be solved fully by Jonathan Schaeffer's team at University of
Alberta within two years (he actually says one year). This does not
mean it will no longer be a deep and challenging game for humans,
although correspondence play allowing the use of computers as an aid
will likely die off.
This is true as well of Chess and Shogi. It will take much longer to
"solve" chess but I would say ten years is certainly long enough. Will
people then find it boring and stop playing? Not a chance.
But it is true that computer-playing Go programs aren't really very
good yet, perhaps not much beyond the level of a good intermediate or
lower expert human. The game just doesn't seem quite so amenable to
computer solution.
|
| | |
Date: 18 Sep 2006 19:27:25
From: David Richerby
Subject: Re: as a chess player,what do you think of eastern game Go?
|
[email protected] <[email protected] > wrote:
> This is true as well of Chess and Shogi. It will take much longer to
> "solve" chess but I would say ten years is certainly long enough.
Do you have any evidence for that? I'm not aware of any techniques
which make the solving of chess remotely close.
> Will people then find it boring and stop playing? Not a chance.
Exactly.
> But it is true that computer-playing Go programs aren't really very
> good yet, perhaps not much beyond the level of a good intermediate or
> lower expert human. The game just doesn't seem quite so amenable to
> computer solution.
Exactly.
Dave.
--
David Richerby Miniature Mentholated Soap (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a personal hygiene product but
it's invigorating and you can hold in
it your hand!
|
| |
Date: 12 Sep 2006 20:06:49
From: Bark!
Subject: Re: as a chess player,what do you think of eastern game Go?
|
On Tue, 22 Aug 2006 06:43:39 +0200, voil <[email protected] > wrote:
> it's question to only those who play chess only while
> knowing Go but not plays it..
>
>
As Morbo from Futurama would say: - I have no particular opinion what so=
=
ever.
-- =
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
|
| |
Date: 23 Aug 2006 08:16:15
From: Alkhimey
Subject: Re: as a chess player,what do you think of eastern game Go?
|
> On the other hand, the increasing number of draws
> makes me wonder if most chess players will end up
> switching to Go or Shogi eventually.
Thi hight number of draws in chess is becouse most pro chess players
are lazy. Most players play games " just for the protocol". Notice
that when a game of death and life being played (a game that decides
the first place in a major turney for example) the game turns suddenly
interesting because each player wants very much to win.
What I like in go is the handicups, weak players can play interesting
and not frustrating games with better ones.
Also, go is very abstract game. Chess and shogi on the other side are
games with strict rules, you can not implement them in real life and
they most likely to be "solved" (that means computers will become too
good, and games will become too boring) in the next half century.
[email protected] wrote:
> cmcanulty wrote (22 Aug 2006 06:54:53 -0700):
>
> 7 What i like about go is there isn't as much memorization
> 7 as ches a lot more is just visual and spatial logic, but I'm
> 7 a beginner at GO.
>
> _
> I get the impression that, past the beginner level, there
> is a lot of stuff to memorize - josekis, status of certain
> shapes of territory, even openings, ...
> _
> One of the things I like about chess is the possibility
> for a relatively weak player (such as myself) to enjoy,
> at least some famous games, such as Edward Lasker
> against George Thomas:
> _
> 1 d4 e6 2 Nf3 f5 3 Nc3 Nf6 4 Bg5 Be7 5 Bxf6 Bxf6
> 6 e4 fxe4 7 Nxe4 b6 8 Ne5 0-0 9 Bd3 Bb7 10 Qh5 Qe7
> 11 Qxh7+ Kxh7 12 Nxf6+ Kh6 13 Neg4+ Kg5
> 14 h4+ Kf4 15 g3+ Kf3 16 Be2+ Kg2 17 Rh2+ Kg1
> 18 Kd2#
> _
> On the other hand, the increasing number of draws
> makes me wonder if most chess players will end up
> switching to Go or Shogi eventually.
|
| |
Date: 23 Aug 2006 01:33:31
From:
Subject: Re: as a chess player,what do you think of eastern game Go?
|
cmcanulty wrote (22 Aug 2006 06:54:53 -0700):
7 What i like about go is there isn't as much memorization
7 as ches a lot more is just visual and spatial logic, but I'm
7 a beginner at GO.
_
I get the impression that, past the beginner level, there
is a lot of stuff to memorize - josekis, status of certain
shapes of territory, even openings, ...
_
One of the things I like about chess is the possibility
for a relatively weak player (such as myself) to enjoy,
at least some famous games, such as Edward Lasker
against George Thomas:
_
1 d4 e6 2 Nf3 f5 3 Nc3 Nf6 4 Bg5 Be7 5 Bxf6 Bxf6
6 e4 fxe4 7 Nxe4 b6 8 Ne5 0-0 9 Bd3 Bb7 10 Qh5 Qe7
11 Qxh7+ Kxh7 12 Nxf6+ Kh6 13 Neg4+ Kg5
14 h4+ Kf4 15 g3+ Kf3 16 Be2+ Kg2 17 Rh2+ Kg1
18 Kd2#
_
On the other hand, the increasing number of draws
makes me wonder if most chess players will end up
switching to Go or Shogi eventually.
|
| |
Date: 23 Aug 2006 01:31:12
From:
Subject: Re: as a chess player,what do you think of eastern game Go?
|
cmcanulty wrote (22 Aug 2006 06:54:53 -0700):
7 What i like about go is there isn't as much memorization
7 as ches a lot more is just visual and spatial logic, but I'm
7 a beginner at GO.
_
I get the impression that, past the beginner level, there
is a lot of stuff to memorize josekis, status of certain
shapes of territory, even openings, ...
_
One of the things I like about chess is the possibility
for a relatively weak player (such as myself) to enjoy,
at least some famous games, such as Edward Lasker
against George Thomas:
_
1 d4 e6 2 Nf3 f5 3 Nc3 Nf6 4 Bg5 Be7 5 Bxf6 Bxf6
6 e4 fxe4 7 Nxe4 b6 8 Ne5 0-0 9 Bd3 Bb7 10 Qh5 Qe7
11 Qxh7+ Kxh7 12 Nxf6+ Kh6 13 Neg4+ Kg5
14 h4+ Kf4 15 g3+ Kf3 16 Be2+ Kg2 17 Rh2+ Kg1
18 Kd2#
_
On the other hand, the increasing number of draws
makes me wonder if most chess players will end up
switching to Go or Shogi eventually.
|
| |
Date: 22 Aug 2006 13:10:27
From:
Subject: Re: as a chess player,what do you think of eastern game Go?
|
voil wrote:
> it's question to only those who play chess only while
> knowing Go but not plays it..
Go's popular in East Asia. You take turns placing white and black
stones. Choosing where to place your stones takes skill.
The rest of the world plays Chess, a game with complex moves, where
knights lay siege to castled kings. It also takes skill.
Which you prefer probably depends on where you were born.
|
| |
Date: 22 Aug 2006 06:54:53
From: cmcanulty
Subject: Re: as a chess player,what do you think of eastern game Go?
|
What i like about go is there isn't as much memorization as ches a lot
more is just visual and spatial logic, but I'm a beginner at GO.
|
| | |
Date: 22 Aug 2006 15:51:06
From: David Richerby
Subject: Re: as a chess player,what do you think of eastern game Go?
|
cmcanulty <[email protected] > wrote:
> What i like about go is there isn't as much memorization as ches a lot
> more is just visual and spatial logic, but I'm a beginner at GO.
It's possible to play very strong chess with almost no memorization.
Granted, you won't get to the very highest levels without some opening
theory but, frankly, most people aren't getting there anyway.
Dave.
--
David Richerby Incredible Devil Flower (TM): it's
www.chiark.greenend.org.uk/~davidr/ like a flower that's possessed by
Satan but it'll blow your mind!
|
|
