Go game lessons for chess players

Main

Date: 21 Dec 2005 02:46:12
From:
Subject: Go game lessons for chess players

6-times European Go Champion Alexandre Dinerchtein is looking for new
students.
His prices are even lower than the prices of some amateur teachers.
Don't waste your time and money from learning amateurish mistakes!
Learn Go from a pro master!
The fee of the standard lesson (1.5 hour each) is only $25-$30 ! Group
lessons are available too.
He is also making free tournament game commentaries for regular
students. It's a good bonus!
Alexandre is also offering simultaneous games and life commentaries for

tournament organizers.
More details and contact info on:
http://breakfast.go4go.net/

Try this game! It's much more interesting than chess. Computers cannot
play it at all!

Date: 29 Dec 2005 16:46:11
From: Nick
Subject: Re: Go game lessons for chess players

David Richerby wrote:
> In order to allow interesting games between players of unequal strength,
> it is possible to play with handicap (similar to an odds game in chess).
> Black places an agreed number of stones on the board in standard positions
> as his first move and then the game continues as normal. The placement of
> the handicap stones is given at
>
> http://senseis.xmp.net/?path=Handicap&page=HandicapPlacement
> http://senseis.xmp.net/?HandicapsBeyondNineStones
>
> Somebody somewhere in rec.games.chess.* suggested a way of using this
> idea to give a better handicap system to chess than odds games. Essentially,
> the Go system is to let the first player make more than one move before
> the second player starts (though these moves have to be in standard
> places). To transfer this to chess, both players start with all their pieces
> but white gets to make some agreed number of moves (no captures allowed)
> before black's first move. After that, the players alternate as usual. I think
> this is better than removing pieces from the board at the start as that alters
> the character of the game too much and in unexpected ways.

This proposal could make sense only if White's permitted to make no
more than four moves. Let's suppose White could make five moves:
1 e4 2 Bc4 3 Nf3 4 Ne5 5 Qh5. Then what could Black do to avoid
being mated? 5...Nh6 6 Bxf7+ Nxf7 7 Qxf7#

--Nick

Date: 30 Dec 2005 10:25:13
From: David Richerby
Subject: Re: Go game lessons for chess players

Nick <nickbourbaki3@yahoo.co.uk > wrote:
> David Richerby wrote:
>> To transfer this to chess, both players start with all their pieces
>> but white gets to make some agreed number of moves (no captures
>> allowed) before black's first move. After that, the players alternate
>> as usual.
>
> This proposal could make sense only if White's permitted to make no
> more than four moves. Let's suppose White could make five moves:
> 1 e4 2 Bc4 3 Nf3 4 Ne5 5 Qh5. Then what could Black do to avoid
> being mated? 5...Nh6 6 Bxf7+ Nxf7 7 Qxf7#

The proposal makes perfect sense for any number of moves. Once a player
is good enough to spot this attack, though, you should stop giving them
more than four moves!

Dave.

--
David Richerby Mentholated Dangerous Game (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a family board game but
it could explode at any minute and
it's invigorating!

Date: 28 Dec 2005 19:23:41
From: Nick
Subject: Re: Go game lessons for chess players

David Richerby wrote:
> <superchess@wanadoo.nl> wrote:
> > superchess@wanadoo.nl wrote:
> >> David Richerby wrote:
> >>> The extra half point is to guarantee that the game cannot be drawn.
> >>
> > let me explain a bit more.
> > on *average* white and black seem to have equal winning chances in Go,
> > even although one single game cannot be drawn.
>
> That's not actually the case. You seem to be confusing handicap and komi
> (and the two aspects of komi). Allow me to explain.
>
> In order to allow interesting games between players of unequal strength,
> it is possible to play with handicap (similar to an odds game in chess).
> Black places an agreed number of stones on the board in standard positions
> as his first move and then the game continues as normal. The placement of
> the handicap stones is given at
>
> http://senseis.xmp.net/?path=Handicap&page=HandicapPlacement
> http://senseis.xmp.net/?HandicapsBeyondNineStones
>
> Somebody somewhere in rec.games.chess.* suggested a way of using this idea
> to give a better handicap system to chess than odds games. Essentially,
> the Go system is to let the first player make more than one move before
> the second player starts (though these moves have to be in standard
> places). To transfer this to chess, both players start with all their
> pieces but white gets to make some agreed number of moves (no captures
> allowed) before black's first move. After that, the players alternate as
> usual. I think this is better than removing pieces from the board at the
> start as that alters the character of the game too much and in unexpected
> ways. (For example, IIRC Silman says that he finds it easier to give rook
> odds than knight odds because, with rook odds, a strong player can often
> checkmate before the rooks come into the game but the absence of a knight
> is felt immediately.)
>
> Anyway, back to Go. Komi exists to compensate the second player (white)
> for the disadvantage of going second. Professional practice indicates
> that black's advantage from the first move is about six points of
> territory which doesn't sound very significant out of 381 intersections

19 x 19 = 361, not 381.

> on the board but actually is: most professional games are decided by the odd
> point. Playing with n points of komi means that black has to end up with
> at least n points more than white to win the game. For integer values of
> n, this means that the game can be drawn so it is usual to play with
> fractional komi -- in effect, this means that white wins `drawn' games.
> Standard komi is 6.5 points. To avoid drawn games, playing `without'
> komi tends to mean playing with komi of 0.5.

Here's an article from the American Go Association about komi:
http://www.usgo.org/org/komi.asp

> This can't really be applied to chess because the endgame changes
> dramatically if draws are won for black: for example, many positions in
> K+P vs K would be lost for white, even if he had the pawn.
> This seems silly.

Some people seem to believe that otherwise drawn positions in chess
*should* be won or lost on the basis of being able to click a mouse
more rapidly in an interval of a few seconds.

--Nick

Date: 24 Dec 2005 21:47:59
From:
Subject: Re: Go game lessons for chess players

qp@omen.ru wrote:
>
> Try this game! It's much more interesting than chess. Computers cannot
> play it at all!

well they can beat amateurs. and in a few
hundred years i guess some of the top computer
programs will beat the top Go players.

but at least you're right that learning Go can
improve your middlegame planning& endgame intuition
in chess, even although it's totally different
than chess (it worked for me a little bit, even
although at first i found go to be a boring game)

and Go won't be solved by computers
as there's the handicap system
which currently is not fashionable
in chess; maybe we should take
out the f3 pawn for white in chess
to compensate the white advantage
:)
best regards
& seaons greetings
www.superchess.com

Date: 24 Dec 2005 23:23:02
From: Jerzy
Subject: Re: Go game lessons for chess players

> and Go won't be solved by computers
> as there's the handicap system
> which currently is not fashionable
> in chess; maybe we should take
> out the f3 pawn for white in chess
> to compensate the white advantage

Not f3 but f2 pawn.
And AFAIK it`s not the handicap that makes it difficult for go programs to
win with protein brains.
Although go rules are simple, it`s difficult to count the territory for
silicon brains. So there`s no simple function that estimates which side has
an
advantage as it is in chess.

Date: 25 Dec 2005 11:09:07
From: David Richerby
Subject: Re: Go game lessons for chess players

Jerzy <jciruk@poczta.fm > wrote:
> And AFAIK it`s not the handicap that makes it difficult for go programs
> to win with protein brains.

No, the handicap has no effect on any of this.

> Although go rules are simple, it`s difficult to count the territory for
> silicon brains. So there`s no simple function that estimates which side
> has an advantage as it is in chess.

The main problem is that Go has a much higher branching factor than chess:
there are approximately 289-n choices for the nth move (maybe less because
of the illegality of suicide and the possibility of ko; maybe more because
of the possibility of stones having been captured). In addition to this,
it's harder to do good move ordering for Go. In chess, you can look at
captures first (in the order of least valuable piece capturing most
valuable piece first) and then checks and then everything else, say; in
Go, it's easy to order captures first but there are only likely to be a
couple of those in any one position and all the other moves can look
rather similar.

Dave.

--
David Richerby Perforated Indelible Apple (TM):
www.chiark.greenend.org.uk/~davidr/ it's like a tasty fruit but it can't
be erased and it's full of holes!

Date: 25 Dec 2005 13:02:58
From: Jerzy
Subject: Re: Go game lessons for chess players

>> Although go rules are simple, it`s difficult to count the territory for
>> silicon brains. So there`s no simple function that estimates which side
>> has an advantage as it is in chess.
>
> The main problem is that Go has a much higher branching factor than chess:

If you are trying to solve go in a similar manner as chess then you are
correct but as I mentioned above the main problem with go programs is that
there is no good evaluation function for go positions. Maybe go programs
should be constructed in a different manner, based on good and bad shapes.

Date: 24 Dec 2005 23:49:07
From:
Subject: Re: Go game lessons for chess players

Jerzy wrote:

>>in chess; maybe we should take
>>out the f3 pawn for white in chess
>>to compensate the white advantage
>
>
> Not f3 but f2 pawn.

yep,
just a little typo

>>and Go won't be solved by computers
>>as there's the handicap system
>>which currently is not fashionable

> And AFAIK it`s not the handicap that makes it difficult for go programs

well i was mixing winning from humans (easy, in tow centuries)
and solving Go for the first (computer) player (not so easy).
the latter won't be possible anyway with a handicap system
that was what i meant. a slightly crazy remark maybe,
for real Go players, but not from the perspective
of (computer)game theory

:)

Date: 25 Dec 2005 11:14:37
From: David Richerby
Subject: Re: Go game lessons for chess players

<superchess@wanadoo.nl > wrote:
> well i was mixing winning from humans (easy, in tow centuries) and
> solving Go for the first (computer) player (not so easy). the latter
> won't be possible anyway with a handicap system that was what i meant.

The handicap system doesn't make Go significantly harder to solve. In
practice, there are very few handicap starting positions -- are there any
with more than nine stones on the board? In practice, the effect of
multiplying by ten (or even ten thousand) the astronomically huge number
that represents the difficulty of solving Go starting from an empty board
is negligible.

Dave.

--
David Richerby Hungry Microsoft Toy (TM): it's like
www.chiark.greenend.org.uk/~davidr/ a fun child's toy that's really hard
to use but it'll eat you!

Date: 25 Dec 2005 15:55:40
From:
Subject: Re: Go game lessons for chess players

David Richerby wrote:
>
>
> The handicap system doesn't make Go significantly harder to solve.

yes, but the compensation of 0.5 stone (i think it's
called byomi or so) makes it more equal between black &
white; which is different in chess.

PS yes i know that you would call proving it's a draw
also is 'solving' the game, so Go might be easier
to solve than chess, as it most likely is a draw.

whereas chess *might* be a win for white..
:)

Date: 25 Dec 2005 18:19:58
From: David Richerby
Subject: Re: Go game lessons for chess players

<superchess@wanadoo.nl > wrote:
> David Richerby wrote:
>> The handicap system doesn't make Go significantly harder to solve.
>
> yes, but the compensation of 0.5 stone (i think it's
> called byomi or so) makes it more equal between black &
> white; which is different in chess.

The extra half point is to guarantee that the game cannot be drawn.

> PS yes i know that you would call proving it's a draw
> also is 'solving' the game, so Go might be easier
> to solve than chess, as it most likely is a draw.

Yes, proving that a game is drawn (or even won for the second
player) still counts as (weakly) solving the game.

Dave.

--
David Richerby Hilarious Boss (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ middle manager but it's a bundle
of laughs!

Date: 25 Dec 2005 23:21:15
From:
Subject: Re: Go game lessons for chess players

David Richerby wrote:
>
>
> The extra half point is to guarantee that the game cannot be drawn.
>
yes but if gives the first player a handicap,
which is not done in chess
>
>
> Yes, proving that a game is drawn (or even won for the second
> player) still counts as (weakly) solving the game.
>
no problem with that.

so Go is solved now ?
:)

Date: 26 Dec 2005 12:20:33
From: David Richerby
Subject: Re: Go game lessons for chess players

<superchess@wanadoo.nl > wrote:
> David Richerby wrote:
>> Yes, proving that a game is drawn (or even won for the second
>> player) still counts as (weakly) solving the game.
>>
> no problem with that.
>
> so Go is solved now ?

Alas, it does not suffice to prove that the game is either a draw or a win
for white or a win for black. You have to prove that it's a draw (not
possible in Go, as discussed) or prove that it's a win for white or prove
that it's a win for black. Pity, really, as that's rather harder. :-)

Dave.

--
David Richerby Accelerated Pants (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ well-tailored pair of trousers but
it's twice as fast!

Date: 25 Dec 2005 23:24:55
From:
Subject: Re: Go game lessons for chess players

superchess@wanadoo.nl wrote:
> David Richerby wrote:
>
>>
>>
>> The extra half point is to guarantee that the game cannot be drawn.
>>
let me explain a bit more.
on *average* white and black seem to have equal winning chances
in Go, even although one single game cannot be drawn.

now with perfect deterministics, this would mean
for Go that either white or black would win,
but in practice i believe the chances are equal.
far different than in chess

Date: 26 Dec 2005 12:48:33
From: David Richerby
Subject: Re: Go game lessons for chess players

<superchess@wanadoo.nl > wrote:
> superchess@wanadoo.nl wrote:
>> David Richerby wrote:
>>> The extra half point is to guarantee that the game cannot be drawn.
>>
> let me explain a bit more.
> on *average* white and black seem to have equal winning chances in Go,
> even although one single game cannot be drawn.

That's not actually the case. You seem to be confusing handicap and komi
(and the two aspects of komi). Allow me to explain.

In order to allow interesting games between players of unequal strength,
it is possible to play with handicap (similar to an odds game in chess).
Black places an agreed number of stones on the board in standard positions
as his first move and then the game continues as normal. The placement of
the handicap stones is given at

http://senseis.xmp.net/?path=Handicap&page=HandicapPlacement
http://senseis.xmp.net/?HandicapsBeyondNineStones

Somebody somewhere in rec.games.chess.* suggested a way of using this idea
to give a better handicap system to chess than odds games. Essentially,
the Go system is to let the first player make more than one move before
the second player starts (though these moves have to be in standard
places). To transfer this to chess, both players start with all their
pieces but white gets to make some agreed number of moves (no captures
allowed) before black's first move. After that, the players alternate as
usual. I think this is better than removing pieces from the board at the
start as that alters the character of the game too much and in unexpected
ways. (For example, IIRC Silman says that he finds it easier to give rook
odds than knight odds because, with rook odds, a strong player can often
checkmate before the rooks come into the game but the absence of a knight
is felt immediately.)

Anyway, back to Go. Komi exists to compensate the second player (white)
for the disadvantage of going second. Professional practice indicates
that black's advantage from the first move is about six points of
territory which doesn't sound very significant out of 381 intersections on
the board but actually is: most professional games are decided by the odd
point. Playing with n points of komi means that black has to end up with
at least n points more than white to win the game. For integer values of
n, this means that the game can be drawn so it is usual to play with
fractional komi -- in effect, this means that white wins `drawn' games.
Standard komi is 6.5 points. To avoid drawn games, playing `without' komi
tends to mean playing with komi of 0.5.

This can't really be applied to chess because the endgame changes
dramatically if draws are won for black: for example, many positions in
K+P vs K would be lost for white, even if he had the pawn. This seems
silly.

Dave.

--
David Richerby Natural Whisky (TM): it's like a
www.chiark.greenend.org.uk/~davidr/ single-malt whisky but it's completely
natural!

Date: 22 Dec 2005 00:58:47
From:
Subject: Re: Go game lessons for chess players

I am a middle level amateur in Chess, but I know a lot of former chess
masters in Russia (for example IM Alexandre Bitman) ,starting to play
Go instead of playing chess. Some of them are taking my lessons.

Date: 22 Dec 2005 17:27:22
From: Jerzy
Subject: Re: Go game lessons for chess players

>I am a middle level amateur in Chess, but I know a lot of former chess
> masters in Russia (for example IM Alexandre Bitman) ,starting to play
> Go instead of playing chess. Some of them are taking my lessons.

Therefore you should learn chess instead of spamming this group :D

Date: 21 Dec 2005 18:16:05
From: Jerzy
Subject: Re: Go game lessons for chess players

> Try this game! It's much more interesting than chess. Computers cannot
> play it at all!

What level have you obtained at chess to claim so ? Are you better than
Kasparov at chess ?