Go problem: linking up stones

alisbarbe · Post by **alisbarbe** » Wed Jan 07, 2015 10:20 am

Hello,

I'm reading Graded Go Problems for beginners, Volume 2 and encountered a problem for whichs I don't understand the proposed answer.

The author says in his introduction to try and refute the correct answer until you find you can't refute it and thus accept the correct answer. I've been trying to do just that but I always seem to be able to refute his answer. As I'm not good at Go, I suppose there's something that I don't understand.

The question in the book is: What should black do to link the stones at the top with the stones at the bottom. The correct answer from the book is in the sgf below. The comment with the answer says that there is no way that qhite can prevent black from connecting:

If I continue to try and link the top black stones with the lower black stones I always seem to find a way to block with White, at least in the first couple of moves.
After that it depends on which variant White uses:
- I can then take White's blocking stones and connect the black groups or,
- I can't take White, he escapes and I can't connect the black groups.
When I can link the black stones this is after about 20 steps, but that seems a lot to me considering the other solutions in the book. Usually there is a firm answer within just a couple of moves.

Maybe I shouldn't understand 'link' as a true physical link but rather the potential link? Or might there be a more elegant solution, meaning can you always link/connect the two black groups as stated in the answer?

Thanks a lot!

skydyr · Post by **skydyr** » Wed Jan 07, 2015 10:52 am

I don't see the SGF, and when I try to view it directly, it says that the attachment no longer exists. Could you repost it?

alisbarbe · Post by **alisbarbe** » Wed Jan 07, 2015 11:52 am

problem74.sgf: (288 Bytes) Downloaded 786 times

Hopefully it is there now, in the preview it seems okay.

Uberdude · Post by **Uberdude** » Wed Jan 07, 2015 12:19 pm

Where did you manage to create a twenty move sequence?

schawipp · Post by **schawipp** » Wed Jan 07, 2015 12:34 pm

I don't see a problem. When e.g. white plays B14 black simply answers B13. If white then connects via C14, black connects at A14. On the top side its even more simple.

Click Here To Show Diagram Code: [go]$$Wcm1 $$ ----------- $$ |.......... $$ |.8........ $$ |.76X.O.... $$ |.X5X...... $$ |.XOO.O.... $$ |413O.O.... $$ |.2XXX..... $$ |.......... $$ |.......... $$ |..........[/go]

If white doesn't connect via C14 its also simple:

Click Here To Show Diagram Code: [go]$$Wcm1 $$ ----------- $$ |.......... $$ |.......... $$ |...X.O.... $$ |.X.X...... $$ |.XOO.O.... $$ |314O.O.... $$ |.2XXX..... $$ |.......... $$ |.......... $$ |..........[/go]

Could you post a variation where you think white can escape and block the connection?

amatterof · Post by **amatterof** » Wed Jan 07, 2015 12:55 pm

schawipp wrote:When e.g. white plays B14 black simply answers B13.

I think Black can resist even more strongly, by playing at C14.

Click Here To Show Diagram Code: [go]$$Wcm1 $$ ----------- $$ |.......... $$ |.8........ $$ |.76X.O.... $$ |.X5X...... $$ |.XOO.O.... $$ |.12O.O.... $$ |.3XXX..... $$ |.4........ $$ |.......... $$ |..........[/go]

The White stones ( :w1:

and

) cannot escape.

alisbarbe · Post by **alisbarbe** » Wed Jan 07, 2015 4:44 pm

Thank you for your answers.
My '20-move' solution is because I want Black to physically connect and also White is maybe a little exagerating trying to stop Black from doing just that. Maybe White wouldn't do this in 'real' life?
In my solution for White to escape, Black made a mistake.
So I suppose these are part of my problem, how good are White and Black in these problems? Do they make mistakes? This kind of thinking could lead to absurd situations, like Black or White playing somewhere completely irrelevant (for the tsumego that has to be solved) and we don't want that, but how far should I go?

So my total connection solutions are below with around 20 moves to get total connection if White is 'throwing bad stones after good ones' . Also in the first one White escapes in the variation.

problem74_solution-White-B14_Black-B13_White-escapes_variation-Black-wins22moves.sgf: (377 Bytes) Downloaded 813 times

problem74_solution-White-B14_Black-B13.sgf: (297 Bytes) Downloaded 815 times

problem74_solution-Wb14-b13_Bc14-c12.sgf: (384 Bytes) Downloaded 785 times

If I make White smarter and still try to make a total connection, the number of moves is only 12.

Any suggestions?

Thanks in advance!

EdLee · Post by **EdLee** » Wed Jan 07, 2015 4:49 pm

alisbarbe wrote:how good are White and Black in these problems? Do they make mistakes?

Hi alisbarbe, good question. We try to find the absolute best moves for both sides.

Uberdude · Post by **Uberdude** » Wed Jan 07, 2015 5:48 pm

alisbarbe wrote:My '20-move' solution is because I want Black to physically connect and also White is maybe a little exagerating trying to stop Black from doing just that. Maybe White wouldn't do this in 'real' life?
In my solution for White to escape, Black made a mistake.

In go when we talk about connection like this we don't usually mean connecting to form a single chain of stones that all share the same liberties, but to connect stones such that, with alternating play you can, by playing good moves, always maintain that connection even if your opponent tries to cut you. If you choose to play bad moves obviously you won't do as well as you could.

It seems to me you need to practice capturing a stone in atari on the 2nd line. If you are confident in that, and how things can change as liberties of the surrounding stones are reduced, then you will see you can answer b14 at c14 and white's stone is simply dead. Take a look at these problems: viewtopic.php?p=172664#p172664.

Bill Spight · Post by **Bill Spight** » Wed Jan 07, 2015 6:46 pm

alisbarbe wrote:Thank you for your answers.
My '20-move' solution is because I want Black to physically connect and also White is maybe a little exagerating trying to stop Black from doing just that. Maybe White wouldn't do this in 'real' life?
In my solution for White to escape, Black made a mistake.
So I suppose these are part of my problem, how good are White and Black in these problems? Do they make mistakes? This kind of thinking could lead to absurd situations, like Black or White playing somewhere completely irrelevant (for the tsumego that has to be solved) and we don't want that, but how far should I go?

The player to play and succeed (the first player) has to play perfectly. The other player (the second player) can play anything, and the first player has to be able to succeed, no matter what the second player plays. Since the first player has to be able to succeed against perfect play, we can consider both players to be perfect, as Ed Lee says.

In most situations we do not explore all the possible plays for the second player, trusting that the first player can succeed unless the second player makes certain plays. Sometimes that assumption is wrong. :-|

This particular problem is one that I learned when I was around 3 kyu. I do not think of it as a beginner problem. For one thing, the possible number of plays for the second player is quite large. It takes some experience, I think, to be able to prune the possibilities. I have looked at a few of your variations, and I salute your thoroughness. :salute:

Take nothing for granted. But I think that this problem was probably too hard for you right now.

Uberdude · Post by **Uberdude** » Thu Jan 08, 2015 3:25 am

schawipp wrote:I don't see a problem. When e.g. white plays B14 black simply answers B13. If white then connects via C14, black connects at A14. On the top side its even more simple.

$$Wcm1
$$ -----------
$$ |..........
$$ |.8........
$$ |.76X.O....
$$ |.X5X......
$$ |.XOO.O....
$$ |413O.O....
$$ |.2XXX.....
$$ |..........
$$ |..........
$$ |..........

Click Here To Show Diagram Code
[go]$$Wcm1 $$ ----------- $$ |.......... $$ |.8........ $$ |.76X.O.... $$ |.X5X...... $$ |.XOO.O.... $$ |413O.O.... $$ |.2XXX..... $$ |.......... $$ |.......... $$ |..........[/go]

schawipp, seems you need to practice capturing stones on the second line too

You should cut at c14, your falling back to b13 has this problem:

Click Here To Show Diagram Code: [go]$$Wcm1 $$ ----------- $$ |.......... $$ |.6........ $$ |854X.O.... $$ |.X3X...... $$ |7XOO.O.... $$ |.19O.O.... $$ |.2XXX..... $$ |.......... $$ |.......... $$ |..........[/go]

Now black could actually play 6 at 9, and then white 6 doesn't quite work, but with a little more support on the top side around e18 it could. Cutting at c14 just works.

schawipp · Post by **schawipp** » Thu Jan 08, 2015 2:56 pm

Uberdude wrote:schawipp, seems you need to practice capturing stones on the second line too You should cut at c14, your falling back to b13 has this problem:

Yes, you hit the point. As a former chessplayer I initially was too much focused on killing/rescuing even single stones. Lately I am sometimes too relaxed on such stones and then end up losing games by 1.5 points ;-)

alisbarbe · Post by **alisbarbe** » Thu Jan 08, 2015 3:31 pm

Thanks for all your help.

I will remember to try and play Black and White as good as possible & not always play till the end (stones that are clearly dead, groups that are alive).

@Bill, thanks for your compliments but I must say that my thoroughness was spiked by others in this thread who asked me to elaborate my questions & who also gave me a base to work on (Uberdude, shawipp, amatterof) ;-)

Life In 19x19

Go problem: linking up stones

Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones

Re: Go problem: linking up stones