Lately my go has been getting worse rather than better, and I'm really not sure what I'm doing wrong. When I try to review I see tactical errors, but I those tend to be after I've already got myself in to a bad position and I can't see which moves brought that about. I decided to play and analyse a bunch of slow games to try and get myself back on track. This is the first of those. I've included some comments on where I think things went wrong.



Thanks for any help.

I made some comments, but don't trust them too much :p

Edit: By the way, free free to ask questions or point out mistakes

You had a decent game - maybe a won game up through move 122. At move 123 you made a classic mistake of attacking from a weak group. Just capture. Then you are strong. THEN attack.

_________________
Help make L19 more organized. Make an index: https://lifein19x19.com/viewtopic.php?f=14&t=5207

Oh, by the way, forgot to mention that I confused who you are for the sequence around 60, but opted to keep most of my original comments because the basic ideas are still there.

To me, move 34 was a huge mistake. Perhaps 35 too? White should play in the top left, as you suggest. How to continue seems like an exciting question. My first feeling's the following.

Click Here To Show Diagram Code

[go]$$Bcm35
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . X , . . . 1 . O . . 2 . . , . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 3 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O X X X . X . . . . . . . . O . . |
$$ | . X X X O O X . . . . . . . . . . . . |
$$ | . O O O X O O X X X O . . . X , O . . |
$$ | . . . . . . O O O X O . . . . . . . . |
$$ | . . . . . . . . . O . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

I agree with illluck (probably about everything) that 39 should be E10. Remember success for white isn't to live small in gote (or die big, as the case may be). A good result would be to limit/overconcentrate black and get a flexible shape. 40 here looks interesting to me (though it continues to seem hard for white to get a good result on the board after 33); none of the stones in and of themselves are important.

Click Here To Show Diagram Code

[go]$$Bcm39
$$ +---------------------------------------+
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . X . . . |
$$ | . . X , . . . . . O . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . X . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . 2 X . . . . . . . . . . . . . . . . |
$$ | . . O X . . . . . , . . . . . , . . . |
$$ | . . . O 1 . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . O X X X . X . . . . . . . . O . . |
$$ | . X X X O O X . . . . . . . . . . . . |
$$ | . O O O X O O X X X O . . . X , O . . |
$$ | . . . . . . O O O X O . . . . . . . . |
$$ | . . . . . . . . . O . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ +---------------------------------------+[/go]

_________________
Revisiting Go - Study Journal
My Programming Blog - About the evolution of my go bot.

Thanks everyone this is all really useful. Just a few questions.

Move 39: I just want to make sure I'm understanding the reasoning here. Is this bad because black's main goal here should be to build influence by attacking white's group (and white's is to stop that), and so the solid move is better than the more severe attack?

Move 49:



Are you saying there was a sequence for white to live after B7, or that B8 is a better way to kill?



Were you asking why I suggested Q9 or suggesting it yourself?

For move 39: The game move is not more severe - it gives white more cutting points to play around with. It's often better to play solid when you are trying to kill - your game move is more suitable if you don't want to kill and want to build outside.

Move 49: A cleaner kill probably - not sure if B7 kills, but it looks off to me just because it leaves more weaknesses and thus gives more chance for errors (the variation around move 60 would not have been an issue with B8).

Sorry, got confused there (reading numbers is hard, man XD). Yes, I was suggesting Q9 without realizing that was your suggestion as well, great minds think alike? :p

The exact opposite of what I thought... Orz

I'll take another look and see if I can see it. I'm actually pleased that I've now found something that seems to highlight one of my general problems rather than just a specific tactical error,

Thanks for all the help Illuck.

I'm about your rank, so don't take this as a teaching but:

I recently started thinking about those moments when I have to punish a too deep invasion in terms of "given stones". I concentrate on not making moves that allow the opponent to add free stones to the invasion; sublimated in preferring nobi to hane.

I have one chance to surround after each of my opponent moves, instead of once every two or three stones (the rest being responses to cuts and pokes).

In short, I now try to attack overvaluing aji and I'm getting good results.

The capturing race after move 150 may be worth some discussion. This sort of position should be pretty easy to read out, but maybe a few pointers will help.

First, do *not* try the brute force method of mentally imagining "I play here, he plays there, ...". That gets very confusing very quickly. Instead, try to count liberties directly, without worrying about the order of moves.

(First digression -- the order of moves does matter if either side has a way to make eyes or to increase or decrease liberties. In this position, W has some forcing moves at Q16, but actually making an eye there costs W more liberties than it gains. Similarly, B can play around L15 but cannot make an eye or gain a liberty. So this is a straight capturing race.)

(Second digression -- there are two shared interior liberties, so seki is conceivable. We can come back to that if necessary, but first let's see if B can actually capture outright.)

W clearly has 7 liberties. Counting B liberties is slightly trickier. In order to capture, B needs to play moves at N19 and N15, so imagine those stones are already on the board. Now how many liberties does B have? The answer is 7 liberties. It might look like only 6, but to play those 6 moves, W needs an extra approach move (K19 to approach at the top or L15 to approach at the side). Since it is B to play, he should be able to capture W, simply by filling liberties.

In the game, B went astray at move 153, which did not gain a liberty for B or take away a W liberty. Counting after move 155, we see that now W has 6 liberties and B also has 6 liberties. (A B stone at N19 leaves 5 physical liberties plus one approach move.) Since it is now W to play, W wins the capturing race.

Thanks Alguien, that sounds like a really useful way of thinking about it.



That is exactly what I tried to do and I my brain just turned to mush. Thanks for showing me the light.

By the way, You guys did help me stop my downward spiral; I've won all three games I played since this one.

Edit: I of course immediately went and lost a game after posting this.

One step forwards, two steps back.

That's why I don't evaluate my Go strength by the strength of my strongest moves but by how bad are my worst.

I am interested in this approach, as currently I pretty much use the brute force method for situations like this. You mentioned imagining N19 and N15 are already on the board. How exactly does this work? Can you always do this in a capturing race and still count the correct number of liberties?

Thanks so much,
Euphony

My understanding was:

Imagining those stones are already on the board when counting black's liberties is the same as not counting shared liberties. It makes it slightly easier to count because then you can more easily see how many liberties black has left, but since two liberties = two moves, it doesn't affect the calculation. We can see that white has 7 liberties including shared ones, and black has 7 excluding shared liberties, so if black goes first he wins.

That's interesting, Splatted. So one way of calculating a fight is to calculate the total liberties of the opponent's group, and the liberties of your group after all of the moves are made on the inside, and then compare.

That seems like a sound strategy. I'll have to try it out.

Thanks for the explanation.

Wait, that strategy doesn't work in general at all. Filling internal liberties not only takes a move, it also reduces the total liberties of the group. It works in this situation to assume those stones are there because there are two internal liberties, and filling those internal liberties actually adds a liberty to the group in both cases. One is a bamboo joint, so filling the internal liberty adds a liberty. The other adds a liberty because it is on the edge of the board, and the opponent must use an additional move after the hane.

I am going back to brute force.

I think you've misunderstood. You Should only count the internal moves as played when counting one side. In this game we counted white as having 2 liberties on the inside, but when counting black we didn't count any inner liberties. This means that it is irrelevant that playing a shared liberty takes one from black because we didn't count those anyway, and by counting them as white liberties we know that black has time to play those imaginary inner moves.

Also, white would only have to make one extra approach move because the final one would be a capture, and that's irrelevant to whether or not this works because if the inside plays didn't force an extra approach move we would have counted black as having one less liberty and seen that he couldn't capture white.

Edit: From re-reading your earlier post it seems you were doing it the way I described so maybe this post is useless...

Ok, I think I didn't apply it right in the game I played. I'll have to try again.