I've been learning to play Go for the past few days, and in this game I'm playing I don't understand exactly how I captured the whole board in one move. These images show what happened.

enter image description here enter image description here

  • 4
    Looks like the AI could tell there's no way for white to defend that last group. To see, just play it out yourself. (which is what players do anyhow when there's no agreement as to who would win a certain area) – L. Scott Johnson Nov 1 '20 at 18:23
  • I strongly suggest you increase the board size to at least 11 x 11 instead of 9 x 9. The AI you're playing against is making terrible first moves in the corners, so clearly thinks the board is bigger than it is. even an 11 x 11 board has all edge and corners, so a first move should never be further out than the 3-3 or 3-4 point. Allowing an opposing corner invasion at 3-3 on these small board sizes is suicide. – Forget I was ever here Nov 1 '20 at 22:27
  • 1
    Even if this move does kill the white group you arguably only captured the rest of the board (about a third) rather than the entire board with that move, as you had apparently already captured the other c. two thirds before you moved. This may seem like nitpicking, but it is important to think about how much difference each move makes. – PJTraill Nov 2 '20 at 10:36

A corner invasion can in some cases can make a huge difference. In this case the play on th 3-3 point threatens to cut the white group in the upper left into three pieces, two of which have no chance at life and the third is in bad shape.

That siad, I think the AI is a bit premature in declaring all these stones dead. A play by white to connect at the 4-3 point or to undercut at the 3-2 point could lead to life for at least the largest part of the white group, depending on the further line of play by both players. I am not an expert (I was about 10 Kyu at my best, years ago), but I would not concede with the white position after this move,

I agree with the comment that a larger board size may avoids tactical oddities with this AI.

Addition: A comment by @ilkkachu points out that in the 2nd image it is black's turn, not white's, Given that, white doesn't seem to have much chance. If as the comment suggests black extends to the left of the previous store, ther does not seem a plausible line for any white group to form eyes, assuming reasonable play by black. I had incorrectly read the 2nd image as white to play, that makes a sizable difference.

  • In the second image, white is shown to already have answered below the black stone. After that, the only move is that black extends to the right, taking either the lone white stone in the middle or the chain to the lower left. Either way, white has no space to make eyes anymore. Had white answered to the right of the black stone (the other 4-3 point), there might be something, but that's not what the image shows. – ilkkachu Nov 2 '20 at 21:56
  • @ilkkachu see edited answer above. – David Siegel Nov 2 '20 at 22:47
  • Yep, thanks. And yes, the image is a bit odd. I don't think it makes much sense to mark stones as dead before both have passed, and you wouldn't do that there... – ilkkachu Nov 3 '20 at 10:09
  • In the second image, the strongest attack for black seems to be the 3-4 atari on the single stone: This gives the invading stones a third liberty while the big white group has only two, as such, that white group can be considered dead. And the remaining white stones do not have enough eye space to make a living group. As such, I think the AIs conclusion is perfectly ok. – cmaster - reinstate monica Dec 1 '20 at 12:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.