Home > Steve > Go > Problem of the month

November 2002's Go Answer of the Month

Dia 1 - A game from Zagreb.

This is a game I played at Zagreb. Black is ahead by around 10 according to Many Faces of Go, even allowing for 8 komi and assuming the top black group is dead and the middle lower white group alive.

If you think the lower white group dies, which Many Faces does, it is a win by around 60.

In a typical fashion, I managed to lose this game against Thomas Scholz, 1k, Hannover .

What is going on and what should I have done?

Define the groups

In discussing this situation, I'll refer to the various groups and chains on the board using compass settings. So there are white groups in the NW, N, upper W, lower W, centre (C) to S, SW and W. There are black groups in the left N, right N which may join a group spawling to the E, also a C group which may join the SE and S sets of stones and lastly a group to the SW but cut of from the corner itself.

I suppose this month's question is really "how on earth did Black manage to lose this game"? MFGO believes black to be ahead by 60 points and reckons that white's C-S and lower E groups and black's left N group are all dead. Everything else is alive.

You can dispute NFGO's scoring - the likelihood of the C-S white group dying is small. Black may manage to live in the left N. Letting the white C-S group live changes the score to a black win by 17. Killing the black SW group really makes a difference - a white win by 54!

Review each group.

• The NE black group has enough shape to live.
• The white stone at K9 can be taken off easily, so the C black stones are connected. They also connect to the S stones either directly or by capturing K6. The southern boundary is secure, black M2 being ok y connection or capture of L2. Thus we have a strong C-S black 'wall'.
• The black SE stones are strong, having captured the 3 SE-most white stones.
• The black C-E stones separate the eastern borderand can guarantee to connect to either the C stones (capture M11) or to the NE group (play at O14). Therfore this group is also strong. This means that the SE white stones are all dead - at best they can just make one eye.
• The white upper E group is weak and potentially might have to sacrifice the two white stones on row 10 to live if black attacks. In fact this is unnecessary since there is a black weakness at R16 and throw-ins and placements and such-like all let white scrape a life. Thus this white group is ok.
• The white NW group is stong.
• The left N black group is weak and probably dead.
• The white N group has not yet fromed eyes, but can either do so or can connect out (whilst attacking the balck left N group).
• The C-S white group has no eyes at present, but can easily make one and can easily connect to the N white group. I therefore disagree with MFGO's analysis that these are dead. Assuming that they are alive the the black left N group is even weaker (deader?).
• The white SW group has enough space to make two eyes.
• The black SW group has one eye (the two white stones) but the second is not guaranteed without another move. The weakness at C9 makes D10 a stronger black play and should let these live happily.

The biggest undefined area is the C-upper W part. If this becomes neutral territory then black should maintain a lead. If the entire NW quarter of the board becomes white then it may be different.

So what went wrong?

Black makes an abortive attempt to live the left N group. In doing so he connects the right N to the C-S white groups, a strong combination. He aslo adds prisoners in white's teritory which grows whilst black's territory remains static.

Four moves after Dia 1, MFGO has white winning by 10.

Some clever footwork by white keeps black's left N dead and separated and still black tries to live. White makes territory galore on the W. Dia 2 shows the end of the recorded game with white in the lead.

Dia 2 - How not to do it!

White wins by around 23.

(There was no problem in December 2002)

Home > Steve > Go > Problem of the month
British Go Association
Last updated 2004-08-10