How to approach a life and death problem
When confronted with a life and death problem, whether in a problem book or in a real game, it is a good habit to work systematically through all the possible moves that may work.
|Table of contents|
Here is a systematic approach for killing groups:
- Seal in the group
- Reduce the eye space of the group under attack
- Look at the cuts
- Look at the vital point
Seal in the group
If the group still has access to the side or center, it has a possibility to link up to friendly forces. In short, one should seal in a group first.
Reduce the eye space of the group under attack
Look at the cuts
If there is a defect remaining in the group under attack, it may become a good deal easier to kill it if you manage to divide it into separate pieces. ("Divide and conquer".)
Placement at the vital point
The rationale is that the placement will mess up the shape of the group under attack. However if you start with this technique, the group under attack may still have options of running to the center or becoming a bigger group, i.e. a group with more space to create eyes. This is why you should first look at isolating it and reducing the eye space.
In order to live, one can reverse the above list
- Move out the group, often into the centre
- Expand the eye space of the group under attack
- Stay connected
- Play at the vital point
In real life, there is even more involved. There is a proverb 'play kikashi before living' (they may not be sente later). And there is a possibly conflicting piece of advice: live in a way that improves the position on ko threats. It is even possible that you deliberately live in a way that allows seki later, for the sake of other aji. This came up in a famous consultation game where Kitani and Go Seigen made up one of the teams. Where the status of a group depends on some aji on the outside - which is often the case - one cannot really just read the position once and leave it.
If there is a move which consistently refutes your attack, try that move as your first move.