![]() StartingPoints Paths Referenced by
|
One-Two-Three
Path: SequencingQuestions · Prev: OrderOfPlay · Next: KoThreatPlayingOrder
Difficulty: Dan level
Keywords: Tactics
There are a few ideas about three-move sequences that come up quite frequently. 123 principleDon't play Black 1, White 2, Black 3: just play Black 3. For example, in atari-connect combinations, the atari should often be omitted and only the move that would connect should be played. Confused: As a rather weak player, I have trouble understanding where this proverb should be applied and why. Does this proverb cover the following situation?
Charles This hanetsugi is actually a case discussed on the reversible page. I've put a comment lower on this page. Well, this page is marked 'dan-level' and there's a reason for that. These ideas have more holes in them than most proverbs, if you insist on them as rules. The discussion here, to be full enough to be useful, ought to include the crosscut then extend stuff (see discussion pages linked from there); because that's really just an important special case. But that hasn't been brought to a satisfactory conclusion, in my view. What seems useful to say is: The '123 principle' is a small-scale version of a key proverb beware of going back to patch up. An example from a recent article I wrote for the American Go Journal.
Here White must get out to the centre. But it is better simply to play White 1, than to play atari on the marked black stone.
White 1 here is poor style. White has lost the chance of a ladder with White at 2, and still must play 3.
Black 2 here looks like a possible mistake of the same type. Omitting Black 2 and simply playing 4 is normally better style in attacking.
Using the tewari technique, we can ask whether Black would really play Black 1 here? There seem to be many better points, such as a to d. A more involved case of the same idea is: ABC principleIf Black has the choice of Black a, White b, Black c or Black b, White a, Black c, perhaps Black should simply play c.
This staircase shape is a reasonable example of the ABC principle. Black 1 will often be correct, rather than Black's atari play at either of a or b. This is one aspect of: Don't play out miaiIn the absence of a good reason, true miai points should probably not be played out, as an unmotivated exchange a-b. Some more related ideas: Forced answer adviceIf you play a which you expect the opponent to answer at b, treating a as a forcing move, you should already know your follow-up play c. (From Tokimoto Hajime 8 dan.) James Kerwin on urgent plays:Treat a play at c as urgent if the opponent's play at b otherwise puts your earlier play at a at risk of being made meaningless. (Noted on play urgent moves before big moves.) Theory of reversible playsFrom CGT there is the quite profound idea of a reversible play. It again relates to thinking about a three-move 'block': Black a gives White an answer b which (provably) gives a position at least as good as the initial one, so Black ought to have the next play c lined up.
In this example the point is that Black has gained nothing yet, if we're just talking endgame. Simply playing Black 1 and White 2 isn't typically kikashi - effective forcing play - because White a is now better than before. Dieter: Can we say that this subject falls under the general subject of order of play ? Charles There is certainly a major topic you could call sequencing questions. Kupopo: I'm certainly out of my league in this discussion, but according to the example at the bottom of canonical form, the sequence in the above example must continue with Black 3 at a and White 4 at one of several points inside his own territory, preventing a white invasion. Regardless of whether or not White a is better then before, it's Black's turn to play, and he may have to give up sente (but probably not) to take that point, thus stealing one or two points of territory from White. Am I missing something? Charles Yes. The point is not whether Black may gain something by this play (he may); but whether Black has already gained something by getting White to answer. Path: SequencingQuestions · Prev: OrderOfPlay · Next: KoThreatPlayingOrder This is a copy of the living page "One-Two-Three" at Sensei's Library. ![]() |