Reversible

Path: <= CGT path =>
  Difficulty: Advanced   Keywords: Theory

In CGT terms, a play reverses if the opponent's good response to it can return it to a position (CGT game) no better for the player than the original position. So (except for ko considerations), the player may as well continue, and often should do so.
Go players do not have a term for reversal, but they understand it intuitively.

For instance, the combination of hane and connection is so common that it has a term of its own: hanetsugi.

[Diagram]
Reversal  

If Black 1 reverses, then the position after White 2 should be no better for Black than the original position. We can show that with a difference game.

[Diagram]

Difference game


Can Black with the move win? If not, the hane reverses.

If Black simply connects on the left, White can play hanetsugi on the right and Black cannot win. So let's try a hane on the right.

[Diagram]

Black plays first

White 6 fills. The result is jigo. So Black 1 reverses.


Just for fun, let's see if White to play wins the difference game.

[Diagram]

White plays first

Yes. White wins by one point. Black can do no better.


By Bill Spight; moved by Charles Matthews from canonical form.


Charles It is a powerful general CGT technique, to isolate these plays, based on a clear general definition, of course. I think Bill is correct to say that the reversible move concept isn't something for which go players have a name.

Thinking about a translation, I came up with 'ungratifying', for the player who makes the reversible move. I play in such a way that my opponent has some reply to get a position at least as good as the starting position, from her point of view. It's as if the opponent can say "well, all right then, if you insist: if you play there I have only to play here and I don't see I have lost anything".

One example from common experience that occurs to me: playing a ko threat that gives the opponent a matching ko threat, which is pointless except in terms of the clock.

What you do with reversible moves is to make them into 'block plays'. It's as if, for example, Black can pick up two black stones and a white and place them down, announcing 'hanetsugi'. Black's one option of playing hane is replaced by (in general) many, in which the game is shortened. 'Cut to the chase', that kind of thing.

For example, in the matching ko threat case just mentioned, one could replace the threat play by 'blocks' of seven (threat-answer-take ko-counterthreat-answer-retake ko-whatever next).

Another case is playing in true miai - for example a pair of endgame plays that are a perfect match in size.

There are various bits of go wisdom about three-move sequences, such as the '1-2-3 principle' behind cross-cut then extend and so on. Here the CGT idea corresponds to some less-formulated feelings about go. It would certainly be nice to understand canonical form somewhat better, with a suitable combination of 'accept no absolute loss' and 'avoid redundant plays' as the primary strategy concepts.

It does seem to me that what I wrote about the 'two unacceptables' over at middlegame joseki could tentatively be related.


Kupopo: I'm probably way out of my league in posting to this particular discussion, but after reading through a bunch of pages on CGT, I'm still at a loss as to why Black 1 doesn't gain him anything. If you simply count the surrounded territory at the end, Black has one more point than White, whereas before playing they were equal. I know I'm not looking at it mathematically, but to me it seems like a good move. Also, I'm very unclear as to why Black 1 in the first diagram is a 'reverse', since I can't seem to find a clear definition anywhere. Am I correct in guessing that a reverse is a move which forces a response and causes the game to end up in an equal-valued situation?

Bill: The point of reversible moves is that the result after the opponent replies is no better than the original position. Therefore, unless you are making a ko threat, once you make a reversible play you might as well continue.
In this example, the question is not whether B 1 gains anything, but whether Black has gained anything after B 1 - W 2, (without playing B 3). In fact, on average Black has lost 1/6 point after W 2.

[Diagram]
A loss for Black  

The original position has a count of 0, of course. Suppose that the two marked stones have been played. If Black plays at 1, the local score is +1 (for Black). W 1 moves to a position worth -1 1/3, on average.
Later, if White plays at 2, the local score is -2. If Black plays at 2, the local count is -2/3 (White has captured a stone, but must count the stone at 1 as 1/3 point for Black. All of these plays are gote.

Detailed discussion at reversible play - loss and gain.


Path: <= CGT path =>
Reversible last edited by MrTenuki on July 9, 2006 - 01:47
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
RecentChanges
StartingPoints
About
RandomPage
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Goproblems.com
Login / Prefs
Tools
Sensei's Library