[Welcome to Sensei's Library!]

StartingPoints
ReferenceSection
About


Referenced by
Nakade
Definitionism

 

Nakade / Discussion
Sub-page of Nakade

Warp: I find the definition of nakade in the nakade page a bit confusing, or at least not very exact.

Nakade is defined there as "a play inside an eye space which prevents it becoming two (or more) eyes." By this definition alone one could think that the following is a nakade:

[Diagram]
Is this a nakade?



unkx80: Is this inside an eye space? I thought this looks like the boundary of the eye space.

Bill: The definition on eye space is essentially the same as that in [ext] Eyespace values in go by Howard Landman. Whoever wrote the definition on nakade meant the space inside an opponent's eye. I think we should go with Howard's definition. And Warp is right, the definition of nakade is flawed. (I have since modified it. More revision is probably necessary.)

unkx80: Thanks for the reference to Howard's definition on eye space. However the definition states:

An eyespace for a group X is a set E of connected intersections such that there exists a sequence of moves that results in at least one intersection of E being an eye of a group X', and the stones of X' include all the stones of X.

Since we are aiming at precision here, I would like to ask: Does it make all the points marked x in the following diagrams part of an eye space? Please correct me if I am wrong.

[Diagram]
Eye space?

[Diagram]
Eye space?



Bill: I think that they are part of White's eye space. But since the diagrams are open, it is hard to know what is not in White's eye space.

As for precision, Landman needs it for mathematical and logical reasons. Moi, I am more concerned with accuracy. Language is not all that precise, nor formal. I think that the definition on eye space is a good, informal definition that is faithful to Landman's idea. But nakade are not simply plays withing eye space, as defined there or in Landman's paper. (I see that eye shape is currently an alias for eye space, so I have edited the defenition of nakade again, to say that nakade occur inside the opponent's eye. Eye space is too broad a term.

In Japanese naka means "inside" or "middle", and de means "hand" (move). The name alone would indicate that a nakade is a play inside your opponent's group. However, there are more requisites than this. As far as I can see, there are at least four possible ways of understanding the meaning of nakade:

  1. Any play inside your opponent's territory (probably not correct, but could be deduced from the name).
  2. A play inside your opponent's territory which kills the group (this is probably a too loose definition because it also includes throw-in moves and moves like the one above which are different from nakade).
  3. A play inside your opponent's group which creates a shape such that the group dies.
  4. The shape itself.

The difference between definitions 3 and 4 is whether nakade is a move or a shape. That is, is nakade the move which creates the shape, or is it the shape itself which is called nakade?

[Diagram]
What exactly is the nakade here?

B1 kills the group by creating a certain shape which makes the white group dead (black can eventually atari the white group by making a bulky five shape). However, what exactly is nakade here? Is it B1, the shape formed by all the black stones inside the white group, or have all black moves inside the group been nakade?

(Some people define nakade as a move in the center of the group which kills the group, but clearly B1 is not in the center of any group.)

Charles I'd say that nakade was the special case of the vital point concept; when the vital point for two eyes was an internal point.

Bill: But what is the vital point in the next diagram? White has a choice among four plays.

Charles So, there can be miai or more. Should there be a 'definition'?


[Diagram]
White's nakade wins


A dead group might become alive later in the game (in the above example the white group might be able to connect to a live group later by pushing on the first line). If this is the case, is the move/shape still a nakade? Clearly a move inside an opponent's group which does not kill it is not a nakade.

Charles Disagree. For example it might create a seki.

Bill: Disagree, as well. (Agreeing with Charles again! :-)) It might lose a race to capture.

John F. My comments in Nakade Example 2/Discussion seem relevant here - maybe I misplaced them (one of the problems that would have infuriated me if I were - alas - still young enough to be infuriated - all this instant spaghetti of extended Talk Discussion Examples Lingusitics Giraffes pages).

Warp: The current definition of nakade in the nakade page is "A nakade is a play inside an opponent's eye". Is this really a complete definition? Shouldn't be necessary to give examples on how playing inside an opponent's eye achieves absolutely nothing (except giving your opponent one point in gote) or is only a sente threat (eg. in a ko-fight) but doesn't do anything else.

Bill: I think that is a question of pragmatics, Warp. A text may carry meaning based not simply upon the dictionary meanings of its words and phrases, but upon questions such as relevance and completeness, upon the expectations and intentions of author or speaker and reader or hearer. For instance, someone might say, "Black has a choice of two plays," meaning that there are only two potentially best points to play on in a position, totally ignoring dame or plays that lose a point, or, in some cases, obviously small plays. We do not have to alter the definition of the word play to take care of such texts. Pragmatics covers them. The plays you are talking about are, properly speaking, nakade, but who cares? In practice, cases where somebody will refer to them as such are rare or non-existent.

Warp: Whether or not nakade means technically and explicitly "any play anywhere inside opponent territory regardless of its consequences", I believe the main purpose of sensei's library is to teach, not to be a semantics reference manual. The definition of nakade should say what the term really means when people use it, not what it theoretically might mean if taken explicitly. I still maintain that "a play inside an opponent's eye" is wrong because it's not specific enough. When you say "nakade" you don't mean that; you mean a very specific type of play inside the opponent's eyespace. Do we really want newbies to get the wrong idea about go terms?

Bill: By that point of view, play is not specific enough. I think that the definition given of nakade is correct.

As for trying to define relatively plain go terms in terms of strategy or effectiveness, I have often seen that lead to confusion and inaccuracy because the amateur conception of the strategy involved was inadequate. Yes, such things should be taught, but definition is not the place for it, as a rule.

Good examples, I think, are heavy and light. Understanding the concepts is difficult, and not to be conveyed in a sentence or two, which is what a definition demands.

Really, the problems you find with nakade do not pertain to nakade, but to te (play). When players ask whether there is a play inside the opponent's enclosed space, they do not mean ko threats or plays that simply lose the stone or stones played.

(Later.) Warp, there has already been a fair amount of discussion on this and related pages. The definition in The Go Player's Almanac, which is more specific, saying that a nakade prevents two eyes, does not seem to apply to all examples in the literature, such as Nakade Example 3. See also John Fairbairn's remarks in Nakade Example 2/Discussion, where he refers to the Japanese authority, Hayashi. Hayashi gives two definitions, one quite general, as the one on the nakade page, and one specific to the 3 moku nakade, 4 moku nakade, etc. series. Maybe we should have two definitions, as well.

Warp: It feels a bit contradictory that the definition just says that nakade is a play inside your opponent territory, period, but then all the following examples related to nakade talk about killing. There's not even one single example of a a nakade which is not related to killing. The examples reflect exactly what the term nakade is always used for: To talk about about plays inside your opponent group killing it. I have never heard anyone talking about nakade with any other meaning; for example I have never heard anyone say "playing nakade at the end of the game costs you a point". However, I don't know how Japanese players talk about that: Do they use "nakade" when they say something like that?

Bill: They are more likely to say something like, "Te ni naranai." (That's not a play. = That does no good.) The important thing is that it does not work, not that it is a nakade.

I added a couple of seki examples. :-)


Question: Is the above seki a really a seki? Can't black play in the corner and recapture for life?

Bill: Do you mean this seki?

[Diagram]
Seki



[Diagram]
Seki

B3 at BC.

Black can play like this (called sending two returning one), but when he recaptures W2 he is just back to where he started, minus one net stone.



This is a copy of the living page "Nakade / Discussion" at Sensei's Library.
(OC) 2004 the Authors, published under the OpenContent License V1.0.