Forum for Nakade
WME [#131]
I'd like to WME this page, after suitable discussion, and noting John F.'s comments on Nakade Example 2/Discussion. The basic approach would be to position "nakade" as having three and a half, closely related meanings:
0.5. The base/literal meaning: a move played inside.
1. A move played inside one of the standard killable spaces.
2. By extension, one of the spaces that permits such moves.
3. Further by extension, a play that results in a shape that when captured leaves one of the spaces that permits such moves.
I'd also like to sort out related English terminology (killable, unsettled, pyramid four etc.) and quixotically propose some new ones, such "killable n-pointer".
Thoughts?
-- Bob
I say go for it, Bob. :-)
I definitely think we need two definitions, I don't know about more. But we do need to talk about the phrase, n-moku nakade. As John says in his comment, that's best regarded as referring to the number of inside moves before (possible) capture.
![[Diagram]](diagrams/39/e728c1013ba6e79aebc59b685e89f2f1.png)
Hmm. I guess we better get this straight before starting. I also noted John's comment:
The second Hayashi meaning is the special X-moku nakade phrase. This is the phrase that has tripped up westerners so often, because they interpret it as a shape. The way to understand the original phrase is (a) to remember that Japanese does not usually distinguish singular and plural and (b) to treat moku as a counter. Gomoku nakade can then best be understood as "five inside moves". By association, of course, such phrases are closely bound with life-and-death and the "bulky" shapes, otherwise there's not much point in them. But it's an association, not a definition.
John seems to be making two distinctions; the first is between a "shape" and a "count". But I'm having a hard time understanding this. I've always understood that the "n" in "n-moku nakade" refers to the number of points in the shape (regardless of whether or not they are filled), not the number of moves required to capture.
Trying to parse the second distinction, between association and definition, if something is 100% associated, then it is pretty much a definition. So I guess he's saying that the association is not 100%, that there are non-bulky-five shapes that could also, in some abstract sense, be called "5-moku nakade"--such as five points in a straight line. Is that it? But I've never seen such an odd use of 5-moku nakade, which is not surprising, since it seems nearly useless. In addition, my interpretation of the "naka" part of "nakade" is that the emphasis is on "center", not just "inside"--in other words, the (single) center point in all these shapes. Perhaps I'm missing something here.
With regard to your comment, if the "n" referred to the number of moves required to capture, as you say, then the "n" would change with every move played, and the position right before the final capturing move would be called "1-moku nakade", a word which doesn't even exist. This would also mean there would be no unique Japanese term for the generic straight three shape independent of its emptiness or partially filled status.
Of course, *if* the 5-moku nakade (bulky five) shape is completely empty, and the opponent starts filling it one point at a time, and you don't respond, then it will also take five moves to capture it. But it always takes n moves to fill n points, so what is the point of the distinction? In practice, as in a capturing race, the opponent will take after the first four moves (three net moves); then take after the next three moves (two net moves); then three more moves will be required to capture, for a total of eight.
The "moves" interpretation would also seem to be inconsistent with the normal meaning of the word "moku", which refers to intersections or points. If the intent were to talk about moves, one would expect the term to be "5-te nakade".
![[Diagram]](diagrams/21/18aea1bc2fde28b681964dfff27624f5.png)
The source of Nakade Example 3 says, 隅 は “４目ナカデ” だから、... (Since the corner is 4-moku nakade, ...)
Plainly, it is not referring to remaining moves, since three of the moves have already been made. Also obviously, the fourth move must make a square shape, or White lives.
So the four nakade, at least for four or more, are not just any four inside moves, but four moves that make one of a number of particular shapes.
![[Diagram]](diagrams/29/6716197013501bbeda2c17340ebb33fe.png)
Later the commentary says that here White's shape has become 2-move nakade.
Plainly, if there are already nakade, the number of nakade referred to is not the number of points inside the eye, but a smaller number, usually that number minus one. Does that mean that for an empty eye it is the number of points in the eye? Or is it also a smaller number?
And is the shape itself nakade? Or does nakade modify shape? (The latter, I think.)
Interesting, thanks for pointing this out.
Before getting into nakade, what term is best for these "spaces" in the context of the nakade discussion? We have "eye", "eyeshape", and "big eye" (which seems to a made-up English term).
1. "Eyeshape" has some problems. First, the "shape" part might make one think of stones, whereas we are dealing with empty points. Second, at least in the Japanese usage for the semi-equivalent "gankei", it has a much broader meaning--we could say of a group that has lots of alternatives for making eyes that it has "ample eyeshape" (this meaning is missing from the SL definition of eyeshape; does anyone use the word "eye potential"?).
2. "Eye" is OK, but includes one-point eyes, which are not relevant to nakade, which is all about making larger spaces into multiple eyes or preventing the enemy from doing that, or filling in eye spaces in a capturing race.
3. "Big eye" would seem to be the right idea, although the current SL definition limits its usage to capturing races (why?), as well as stating that big eyes must be at least four points (why again?). Of course, there's another aspect to eyes, at least as related to nakade, which is that there is only one of them. Therefore, in some cases it might be good to clarify with "one big eye".
Provisionally we'll go with "big eye" in a slightly modified sense, namely any single eye greater than two points.
If the big eye is killable, then we call it a "killable big eye". This seems better than killable eyeshapes, which is used elsewhere on SL (although, oddly, that page also includes shapes that are already dead), but could be applied to just about any life-and-death problem. Note that the killability distinction is not made in Japanese, where it's just called nakade.
Is the move played to kill, or live, also "nakade"? After all, the word "te" occurs right there in "nakade". But my interpretation for this is that "nakade" is actually a kind of abbreviation for "[a shape which is vulnerable to] a move inside (in the center of) a big eye." In Japanese you can say "nakade ni utsu" (play in/at the nakade), which might lead one to think that nakade could also be the move itself, but this Japanese could just as easily be interpreted to mean "play in the big eye". To avoid confusion with the form of the eye, for the move itself, "center of the killable big eye" probably works.
With regard to this meaning, putting "n-moku" on the front to get "n-moku nakade" (killable n-point big eye) obviously refers to the specific shapes of such killable big eyes. For this, of course, we already have the various English terms, although it might also nice to tune the terminology to emphasize their commonality: straight-three big eye, bent-three big eye, etc.
What you're pointing out, I think, is a distinct usage in the context of a capturing race. It still refers to a big eye, but now including even ones which are dead locally (such as clumped four). In this usage, the key point is how many moves are required to capture, given the number of internal points. In such cases, the term "big eye" could be used for clumped four or even for two-point big eyes, as you point out in Nakade Example 3. Since the whole point is the number of moves to capture, not killability, "nakade" by itself would not be used in such a case, only "n-moku nakade", which in this case we can simply call "n-point big eye". (See ![[ext]](images/extlink.gif)
http://www2.tokai.or.jp/tuta/semeai3.html for a discussion of nakade in the context of races to capture.)
To be precise. let's break this down, working backwards:
1) Black plays a move within a White big eye, either:
a) killing it, since it's one of the killable types of big eye
and/or
b) filling in a liberty necessary to win a capturing race.
2) Before Black played that move, White had a particular shape of big eye, either killable or dead (in which case Black would play only for reason 1b). That big eye contained "n" points.
3) White may have ended up with that big eye by a sequence where Black had played inside her group in such a way as to leave that big eye after White eventually had to capture.
Then:
1a) (the killing move) is not "nakade", but rather "center of the big eye" (in spite of what the SL nakade page says.
1b) (the filling move) is simply "playing in the big eye".
2) is nakade, a "big eye", or a "killable big eye".
3) can be glossed as "creating a killable big eye" (Japanese: gomoku-nakade ni suru), or "play for a killable big eye".
Before getting into nakade, what term is best for these "spaces" in the context of the nakade discussion? We have "eye", "eyeshape", and "big eye" (which seems to a made-up English term).
1. "Eyeshape" has some problems. First, the "shape" part might make one think of stones, whereas we are dealing with empty points. Second, at least in the Japanese usage for the semi-equivalent "gankei", it has a much broader meaning--we could say of a group that has lots of alternatives for making eyes that it has "ample eyeshape" (this meaning is missing from the SL definition of eyeshape; does anyone use the word "eye potential"?).
I agree that we should avoid eyeshape.
2. "Eye" is OK, but includes one-point eyes, which are not relevant to nakade, which is all about making larger spaces into multiple eyes or preventing the enemy from doing that, or filling in eye spaces in a capturing race.
I like eye for the second definition of nakade, but it needs to be qualified in terms of size and shape. As pointed out on http://www2.tokai.or.jp/tuta/semeai3.html , the shape of a nakade can have defects. Well, open points cannot have defects. Eyes can.
3. "Big eye" would seem to be the right idea, although the current SL definition limits its usage to capturing races (why?), as well as stating that big eyes must be at least four points (why again?). Of course, there's another aspect to eyes, at least as related to nakade, which is that there is only one of them. Therefore, in some cases it might be good to clarify with "one big eye".
Provisionally we'll go with "big eye" in a slightly modified sense, namely any single eye greater than two points.
Well, there are big nakade and small nakade, including 2-point nakade. The big-small distinction occurs between 3 points and 4 points because 4-point and larger eyes are qualitatively different in terms of semeai. Not only do they have extra liberties than the number of open points, in a big eye vs. small eye semeai the shared liberties belong to the big eye (as a rule), as in me ari - me nashi.
So I think that defining nakade with the phrase, big eye, would be confusing.
Personally I think "big eye" works quite well, and the distinction between smaller and bigger ones could be made by saying, ummm, "smaller big eyes" and "bigger big eyes" (or even "very big eyes" for those containing more than seven points). But if you don't like "big eye" for this purpose, are you proposing using "nakade"? Or "eyespace"? Or any other ideas, if we are in neologizing mode, which perhaps should be avoided? I think the basic definition is "a single, completely surrounded empty space of two or more points".
From "An overview of go" (囲碁大観）by Shimamura 8-dan (島村 利博) and Noue (?) 5-dan (野上 彰), pp. 102-3.
眼あり眼なしの攻合に移るがその前にナカテの数を調べておく必要がある。
Rough translation: We'll be shifting the discussion to me-ari me-nashi semeai, but before then it is necessary to investigate the number of "nakate".
(Note the t in te.)
They go on to show the 4-point square eye and say that at first glance it seems to have four moves, but there are really five.
They continue:
四目以上の地を持つ眼を大ナカテと言う。
Eyes with four or more points of territory are called "big nakate".
They go on to give the formula for the number of "nakate": 3-3, 4-5, 5-8, 6-12, 7-17.
すなわち三目ナカテは三手、四目ナカテは五手、...
That is, "sanmoku nakate" has three moves, "yonmoku nakate" has 5 moves, ...
Note the looseness of the language. At first "nakate" seems to refer to the moves (liberties), but quickly shifts to the eyes. I think that the difference between nakate and nakade is simply one of pronunciation, and that this reference justifies a definition of nakade as an eye of a certain size and shape.
And maybe eye rather than eye shape, because on p. 105 they talk about "big nakate - small nakate" semeai for "big eye - little eye" semeai.
Quick note: Searching with Infoseek Japan I have found several sites that discuss nakade. This one in particular refers to a specific shape (形) with 5 empty points as a 5-moku nakade.
In this web page from his Tesuji Bible Segoe is more careful with his language than Shimamura and Noue. He gives the formula for nakade liberties as（ｎ2−３ｎ＋６）／２＝ナカデの手数 . I. e., the formula tells us the number of nakade plays. He refers to n as the number of points in the eye, and says, ５目地は８手 . (5-point territory instead of 5-point nakade.)
With regard to your comment, if the "n" referred to the number of moves required to capture, as you say, then the "n" would change with every move played, and the position right before the final capturing move would be called "1-moku nakade", a word which doesn't even exist. This would also mean there would be no unique Japanese term for the generic straight three shape independent of its emptiness or partially filled status.
Of course, *if* the 5-moku nakade (bulky five) shape is completely empty, and the opponent starts filling it one point at a time, and you don't respond, then it will also take five moves to capture it. But it always takes n moves to fill n points, so what is the point of the distinction? In practice, as in a capturing race, the opponent will take after the first four moves (three net moves); then take after the next three moves (two net moves); then three more moves will be required to capture, for a total of eight.
The "moves" interpretation would also seem to be inconsistent with the normal meaning of the word "moku", which refers to intersections or points. If the intent were to talk about moves, one would expect the term to be "5-te nakade".
As my recent posts indicate, I have come to agree with you that moku refers to the points inside the eye.
I now speculate that the phrase, 5-moku nakade, originally meant the plays inside certain 5-point eyes, but eventually came to refer to the eyes themselves.
Bob,
After looking through my books and searching the web, I think we need at least two, maybe three definitions. (My own thinking is to try to have a few general definitions and to deal with tendencies and frequencies of use with "especially" or "typically".) Anyway, here are three cases we need to cover, at a minimum.
Nakade definition 1:
![[Diagram]](diagrams/38/acb2f13a8647bed670c551178e4a53ed.png)
, 
, 
, and 
are nakade, especially ![[Diagram]](diagrams/41/b159026063ffa67c749d40bfe5e9bad1.png)
![[Diagram]](diagrams/1/db5ddfa39639681d449a8cef6d8bfdc4.png)
I think this Kansai Kiin site gives a very good introduction to nakade, illustrating different uses of the term in a life and death context.
I have some suggestions for the definition section. I have come up with <gasp> six definitions, three for nakade and three for n-point nakade.
Nakade, literally inside play, is a Japanese go term used in the contexts of semeai and life and death. It can refer to both plays and eyes.
Definition:
1. A play inside an opponent's eye.
Especially, b. Such a play that prevents a second eye.
2. A large eye that is vulnerable to an inside play.
N-point nakade is a phrase that refers to specific plays and eyes. N refers to the number of points inside the eye and typically lies in the range of 2 to 7.
Definition:
1. A play inside a perfect eye (one with no defect) with N points that prevents a second eye.
2. A perfect eye with N points such that there is an opponent's play inside it that prevents a second eye or takes away a liberty in a semeai.
3. An eye that will become an N-point nakade or whose stones are dead because it would become such if the position were played out.
Obviously, there are things to be fleshed out and examples to give.
Bob, I hope you like large eye.
Yes:
My preference for such a page would be to have translation as a first goal, and explain the concepts on separate pages. We already have pages like killable eyeshapes, almost fill, basic live shapes etc ...
So strictly speaking, I could do with your list of meanings in as much as they refer to other pages. But, ...
- I know many/most people prefer the Japanese terms to contain the basic diagrammage.
- Even if I like English terms best (nowadays), I still think that technical pages without diagrams are awkward.
I'd like to propose the following page structure for "nakade", based on this discussion.
1. Nakade: introduce as Japanese word for "big eye". "Nakate" as lesser-used alternative. Generic meaning. Etymology etc.
2. BigEye: main page; describe in general terms. Three primary branches: BigEyeStatus?, CenterOfTheBigEye?, and BigEyesinCapturingRaces?. Point to existing TableOfEyeSpaces?, which should be given a mini-WME and possibly renamed to TableOfBigEyes?.
3. BigEyeStatus?: discusses big eyes from the standpoint of their local status: unsettled, dead, or alive. Currently closest page is EyeShape. Break down further into KillableBigEyes?, DeadBigEyes? and LiveBigEyes?.
3.5. CenterOfTheBigEye?: this is what some people seem really, really convinced is the only true meaning of "nakade". In other words, the killing/living move/point itself; "nakade play". Point to KillableBigEyes?, where killing moves are shown.
4. KillableBigEyes?: corresponds to current KillableEyeShapes page. However, the already-dead big eye shapes on that page should be excluded and/or moved to "Dead Big Eyes" page, limiting this page to true killable eyeshapes. Refer to Unsettled, which is a superset of KillableBigEyes?. UnsettledEyeShapes and KillingShapes currently are highly duplicative of KillableEyeShapes and should be integrated.
5. LiveBigEyes?: current BasicLivingEyeShapes.
6. DeadBigEyes?: current DeadEyeshapes.
7. BigEyesinCapturingRaces?: new. Overall discussion of the basic ideas of how many moves are required to capture a group with a big eye when necessary for semeai purposes. Integrate current BigEyeLiberties, EyeLiberties, and BigEyesWinSemiais?, all of which overlap. We also have Eye versus Eye Capturing Race.
The existing pages for specific types of big eyes are in pretty good shape, and just need tweaking: TwistedFour/DoglegFour (overlaps with BentFour, which should probably be eliminated due to the possibility of confusion with BentFourInTheCorner), StraightFour, SquaredFour, BulkyFive, ThreeInARow, BentThree, PyramidFour, CrossedFive, RabbitySix, ButterflySeven. UnsettledThree currently overlaps with/duplicates ThreeInARow and BentThree. This page can probably be removed without harm.
Peripheral issues:
EyeSpace: clarify to emphasize that this is a term for any kind of space within a group in which eyes could be made in a number of ways, as in a life-and-death problem ("futokoro").
EyeShape: rewrite with a focus on its meaning as EyePotential, while noting the term is also widely used to refer to various big eye topologies. If the editing is done correctly, the ideas in EyeShapeEyeSpaceDiscussion should all be incorporated and that page might not be necessary.
Hmm. Now this seems like a lot of work.
How about the term "one-eyed group"? This solves the problem with terms like "killable big eyes" since it's the group being killed, not the eye. I suppose "monocular" (group) is going too far, although it does give us the handy "monoculus" to refer to the eye itself.
![[Welcome to Sensei's Library!]](images/stone-hello.png "Sensei's Library")