Life In 19x19 http://prod.lifein19x19.com/ |
|
Lake, Nakade, Eye, Eyespace and Related Terms http://prod.lifein19x19.com/viewtopic.php?f=15&t=9335 |
Page 1 of 4 |
Author: | RobertJasiek [ Sat Nov 09, 2013 3:58 am ] |
Post subject: | Lake, Nakade, Eye, Eyespace and Related Terms |
As a motivation, here is an example: If considered generously, the 'potential eyespace' of the black group consists of all marked intersections. The squares denote the group's four 'lakes', where each connected part of squared intersections is a lake. - The upper left lake is one eye. - The upper lake is worth 1.5 eyes. - The left lake is worth 0.5 eyes. - The lower right lake is worth 0.5 eyes. The lakes are parts of the eyespace that, in a life and death situation, tend to be the parts with good potential for becoming eyes. If there is only one lake, the defender wants to partition it, while the attacker wants to convert it to (create) a nakade: Black wants to partition the lake, White wants to create a nakade. The outside can be called 'potential eyespace', but usually Black cannot construct additional lakes there. "A LAKE is a connected part of the potential eyespace of the defender's group that is, or can easily become, visually surrounded by his stones [...]" John has mentioned 'futokoro', but I have never heard of this term; maybe it has always been translated to 'eyespace', 'eye' or something similar. Would you now say that it has about the same meaning as 'lake'? An eyespace can, but need not, contain lakes. Presumably, I use nakade with a somewhat different meaning from the one you use. This, however, does not make my use wrong. I have seen different uses in the English literature. Therefore, when I use nakade, I also explain the meaning that I use: "Simply speaking, a NAKADE is a lake, so that 1. the defender can fill all but one of its intersections, 2. the defender cannot partition it and 3. there cannot be a seki or ko in it." Unlike you have guessed, I do not use nakade as a shape term. As you can infer from the approximative definition, I define it as a lake, i.e., as a part of the board, or you might say: "a set of intersections" or "connected part of the potential eyespace". Furthermore, I use the term "create a nakade" for a play that, surprise, does create a nakade. I.e., I would describe the part of the board just before the move as "a part of the board that a play of the kind 'creating a nakade' can transform into a nakade". The previous example is such a part of the board. The clarity of restricting the meaning of 'nakade' to a part after the interesting play allows us to perceive well the kind of play 'creating a nakade'. This is what the attacker wants: to make a play that creates a nakade. With this clear understanding of a move meaning, solving life and death situations is eased. Other uses of 'nakade' exist, but I have found none else similarly useful when thinking about how to solve an LD problem. Whichever use of nakade somebody (or a particular text) uses, it is bound to be in conflict with other uses of the word. This is a bit unfortunate, but it cannot be avoided, because we are in the age of transition from ambiguous terms to terms with clear meanings. During the transition phase, different people might prefer different uses. Well, this is not ideal, but it is much better when everybody, who uses a term regularly, declares the meaning of his use than continue ambiguous use of undefined terms in a text. The meanings of 'eye', 'nakade' etc. are not well established in the go world yet. Also 'eye' has a (great) variety of meanings from "potential eye" (during the middle game) to "final eye" (as in scoring positions after completely finished yose). A potential eye can, but need not, become a final eye. Similarly, there are potential versus final eyespace of a group. Other varieties of meanings exist. |
Author: | Cassandra [ Sat Nov 09, 2013 4:59 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Dear Robert, There are several issues in your "definitions" that do not fit "common understanding". These points can NEVER become part of Black's "eyespace", just because they will not remain empty. At most, they can become part of the outside bordering line of Black's group. These points are Black's "eyespace" for SURE, as long as Black does not fill-in these points. They are NO LONGER part of "potential" eyespace. At least these two points are part of Black's "potential eyespace" (irrelevant here, because Black's group is already alive). If we take "A LAKE is a connected part of the potential eyespace of the defender's group that is, or can easily become, visually surrounded by his stones [...]" as the definition of "lake", then "An eyespace can, BUT NEED NOT, contain lakes." cannot be true (assuming that "eyespace" also includes "potential eyespace"). |
Author: | Bill Spight [ Sat Nov 09, 2013 5:28 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
RobertJasiek wrote: As a motivation, here is an example: If considered generously, the 'potential eyespace' of the black group consists of all marked intersections. The squares denote the group's four 'lakes', where each connected part of squared intersections is a lake. - The upper left lake is one eye. - The upper lake is worth 1.5 eyes. - The left lake is worth 0.5 eyes. - The lower right lake is worth 0.5 eyes. I have altered the diagram above. I think that we have to count ![]() ![]() RobertJasiek wrote: Whichever use of nakade somebody (or a particular text) uses, it is bound to be in conflict with other uses of the word. This is a bit unfortunate, but it cannot be avoided, because we are in the age of transition from ambiguous terms to terms with clear meanings. I am afraid that, as always, language tends to become less precise over time. One reason that go terms are ambiguous (when they are) is that they have been around for long enough to acquire new meanings. For instance, the root meaning of nakade is clear. It is a play inside an opponent's eye (specifically, a region of connected empty points surrounded by stones of one color). But other meanings have accrued. This example comes from the Gokyo Shumyo. ![]() ![]() ![]() |
Author: | RobertJasiek [ Sat Nov 09, 2013 5:38 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Cassandra, ko threats can let more intersections become eyes. Of course, I get your point that my markup was very generous. For eyespaces, I am pragmatic and use a visual approach as a first approximation. For lakes, I am stricter and (was not cited) exclude 'simple boundary defects', i.e., intersections where necessarily the defender needs to fill. With a similar idea, we can be less generous and exclude the border intersections of the wide eyespaces. Ok, if you want to distinguish potential from final eyespace early, you can do so by mentally partitioning an eyespace into its already final parts and its still potential parts. I have perceived the union of all the parts as one eyespace of the group; as a union, it is still a potential eyespace; when the outside shapes will be settled, the upper parts will remain and belong to the final eyespace. IOW, both approaches work, and it is a matter of choice which one wants to use. Here is an example of a black group with a potential eyespace without lakes: |
Author: | RobertJasiek [ Sat Nov 09, 2013 5:41 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Bill, I have wanted to test who understands lakes. You do:) |
Author: | Boidhre [ Sat Nov 09, 2013 5:49 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Robert, may I ask, why the word "lake?" It really doesn't seem like a natural fit for this concept to me and sounds very out of place. |
Author: | Cassandra [ Sat Nov 09, 2013 5:54 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Dear Robert, Again, and again, your terminology is far, far away from "common understanding". Again, and again, you do not understand that it is. Again, and again, you divert into minor side-aspects of what has been written, or open new fields for discussion, as well as presenting childish "counter-examples". |
Author: | Bonobo [ Sat Nov 09, 2013 6:03 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Boidhre wrote: [..] why the word "lake?" It really doesn't seem like a natural fit for this concept to me and sounds very out of place. Well, to me it doesn’t seem more “out of place” than “eyes” or the “racoon’s belly”. I could adapt this word … I’ll have to think more about it, but something in it appeals to me. For me a lake (not “lah-keh” ![]() BTW I like it that you, RJ, are using a geo-/topographical Real Life word as metaphor in Go ![]() Greetz, Tom |
Author: | Bill Spight [ Sat Nov 09, 2013 6:11 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Boidhre wrote: Robert, may I ask, why the word "lake?" It really doesn't seem like a natural fit for this concept to me and sounds very out of place. Oh, it's not a bad metaphor. Who knows, it might catch on. ![]() |
Author: | RobertJasiek [ Sat Nov 09, 2013 6:18 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Why lake? When I needed a term for the object, it was the word immediately occurring to my mind. Since then, I have always liked it. It is descriptive and short. For go, lake was not used as a term for something else. For a term needed frequently, these are very useful features. Cassandra, with terms of "common understanding" / traditional terms, thinking and written text is nasty, clumsy, ambiguous and inefficient. Being far away from this state is a great advantage! LD is a complex topic, and it is essential to start with efficient language already for the basics. Otherwise, the situation remains as it is: LD theory is very weak in comparison other fields of go theory. Bad terms lead to bad theory, such as the proverb "1. reduce, 2. occupy vital point". We must overcome the current state of ca. 2/3 of all teachers teaching little else about LD than things like this bad proverb. Without proper tools (such as efficient terms), things do not change fast enough. |
Author: | John Fairbairn [ Sat Nov 09, 2013 7:38 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Quote: Simply speaking, a NAKADE is a lake, so that 1. the defender can fill all but one of its intersections, 2. the defender cannot partition it and 3. there cannot be a seki or ko in it." Again, there are lots and lots of people in the western go world who know some Japanese, so why be so obtuse as to misuse nakade like this? But my main question is: Where does Gold Jade 7 and its seki fit in with this? .. ... .. |
Author: | RobertJasiek [ Sat Nov 09, 2013 7:51 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
With a seki possible or existing, the thing is not a nakade. When the seki already exists, call it a 'seki'. When the seki does not exist yet, I call it a 'lake with a possible seki'. (If one wants to be more precise, one can state the two cases Black / White starts.) Similarly for ko, except that for historical reasons it would be called 'lake with a hidden ko' (the alternative is '...possible ko'). EDIT: Defining 'nakade' in a specific way is done for the sake of profiting from this definition during its application as well as possible. When nakade is defined as I suggest, application is by far the easiest and most powerful. This is so, because such a nakade has very nice stable properties, which differ from the properties in the position preceding a nakade. E.g., the attacker wants to create the nakade and wants to get rid of the unsettled position preceding it. E.g., one can speak of 'creating a nakade' and, earlier, 'threatening to create a nakade', and these kinds of creating / threatening to create move meanings correspond consistently to the same kinds of other techniques for moves. Using the same words for things that have the same behaviour is much more useful than using the same word for things with different behaviours. (An alternative terminology could be imagined: 'stable nakade' versus 'unstable nakade'. Then, one would need two words for the stable state. I prefer one word: 'nakade'. Simpler!) |
Author: | Cassandra [ Sat Nov 09, 2013 8:37 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
RobertJasiek wrote: Cassandra, with terms of "common understanding" / traditional terms, thinking and written text is nasty, clumsy, ambiguous and inefficient. Being far away from this state is a great advantage! LD is a complex topic, and it is essential to start with efficient language already for the basics. Otherwise, the situation remains as it is: LD theory is very weak in comparison other fields of go theory. Bad terms lead to bad theory, such as the proverb "1. reduce, 2. occupy vital point". We must overcome the current state of ca. 2/3 of all teachers teaching little else about LD than things like this bad proverb. Without proper tools (such as efficient terms), things do not change fast enough. Dear Robert, It is carrying coals to Newcastle to let me know that Tsume-Go is complex. As usual, you did not really understand what I tried to tell you. "Common understanding" has "potential eyespace" as a subset of "eyespace". And knows that two solidly surrounded eyes are "sure" eyespace, not "potential". In "your world", "potential eyespace" is the collecting term for solidly surrounded eyes (= "sure eyespace") and potential eyes. And this will not fit the expectation of the "usual" reader. |
Author: | RobertJasiek [ Sat Nov 09, 2013 8:43 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Ok, you might be right about this. Anyway, the problem occurs only if a group has already 2 safe eyes. When one has determined that much, all further discussion is l'art pour l'art:) |
Author: | hyperpape [ Sat Nov 09, 2013 2:08 pm ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
I like the term "lake", but it seems rather imprecise for you, Robert: why "visually surrounded"? And why "can become"? Under what conditions? |
Author: | RobertJasiek [ Sat Nov 09, 2013 3:05 pm ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
hyperpape, since I may not explain the degree of imprecision in this thread, please see here: viewtopic.php?p=152402#p152402 "Visually surrounded" could be made reasonably precise by referring to hulls constructed from the defender's bounding stones, where bounding is defined trivially a la Pauli. Defining "can easily become" is trickier. I am not sure yet if this needs or can avoid "force". "Visually surrounded" is used, because the concept 'lake' is a pragmatic means meant to help pruning reading. For this purpose, "visual" is good enough precision. If much greater precision is needed, one can fall back to the concept 'potential eyespace' and to more detailed reading. Potential eyespace has another informal definition and is also a pragmatic concept. If necessary, we fall back to even more detailed reading, which is defined precisely somewhere. If a lake IS visually surrounded, things are easy. Not all lakes are easy, but there are lakes with open ends or gaps in the visual surrounding. For these circumstances, "or can easily become" is also needed. The gaps must, however, not be so big, loose and wide that nothing is meaningfully surrounded. There must at least be a principle perception of of surounding the lake easily. Currently, the conditions for such surrounding are undefined. OTOH, one can construct such conditions using hulls again and allowing the defender successive imagined plays along a hull's boundaries. Well, if you really want it done by formal research. As I say, lake is a pragmatic concept, as are 'potential eye' and 'potential eyespace' at the moment. I am satisfied if they are just big enough to include every relevant intersection. it is like for local move selection, which also relies on a big enough set of local intersections to be sure not to miss any relevant intersection. Are you not used to my occasional use of fuzzy concepts?:) Fuzzy is not 'guessing' but is 'safe approximation'. Like nigiri: if you put your hand into the bowl, you will surely fetch at least a few stones, and this is good enough to guarantee a working procedure; it suffices to avoid fetching only zero stones. If only our regions are large enough, the relevant intersections will be in them. |
Author: | cyclops [ Sat Nov 09, 2013 4:28 pm ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
Cassandra wrote: ... "Common understanding" has "potential eyespace" as a subset of "eyespace". And knows that two solidly surrounded eyes are "sure" eyespace, not "potential". ... In mathematics a set is a subset of itself. So even the empty set has a subset. One reasons this choice is made is that it allows to formulate statements more efficiently on the average. In the same way it might be more efficient to include real disasters ( say ) into the concept of potential disasters. Better to state this explicitly, though. |
Author: | RobertJasiek [ Sun Nov 10, 2013 12:41 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
From the Nakade thread viewtopic.php?f=15&t=9025 I cannot infer the meaning of 'nakade' in Japanese. In its list of Japanese terms, The Go Player's Almanac allows the move, a position or a shape to be called a nakade, but presumably "position" shall refer only to a relevant local part of the board, although it might possibly include also surrounding intersections and their stones. Anyway, whichever of these uses might be allowed in a Japanese use of the term, all these uses are related to the simplifying conditions "connected part of the board so that 1. the defender can fill all but one of its intersections, 2. the defender cannot partition it and 3. there cannot be a seki or ko in it." - A nakade move (I say: a move creating a nakade) is a move creating a whole board position with these conditions. - A nakade shape is a) the attacker's strings in a part of the board with these conditions, b) the appearance of empty intersections and black or white stones on a part of the board with these conditions or c) (b) together with the appearance in some neighbourhood. - A nakade position is a) a part of the board with these conditions, b) (a) together with some neighbourhood, c) a part of the board for which a move transforms it into (a) or d) a part of the board for which a move transforms it into (b). |
Author: | Cassandra [ Sun Nov 10, 2013 1:33 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
RobertJasiek wrote: "connected part of the board so that 1. the defender can fill all but one of its intersections, 2. the defender cannot partition it and 3. there cannot be a seki or ko in it." Why do you exclude the possibility of a Ko-fight for completing the Nakade shape ? |
Author: | RobertJasiek [ Sun Nov 10, 2013 2:28 am ] |
Post subject: | Re: Lake, Nakade, Eye, Eyespace and Related Terms |
When a text is not about ko as its main topic, I do not want to write - 98 times "a nakade without the possibility of a seki or ko", - 1 time "a nakade with the possibility of a seki" and - 1 time "a nakade with the possibility of a ko". I prefer to write - 98 times "a nakade", - 1 time "a lake with the possibility of a seki" and - 1 time "a lake with the possibility of a ko". |
Page 1 of 4 | All times are UTC - 8 hours [ DST ] |
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |