Ing-ko Lecture - Summary of a Thread (View with a fixed width font!) Robert Jasiek 2000-06-08 Subject: Ing-ko Lecture (1) Date: Thu, 09 Dec 1999 20:49:34 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 , 3 Antoon Pardon wrote: > What I expect is enough understanding to play the rules should > the occasion arise. Ok, I shall restrict explanations to important and frequent cases unless someone asks for more. > I don't think I will ever like the rules So don't I. They are just theoretical study fun to me:) > I certainly think this is not good news. However the rules seem > to be played, so I like to know how and in so far I can apply > them when somebody invites me to a game with these rules. This is also the case for any other professional rules. However, Ing-ko is the toughest of all. > Do I understand correctly that an equivalent shape is a shape [...] Let me cite my ko mouth definition: "A ko mouth consists of the marked board points in any of the following shapes, in any equivalent shape, in any equivalent shape with reversed colours, or in any such shape of which a part or parts - but not any marked board point - are cut off by at least one edge of the board. In any shape diagram not marked empty points may be of empty, black, or white colour as long as no illegal position arises." I only use "equivalent" here and it refers to dislocation or symmetries. E.g. all stones might be shifted one line higher (a dislocation), or rotated by 90° (a symmetry). Reversed colours should be clear. > that started from one of your examples You should not consider the diagrams here to be examples but they are reference diagrams for the purpose of defining the board points belonging to some ko mouth. E.g. the first reference diagram can be applied to a double ko that has two ko mouths that fit that first reference diagram. > and to which a number of > stones were added, as long as those added stones are not on a > marked empty point? They may not be on a marked empty point, they need not be added, and the resulting position must be legal (all stones have breaths). > The shape can also be shifted to an edge of the board, even so > far as to lose a number of stones that will fall of the edge, as > long as all marked intersections, empty or occupied remain on the > board? Yes. E.g. you all know of the first shape case occurring on an edge or in the corner. > I'm puzzled about the "no other reasonable ko mouth shapes are > known". This suggest there is some kind of definition of which > the illustrations are examples. One can have and knows about examples for each of the reference diagrams so that a fight about the stones on the ko mouth makes sense: There is a real ko fight or at least a real yose fight about the ko mouth points. So far the worst known shape that makes some such sense has triple ko stones, see the official rules for an example. Quadruple ko stones do not make sense. There in all known positions the players might cooperate to cycles through such ko mouths while it remains unclear why they ever should do so. In practice you should be happy with single or double ko stones and do not bother about everything else; it won't occur in practice anyway. E.g. eternal life or round-robin ko occur in every 10000th game or so and have double ko stones. In the definition's reference diagrams you find the proper double ko stones shapes. > But my first impression was that > a ko-mouth was defined as being one of those illustrations. You should indeed wonder why ko mouth is defined in such a nasty way and is the first thing to be defined. (From now on with Ing91 I mean the official rule text and with Newko I mean my interpretation text on my newko.html.) Ing91 starts by introducing "ko stones". However, I have shown that such a definition could ever only make sense by being dependent on perfect play. Thus for Newko we cannot start with ko stones. Ing91 restricts the number of stones in ko strings to 3 anyway, so Newko does the same by its reference diagrams. Listing all possible reference diagrams is more or less straightforward: No one could ever point out another possibility of a ko mouth shape with at most 3 stones in any ko string. Newko starts with ko mouths so that ko stones can be defined and can be defined independently of perfect play. Ko mouths will also be needed for classes of ko types later, but the first object is to introduce ko stones in a way that is more formal than intuition. Ing91 with its "stones that can be captured cyclically or repeatedly" is so imprecise that one would have to resort to intuition. You know, I did not want to rule "consider a stone to be a ko stone if your intuition tells you so";) The definition of ko mouth is nasty because there are a lot of cases for the reference diagrams. First I tried to define ko stones directly until I realized why it was bound to fail. So Newko has to start with ko mouth shapes, there is no simpler formal alternative, unfortunately. In practice it is still a good idea to use intuition to see what a ko stone is. However, for the definitions of ko types ko mouths should be understood as being only those parts of ko situations on that ko stones can be played in some cycle or part of a cycle. Ing91 uses insufficient language for what a ko be. Restricting Newko to the board points of the ko mouths only allows clear definitions of what will be part of a ko position and what not. Thereby ko positions won't intersect, which would be a mess. Why do we need to know what a ko mouth and what a ko stone is? Rules will especially restrict play of ko stones in ko mouths! Are there any questions about ko mouth left? Or do you want to see examples of applications other than on my e1.html? Otherwise I will proceed to the next definitions. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (2) Date: Fri, 10 Dec 1999 14:32:11 +0100 From: Robert Jasiek Newsgroups: rec.games.go Now that everybody has understood what Newko means with ko mouth, we can proceed to ko stones and move-sequences. First I cite Newko definitions that will be used in later definitions and in rules: 1) "A _ko string_ is a string in a ko mouth." 2) "A _ko stone_ is a stone of a ko string." 3) "A _move-sequence_ is a non-empty colour-alternating sequence of board plays or pass plays by both players." 4) "A _ko stone move-sequence_ is a move-sequence of board plays so that only ko stones are played." As you remember from the ko mouth lecture, a ko mouth includes only the central board points of some relevant shape. So only there ko strings do occur. Examples for (1): . . . . . . . . # O . . . # . # O . . . # O . . . . . . . . position (a) . . . . . . . . # O . . . # @ @ O . . . # O . . . . . . . . the ko mouth @ . . . . . . . . # O . . . # . + O . . . # O . . the ko strings + in the ko mouth . . . . . . (there is only one ko string) . . . . . . . . # O . . . # . # O . . . # O . O . . . # O . position (b) . . . . . . . . # O . . . # A A O . all ko mouths: . . # B B O the ko mouth A and the ko mouth B; . . . # O . currently in the lower right corner there is no ko mouth . . . . . . . . # O . . . # . a O . . . # b . O all ko strings: . . . # O . the ko string a and the ko string b . . . . . . . . # O . . . # . # O . . . # . # O . . . # O . position (c) . . . . . . . . # O . . . # . # O . all ko mouths: . . # A A O the ko mouth A and the ko mouth B; . . . # B B the surroundings of ko mouth B are cut short by edges . . . . . . . . # O . . . # . # O . . . # . a O all ko strings: . . . # b . the ko string a and the ko string b # O O . # O # # # O O O . . . . . . position (d) # @ @ @ @ O # # # O O O the ko mouth @: . . . . . . its surroundings are cut short by the upper edge # a a . b O all ko strings: # # # O O O the ko string a and the ko string b; . . . . . . a is a ko string of two stones; the ko mouth contains two ko strings at the same time (2) should be clear. A move-sequence (3) simply describes a sequence of alternating moves. The players alternate and they can play board plays or pass plays. (4) defines ko stone move-sequence. It consists of board plays only and all of them play a ko stone on the board. The definition will be used later especially in rules that restrict ko stone move-sequences. So far they may still consist of one or of several moves as in the following examples: . . . . . . . . # O . . . # . # O . . . # O . . . . . . . . position (a) . . . . . . . . # O . . . # a b O . . . # O . . . . . . . . involved board points Possible ko stone move-sequences [] with O playing first: [a] [ab] [aba] [abab] etc. and [ab]*, where * means an infinitely repeated sequence Obviously we will need some rule to restrict this nonsense. . . . . . . . . # O . . . # . # O . . . # O . O . . . # O . position (b) . . . . . . . . # O . . . # a b O . . . # c d O . . . # e f involved board points Some possible ko stone move-sequences [] with the player playing first denoted: O[a] at this moment only one ko stone is on the board O[abab] after that again two ko strings are on the board #[d] the two lower corner points have become a new ko mouth #[dc] the original ko mouths and ko stones are restored Counter-example with p meaning pass play: #[dpf] is not a ko stone move-sequence even though f is a played ko stone because a ko stone move-sequence may not contain any pass play for the definition's sake. # O O . # O # # # O O O . . . . . . position (d) # a b c d O # # # O O O . . . . . . involved board points A possible ko stone move-sequence: #[cbda] This happens to be a cycle. Such a sequence might occur in an eternal life pattern. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (3) Date: Sun, 12 Dec 1999 10:13:44 +0100 From: Robert Jasiek Newsgroups: rec.games.go This time ko coupling and ko position are introduced. Ing91 does not use the term ko but rather the term ko position. It is not entirely clear what shall be meant. However, we know that a distinction of ko types shall be possible, i.e. for every "ko" in question we have to know whether it is a disturbing ko or a fighting ko. Thus Newko uses the term ko position as a smallest unit for that a ko type can be assigned. Sane rules would simply use any single ko mouth as a ko position. Unfortunately, Ing91 rules are insane: Some ko positions consist of just one ko mouth, other ko positions consist of more than one ko mouth. Ko positions that consist of more than one ko mouth are linked by a ko coupling: all involved ko mouths are surrounded by exactly the same strings. Ing seemed to intend to classify even more complex unifications of ko mouths than those just linked by a ko coupling. However, he never gave any such example. To avoid arbitrarily complex shape definitions, Newko restricts itself to the worst case construction of a ko position: that given due to one ko coupling. This is good enough to cover especially all official examples of ko positions, whether with or without ko coupling. Now let us review the Newko definition of ko coupling: "Given a ko mouth and its adjacent strings. All of at least two ko mouths with all of the strings and none else being adjacent to each of the ko mouths have them as their _ko coupling_." So a ko mouth can or cannot have a ko coupling. A ko mouth always has adjacent strings, however, they are not always a ko coupling. They can only be a ko coupling by assuming this feature at the same time for at least two ko mouths. Two or more ko mouths are linked by a ko coupling if and only if all strings adjacent to the ko mouths are adjacent to each of them. Thus all ko mouths involved are adjacent to the same strings and none of the ko mouths is adjacent to any further string. Next we come to the Newko definition of ko position: "If a ko mouth is adjacent to all strings of a ko coupling and not adjacent to any other string, then the _ko position_ of the ko mouth is the union of the board points of all ko mouths that are adjacent to all strings of the ko coupling and not adjacent to any other string. If a ko mouth is not adjacent to all strings of a ko coupling or adjacent to another string, then the board points of the ko mouth are its _ko position_." Thus there are two cases of ko positions. The first case is a ko position that consists of all ko mouths surrounded by one ko coupling. Hereby two or more ko mouths are linked by their ko coupling and form the ko position together. The ko position consists of all board points of all involved ko mouths. The ko mouths are not adjacent but only linked by the ko coupling. Nevertheless they constitute the ko position. The second case is a ko mouth without any ko coupling. There the ko position consists of just the board points of the one ko mouth. It does not have adjacent strings so that they all would be the adjacent strings of another ko mouth as well and thereby would be forming a ko coupling and thus result in a larger ko position. Examples: . . . . . . . . # O . . . # . # O . . . # O . . . . . . . . (1) . . . . . . . . # O . . . # @ @ O . . . # O . . all ko mouths @; . . . . . . there is only one ko mouth and no ko coupling . . . . . . . . # O . . . # @ @ O . . . # O . . the ko position @ . . . . . . . . . . . . . . . . . # O . O # . . . # . # O . O # . . . # O . O # . . . . . . . . . . . (2) . . . . . . . . . all ko mouths: . . # O . O # . . the ko mouth A and the ko mouth B . # A A O B B # . one O string is adjacent to both ko mouths, however, . . # O . O # . . not all strings adjacent to A are adjacent to B . . . . . . . . . so there is no ko coupling . . . . . . . . . . . # O . O # . . . # A A O B B # . . . # O . O # . . all ko positions: . . . . . . . . . the ko position A and the ko position B . . . . . . O . # # # # # . . # . # . # . . O # O # O . . O O O O O . . . . . . . O (3) . . . . . . O . # # # # # . . # A # B # . . O A O B O . . O O O O O . all ko mouths: . . . . . . O the ko mouth A and the ko mouth B . . . . . . O . @ @ @ @ @ . . @ . @ . @ . . @ # @ # @ . . @ @ @ @ @ . all ko couplings: . . . . . . O there is only one ko coupling @ . . . . . . O . # # # # # . . # @ # @ # . . O @ O @ O . all ko positions: . O O O O O . there is only the ko position @ consisting . . . . . . O of two ko mouths You need not care much about ko type or rules for this kind of position. The players will leave it until the late yose anyway. Then either # connects one ko stone eliminating the ko coupling or O starts by capturing one ko stone and enabling each player to connect one of the two ko stones in the ko coupling then. (It will be a disturbing life and the basic ko rule will prohibit immediate recapture of a single ko stone.) . . . . . . O . . . . . # # # # # . . # O . . # . # . # . # . # O . O # O # O . . # O . . O O O O O . . . . . . . . . . . O . . . . (4) . . . . . . O . . . . . # # # # # . . # O . . # A # A # . # B B O . O A O A O . . # O . . O O O O O . . . . . all ko positions: . . . . . . O . . . . the ko position A and the ko position B . # O . # O O O # # # O . # O . (5) A A O . # O O O # # # O all ko mouths: . # B B the ko mouth A and the ko mouth B . # @ . @ @ @ @ @ @ @ @ all ko couplings: . @ O . the ko coupling @ + + O . # O O O # # # O all ko positions: . # + + the one ko position + # # # # # . # . # O # O # O . O O O O O (6) # # # # # A # B # C A O B O C the three ko mouths A, B, C O O O O O @ @ @ @ @ . @ . @ O # @ # @ . the one ko coupling @ @ @ @ @ @ # # # # # + # + # + + O + O + the one ko position + O O O O O As you can see, ko couplings are useful for double kos, triple kos, etc. Each string of a ko coupling is adjacent to each of the ko mouths in the ko position given by the ko coupling. The ko mouths surrounded by the one ko coupling are the ko position. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (4) Date: 12/13/1999 Author: Robert Jasiek Every ko position has a ko type. First for every ko position its ko type is determined then rules considering that particular ko type are applied. There are the major ko types disturbing ko and fighting ko. Disturbing kos are classified into disturbing life and disturbing death. Disturbing life is classified into disturbing life with a ko coupling and disturbing life without a ko coupling. The latter shall also be called single disturbing life because it consists of just one ko mouth. Below we discuss the former. However, let me first remark that while Ing91 pretends to classify ko types due to shape patterns and fails due to too many theoretically possible patterns, Newko simply classifies due to numbers of breaths of strings adjacent to a ko position. This always works and is good enough for an approximation of Ing91. A ko position that is a disturbing life with a ko coupling can be either in a stable state or in an instable state. In a stable state both players can safely tenuki or pass, in an instable state exactly one player prefers to play another ko stone in the ko position immediately to transform it into a stable state again. Next we look at the Newko definitions, which are slightly simplified, since we do not discuss rare cases here anyway. (On my web page you find a more complete definition.) "A ko position of a ko coupling is in a state of _stable disturbing life_ if all strings of the ko coupling have at least two breaths and if no move of one player can reduce the breath number of at least one of the other player's strings of the ko coupling to one so that the other player has without repeating the board position no possibility to regain another breath for each of the affected strings immediately." The first condition is a general feature of stable disturbing life. (Two breaths of each string of the ko coupling prevent any immediate capture of them.) The second condition ensures an interchange of breaths between the strings of the ko coupling. (Hereby an immediate recapture in a ko mouth of two board points shall be no interchange. An interchange means that any attack by means of a ko stone move can be answered by a defensive ko stone move. There is a breath balance between attack and defense.) "A ko position of a ko coupling is a state of _instable disturbing life_ if the ko position is not in a state of stable disturbing life and a ko stone move in the ko position exists that transforms the ko position into a state of stable disturbing life." The first condition is useful because the strings of the ko coupling of a disturbing life might have more than two breaths each and then any attack of reducing opposing breaths would still leave the disturbing ko in a stable state. It only becomes instable if at least one string of the ko coupling is down to only one breath. The second condition requires the existence of a defending ko stone move that regains two breaths for every string of the ko coupling and thereby achieves stability of the disturbing life again. Examples: . # # # # # # (1) # # . # O # O O O # O . O . ko position with ko coupling and three ko mouths O O O O O O O The ko position is in a state of stable disturbing life because a) each string of the ko coupling has two breaths, b) a O ko stone move can be answered by a # ko stone move in another ko mouth and regain two breaths for each string of the ko coupling immediately, c) a # ko stone move in either of two possible ko mouths can be answered by a O ko stone move in the left ko mouth and regain two breaths for each string of the ko coupling immediately. . # # # # # # (2) # # O # O # O O O . O . O . ko position with ko coupling and three ko mouths O O O O O O O The ko position is in a state of instable disturbing life because a) it is not a stable disturbing life because not all strings of the ko coupling have two breaths, b) e.g. the ko stone move by # that captures in the left ko mouth transforms the ko position into a stable disturbing life. . # # # # # # (3) # # . # . # O O O # O # O . ko position with ko coupling and three ko mouths O O O O O O O The ko position is in a state of instable disturbing life because a) it is not a stable disturbing life because not all strings of the ko coupling have two breaths, b) e.g. the ko stone move by O that captures in the middle ko mouth transforms the ko position into a stable disturbing life. # # # # # # # (4) . # . # O # O # O # O . O . ko position with ko coupling and four ko mouths O O O O O O O The ko position is in a state of stable disturbing life because a) each string of the ko coupling has two breaths, b) a O ko stone move can be answered by a # ko stone move in another ko mouth and regain two breaths for each string of the ko coupling immediately, c) a # ko stone move can be answered by a O ko stone move in another ko mouth and regain two breaths for each string of the ko coupling immediately. # # # # # # # (5) . # O # O # O # O . O . O . ko position with ko coupling and four ko mouths O O O O O O O The ko position is in a state of instable disturbing life because a) it is not a stable disturbing life because not all strings of the ko coupling have two breaths, b) e.g. the ko stone move by # that captures in the right ko mouth transforms the ko position into a stable disturbing life. Quintuple ko works like quadruple ko. The extra ko mouth does not hurt, it might only help one player to maintain stability more easily. # # # # # # # (5) . # . . # O O O O # # O . . ko position with ko coupling and two big ko mouths O O O O O O O The ko position is in a state of stable disturbing life because a) each string of the ko coupling has two breaths, b) no # move reduces the number of breaths of the O string of the ko coupling to one, c) no O move reduces the number of breaths of the # string of the ko coupling to one. # # # # # # # (6) . # . . # O O O O # # O # . ko position with ko coupling and two big ko mouths O O O O O O O The ko position is in a state of instable disturbing life because a) it is not a stable disturbing life because a # move in the ko position cannot be answered by any O move that would regain the O string of the ko coupling any second breath immediately, b) a O ko stone move that throws in exists that transforms the ko position into a state of stable disturbing life. # # # # # # # (7) . # O . # O O O O # # O # . ko position with ko coupling and two big ko mouths O O O O O O O The ko position is in a state of stable disturbing life because a) each string of the ko coupling has two breaths, b) a # ko capture of two stones still allows O to regain at least two breaths for his ko coupling string immediately by capturing in the other ko mouth, c) a O ko capture of two stones still allows # to regain at least two breaths for his ko coupling string immediately by capturing in the other ko mouth. # # # # # # # (8) . # O . # . . O O # # O # # ko position with ko coupling and two big ko mouths O O O O O O O The ko position is in a state of instable disturbing life because a) it is not a stable disturbing life because not all strings of the ko coupling have two breaths, b) a O move capturing two stones achieves stability again. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (5) Date: Mon, 13 Dec 1999 19:54:58 +0100 From: Robert Jasiek Newsgroups: rec.games.go To recall, a disturbing life can have a ko coupling or can have none. Now we treat the second case and call the ko position that consists of one ko mouth only a single disturbing life. The following text is a slight simplification of the Newko text so that consideration of rare ko mouth cases is avoided and we can concentrate on the essential features. "A ko position which is not adjacent to all strings of a ko coupling or is adjacent to another string is a _single disturbing life_ if - with each adjacent string to the ko position having as an assumption at least one breath not in the ko mouth - on it a ko stone move-sequence of infinite length could not be played." The decisive characteristic is the impossibility of an infinite ko stone move-sequence within the ko mouth. All possible ko stone move-sequences are finite because alternation within them would require pass plays, which may not be part of ko stone move-sequences by definition. Examples: . . . . . . . # # # # . . # . . # . . O # # O . . O O O O . . . . . . . (1) O can play a ko stone in the block-of-4 ko mouth. However, then # cannot play any ko stone in it. So an infinite ko stone move-sequence within the one ko mouth cannot be played and this ko position without a ko coupling is a single disturbing life. In practice this shape is of little relevance because # would rather capture any thrown-in O ko stone. . # # . O # O O O O . O O O O O O O O O (2) + # # . O + O O O O + O O O O the ko position + consisting of only one ko mouth O O O O O 2 # # . O #3 recaptures below 2, O4 passes. # O O O O The sequence 1..3 consists of ko stone moves, however, 1 O O O O then O passes since he cannot play another ko stone O O O O O immediately. Thus in the ko position an infinite ko stone move-sequence cannot be played and it is a single disturbing life. In this case it is of some practical relevance: Life of the involved strings is hardly changed while the ko mouth can only be used for a ko threat consisting of the sequence 1..3. (When any player would like to use a gote ko threat is another matter...) -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (6) Date: Tue, 14 Dec 1999 14:07:41 +0100 From: Robert Jasiek Newsgroups: rec.games.go Like disturbing life, disturbing death can or cannot have a ko coupling. Disturbing death with a ko coupling requires surprisingly many conditions in Newko definitions to distinguish it from disturbing life and from fighting kos properly. In fact, only one shape pattern seems to be relevant. The reason is that a string of the ko coupling of one colour shall be described as being on the verge of removal (when in a stable state) or as temporarily escaping removal (when in an instable state). Only such conditions are helpful in the definitions that suitably describe a change between one breath and two breaths for at least one string of one colour of the ko coupling. This can only be easily achieved for a ko position of exactly two ko mouths of two board points each. Ing has never presented any other possible case, so such narrow definitions are good enough as an approximation in Newko. Its defintions are as follows: "A ko position of a ko coupling is a state of _stable disturbing death_ if - the ko position consists of exactly two ko mouths, - the ko mouths of the ko position have exactly two board points each, - the ko position has a ko stone of each colour, - at least one string of one colour of the ko coupling has exactly one breath, and - all strings of the other colour of the ko coupling have at least two breaths." "A ko position of a ko coupling is a state of _instable disturbing death_ if - the ko position is not in a state of stable disturbing life, - the ko position consists of exactly two ko mouths, - the ko mouths of the ko position have exactly two board points each, and - at least one string of the ko coupling, but all of one colour, has exactly the breaths of the ko position as its breaths." Examples: . # # # # # # . # O O O # O . O O O O O (1) The ko position is in a state of stable disturbing death because, - it has a ko coupling, - it has exactly two ko mouths, - each ko mouth has exactly two board points, - it has a ko stone of each colour, - the O string of the ko coupling has only one breath, and - all # strings (only one) of the ko coupling have at least two breaths each. O is sometimes called dead in disturbing death. . # # # # # # O # O O O . O . O O O O O (2) The ko position is in a state of instable disturbing death because, - it is not a stable disturbing life because not all strings of the ko coupling have two breaths, - it has exactly two ko mouths, - each ko mouth has exactly two board points, and - the O ko coupling string has exactly the breaths of the ko position as its breaths. As one can see, from the temporary attempt to escape being threatened to be captured as in the instable state just another ko capture (in the other ko mouth, if necessary) leads to a stable dead state with one breath only again immediately. # # . . . . # # # # # # . # O O O # O . O O O O O (3) More breaths for the opponent of the dead strings' player do not change analysis. It is still a disturbing death (here in a stable state). *** In the case of disturbing death there is also the single disturbing death without ko coupling: "A ko position consisting of exactly two board points is a _single disturbing death_ if in the cases of a ko stone of one respectively the other colour in the ko mouth all adjacent strings to the ko stone could have after a move of the player of these strings at least two breaths respectively at most one breath provided that the move would neither capture the ko stone nor reduce the breaths of at least one of these strings." The adjacent strings to the ko mouth of one colour waver between being on the verge of immediate removal and a postponement. The ko mouth shall consist of exactly two board points. Indepedently of which ko stone is in the ko mouth the ko position is analysed for both cases at the same time. One of the cases is just imagined. For each case of ko stone in the ko mouth an analysis takes place. The (possibly imagined) ko stone is left as is and may not be removed. The strings (all of one colour) adjacent to the ko stone are examined as to their breaths. Herefore the player of these strings gets an imagined free move (board play or pass play) outside the ko mouth for possibly gaining more breaths in analysis. So the possible breaths of the strings adjacent to the actually imagined ko stone in the ko mouth are counted. For one player there can be at least two adjacent breaths for each string in question. For the other player there can be at most one breath each. Examples: O . O # . . O # # # O O # # . (1A) For analysis the following other ko stone case is imagined as well: O # . # . . O # # # O O # # . (1B) In (1A) the # string is adjacent to the ko stone and thus its breaths are evaluated. Even after an imagined pass play it has already two breaths. In (1B) the O strings are adjacent to the ko stone and thus their breaths are evaluated. No O board play outside the ko mouth could increase their breaths, so O best move is, sadly, a pass play. After that imagined move each of the O strings in question still has only one breath. Both subanalyses together show that the ko position is a single disturbing death. (One might call O to be dead in disturbing death.) O . O # . O O # # # O . # # . (2) Analysis is as before (but only one O string adjacent to the ko mouth need to be considered). So this is a single disturbing death, too. . . . . . . (3A) . . # O . . . # . # O . . . # O . . . . . . . . counter-example . . . . . . (3B) . . # O . . . # O . O . . . # O . . . . . . . . other analysis case In (3A) and in (3B) the adjacent strings can have two breaths each, so the ko position is not a single distubing death. . . . . . . (4) . . # O # . . # . # O # . . # O . O . . . # # . counter-example The O adjacent strings are the ones that might all remain at only one breath each. However, O could extend his upper stone, e.g., and thus prevent that then all his adjacent strings have only one breath each. (The ko position will be a fighting ko.) O . O # . (5) O O # # # O . # # . . # # # # counter-example Presumbly Ing would have called this a disturbing death, too, because he did not just count breaths but tried to evaluate breath patterns functionally. Whatever O does, he cannot gain two breaths of a kind that Ing would have termed permanent. Newko, however, ignores shapes, because there will always be shape patterns that cannot be described in theory. The O string adjacent to the ko mouth still has two breaths outside the ko mouth and thus the ko position still is a fighting ko rather than a single disturbing death. However, in practice this makes no relevant difference because # can easily approach breaths other than the ko mouth breath first and then finally remove that one ko mouth breath as well. How long the ko position is a fighting ko under Newko is just an academic question. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (7) Date: Tue, 14 Dec 1999 16:56:18 +0100 From: Robert Jasiek Newsgroups: rec.games.go After all the trouble of defining various types of disturbing kos Newko can be most pragmatic about fighting kos: "A fighting ko is a ko position that is no disturbing ko." Everything that shall not be a fighting ko has been defined to be a disturbing ko. Still fighting kos can or cannot have a ko coupling. Examples: . . . . . . . . # O . . . # . # O . . . # O . . . . . . . . (1) This is not a single disturbing life because endlessly capturing and recapturing the ko stone is an infinite ko stone move-sequence. We have also already seen that it is not a single disturbing death. Thus it is a fighting ko. Surprise:) # # # # # . # . # O # O # O . O O O O O (2) This is not a stable disturbing life because not all strings of the ko coupling have two breaths. This is not an instable disturbing life because even if O moves next, still the resulting position would not be a stable disturbing life because not all strings of the ko coupling would then have two breaths. This is not a stable disturbing death because it consists of more than two ko mouths. For the same reason it is not an instable disturbing death. Thus it is a fighting ko. *** If you have any difficult shape for that you cannot identify its ko type, then you might ask for it. Especially I appreciate spectacular shapes never seen before:( Otherwise I shall proceed to the nature of disturbing moves. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (7) Date: Wed, 15 Dec 1999 11:50:17 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 Les Fables wrote: > # O . > . # O > # O . > It has the basic shape of a disturbing death with ko coupling, but it > is not played the same way. > How would Ing91 and NewKo treat this case? It does not have a ko coupling because the upper left string is not adjacent to the lower right ko mouth. Newko: Is the upper right ko mouth a single disturbing death? # can play in the lower right corner to provide at least his middle string with a second breath. However, he cannot gain a second breath for his upper left string by playing outside the ko mouth. With an imagined # stone in the upper right ko mouth O has no move outside the ko mouth that would provide his right string with a second breath. Thus for no player all his adjacent strings could get two breaths. The defintion of single disturbing life requires this for one player. Hence the upper right ko position is not a single disturbing life but a fighting ko. Likewise, the lower right ko position is a fighting ko. After any move on the board the ko types of any then existing ko position must be reevaluated... (This is one of the really troublesome patterns.) Ing91: O moving first would kill #. # moving first would play through the board position in a windmill-like manner. (Ing does not presume other restrictions than the basic ko invariation when analysing ko types.) So O never gets to kill all # stones. Thus playing first is decisive for life and death. Thus everything is "unsettled". Thus all involved ko positions are fighting kos. (The Ing91 explanation is really troublesome, of course...) Comparison: Whether both Newko and Ing91 treat the board position the same way will depend on the ko types under Newko after some move. If they should turn out to be fighting kos all the time, then treatment under Ing91 will be the same. Epilogue: # . # . # O # O . Consider the upper right ko mouth, e.g. O cannot play any move outside it to get two breaths for his right string. With an imagined O ko stone in it # cannot play any move outside it to get two breaths for his upper left string. Thus it is not a single disturbing life but a fighting ko. So Ing91 and Newko will treat the board the same way. (Newko does not consider hot stones, but this makes no difference, because in basic ko mouths all ko stones could be hot anyway.) Before determining a result we need to wait for the fighting ko rule, however... -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (7) Date: Wed, 15 Dec 1999 14:51:26 +0100 From: Robert Jasiek To: [...] References: 1 > I appreciate your long and exhaustive discussion of ing-ko Thx! > but I have to admit that it only convices me that the > whole business is a pile of crap. how could Ing and his > cohort ever come up with this? terrible! Ing's character was not that simple, e.g. he was mainly responsible for the one and only logical pro rule set ever (Taiwan 1975-1986). - One of my theories for his late Ing-ko rules is that he was secretly only interested in strange and wonderful shapes. Spending a few millions on rules is cheaper than buying famous paintings nowadays;) Of course, the current Ing rules _are_ terrible. BTW, I have not seen your mail on RGG yet. Kindly, -- robert jasiek Subject: Ing-ko Lecture (8) Date: Wed, 15 Dec 1999 10:23:11 +0100 From: Robert Jasiek Newsgroups: rec.games.go Before we can go to the rules we still need some preparations for the disturber rule to come: It must be clear exactly which moves are disturbing. Ing91 does not define this at all, so Newko has some freedom here. For each disturbing ko type disturbing is defined. However, one can only disturb stable, balanced states, not instable states, which are already unbalanced. The Newko text is slightly simplified here to avoid a discussion of arcane ko mouths. "A player is _disturbing_ a disturbing ko if 1) he plays a ko stone in a ko position that is in a state of a stable disturbing life and reduces the breath number of at least one string of the opponent, 2) he adds a ko stone to a single ko stone in a single disturbing life, 3) he has at least one string that has exactly one breath of the ko coupling of a ko position that is in a state of stable disturbing death and plays a ko stone in the ko position, or 4) he plays a ko stone in a single disturbing death and could not have played any other move that would have provided all his strings that are adjacent to the ko mouth with at least two breaths." (1) is for playing a ko stone in a stable disturbing life. There disturbing is reducing the breath number of at least one opposing string. Example: # # # # # # # (a) . # . # O # O # O # O . O . O O O O O O O stable disturbing life # # # # # # # . # . # . # O # O # O 1 O . O O O O O O O # plays a ko stone at 1 Move 1 disturbs the ko position because it reduces the O ko coupling string from two breaths to one breath. # # # # # # # (b) . # . # . # O # O # O # O . counter-example O O O O O O O instable disturbing life # # # # # # # 1 # . # . # O . O # O # O . O O O O O O O O plays a ko stone at 1 Move 1 does not disturb because it is played in an instable ko position rather than in a stable ko position. *** (2) Characterizes single disturbing lifes. There a move disturbs if it adds a ko stone to a single ko stone. . # # . O (c) # O O O O . O O O O O O O O O single disturbing life . # # . O # O O O O 1 O O O O O O O O O # plays at 1 Move 1 disturbs the ko position because it adds a ko stone the #'s prior single ko stone. . # # . O (d) # O O O O # O O O O counter-example: O O O O O single disturbing life 1 # # . O . O O O O . O O O O O O O O O O plays at 1 Move 1 does not disturb because it does not add a ko stone to a single ko stone. (Likewise, #'s recapture will not be a disturbing move, either.) *** (3) treats stable disturbing death. Only the player that has at least one ko coupling's string with exactly one breath can be the one to play some disturbing move at all. Playing a ko stone in the ko position is a disturbing action. It disturbs the opponent's immediate possibility of removal of the own strings. . # # # # (e) # # . # O O O # O . O O O O O stable disturbing death . # # # # # # 1 # O O O . O . O O O O O O plays at 1 Move 1 is disturbing because it is a ko stone played in a stable disturbing death and by the player who has a string of the ko coupling with only one breath. (In (e) # cannot play any move that would be disturbing. A # move there would just dissolve the ko position.) . # # # # (f) # # O # O O O . O . counter-example: O O O O O instable disturbing death . # # # # # # O # . O O . O 1 O O O O O # plays at 1 Move 1 does not disturb because the ko position is instable before. Gaining stability is not a disturbing action. [As can be seen, Ing had a preference for considering immediate recaptures in a basic ko mouth. Newko is not ready yet for rules, however, Ing was careless about what is or is not a presupposition. In his intuition so called invariation in a basic ko mouth was somehow presumed. He then saw the behaviour in a disturbing death with a ko coupling, e.g., as a consequence, where due to the basic ko invariation move 1 was just a natural reply to the O move in (e) and should not be recaptured immediately. Without such an intuitive presupposition it is not easily clear why exactly move 1 shall not be disturbing, too. Ing saw a shape pattern and wanted it to have a certain desired outcome, so he went ahead and ruled accordingly.] *** (4) is about single disturbing death. Some player has one or more strings adjacent to the ko mouth with only one breath. His ko stone move is disturbing if it is his only possble move to regain at least two breaths for all his adjacent strings. The ko breath is his last resort. O # . # . (g) O O # # # . O # # . single disturbing death O . 1 # . O O # # # . O # # . O plays at 1 A O suicide move would leave the O string even without any breaths, so the ko stone move is O's only choice to provide his adjacent string with a second breath. This is a disturbing move. O . O # . (h) O O # # # counter-example: . O # # . single disturbing death O 1 . # . O O # # # . O # # . # plays at 1 # might have used a pass play as his move to provide his adjacent string with two breaths! Thus the ko stone move 1 is not his only move to have two breaths for his string adjacent to the ko mouth. Hence 1 is not disturbing. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (9) Date: Wed, 15 Dec 1999 17:56:52 +0100 From: Robert Jasiek Newsgroups: rec.games.go After some preparations we are now ready to discuss rules. Newko has four rules: 1) basic ko 2) fighting 3) disturber 4) prohibition (2) applies to each fighting ko, (3) applies to each disturbing ko, (4) gouverns play through several ko positions on the board. Ing was not a mathematically skilled person and thus failed to do better than to hide the rules (1) and (4) besides further connotations in the one word invariation. He was a better PR-man: "Ko prevents invariation!" *** The basic ko rule: "A single ko stone that has captured a single ko stone may not be recaptured immediately." A recapture must be preceded by at least one board play or pass play of at least one player. The rule is only relevant for basic kos, sometimes called standard kos. Since a basic ko normally is a fighting ko, you may wonder why the basic ko rule is needed at all. The reason is that basic ko mouths can be part of disturbing kos with a ko coupling and that the disturber rule does not include the basic ko restriction. Example: # # # # # (1) . # . # O O O # O . O O O O O disturbing life After a ko stone capture in one ko mouth by one player the basic ko rule prohibits immediate recapture in the same ko mouth by the other player. Thus the opponent must instead play a ko stone capture in the other ko mouth. Then the first player is also confronted with the basic ko rule and will then prefer to pass in the above ko position. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Ing-ko Lecture (10) Date: Thu, 16 Dec 1999 10:00:18 +0100 From: Robert Jasiek Newsgroups: rec.games.go Within each fighting ko an extra rule must be considered with Newko: "If in a fighting ko a ko stone is played, then the opponent's next move may not play a ko stone in it." Ing91 uses complicated hot stones instead. The effect is almost the same, however, eternal life and triple ko with double ko stones are treated slightly differently. Considering that both positions could be treated by the easy fighting rule above, they should be called exceptions under Ing91. It remains unclear why Ing wanted matters to be several times as complex as necessary here. Example: # # # # # (1) . # . # O # O # O . O O O O O fighting ko # # # # # . # 1 # O # O a O b O O O O O O plays at 1 Move 1 is a ko stone in the ko position, so the fighting rule prohibits # to use his next move to be a ko stone at a or b. # must first play tenuki or pass before capturing the next ko stone in the ko position. *** I refrain from giving more fighting ko examples because all further examples (other than basic ko) would be complicated. However, you may of course still ask for trouble... -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (10) Date: Thu, 16 Dec 1999 13:01:24 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 Les Fables wrote: > Coming back to the circular ko As to the fighting rule it is quite easy. > we have two independent fighting kos here, right? Yes. > For the final result (whether or not eternal "windmilling" will be > allowed), we will have to wait for a board repetition rule of some > kind. Yes, you got it! This will be the prohibition rule. So you have to be a little more patient... -- robert jasiek Subject: Re: Ing-ko Lecture (10) Date: Thu, 16 Dec 1999 14:32:39 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 Tim Hunt wrote: > So far you have been telling us about individual ko positions. At some > point it would be nice to see a worked example of a full board position, More than one ko position on the board requires the prohibition rule. I have to treat that first. One ko position in a full board position is no big problem now, unless it is a disturbing ko, because I have still to treat the disturber rule. However, setting full board diagrams is a lot of work, so I prefer you to present me with all the full board diagrams that I shall comment on. Commenting is time consuming enough:) > perhaps one from a real game where these rules made a difference. I guess, BS or BP can easily construct me such question diagrams. I could as well, OC, but please see above:) > I don't know if such a thing exists. Most obviously: Japanese rules have the no outcome exception, e.g. > Could I also ask: what was Mr Ing trying to achieve when he invented > these rules? What are these rules tring to achive? My guesses: 1) Ing wanted to exclude the possibility of long cycles even in theory. 2) Ing wanted to exclude the possibility of long cycles at least in practice. 3) Ing wanted certain precedents to have a desired behaviour or outcome. 4) Ing thought that investing a few millions in the art of strange and wonderful go board shapes (by means of criticism on his late rules) was cheaper than buying famous paintings. 5) Ing deceiptively thought that his ko type classification would have been complete and sufficient. 6) Ing did not realize that superko positions can be analysed just by ko threat variation most of the time in practice and wanted to have rules of that he was surer about this behaviour. 7) Ing thought of disturbing kos as some sort of cooled down local shapes while the real fight takes place outside them. > First: area scoring with some sort of superko rule. Almost never the variant of the superko rule makes a difference in global behaviour, however, local internal sequences can differ considerably. Within these Ing-ko threads I shall restrict analysis to positional superko, for simplicity. > Here, before the > game ends every ko can be played out at no cost so by the end of the > game there are no unresolved kos left and it is easy to count. Ok, we may discuss positions before a game stop or after a game stop. > Second: Japanese rules. For simplicity, I shall assume that game stop equals game end and that there is no confirmation phase. > A, B I comment after the disturber rule. > B) This is some sort of a seki. What happens under superko? does someone > end up captured? If so I would consider it a slight shortcoming in > superko rules. This kind of seki is nice! Are you another Ing who desparately wants a desired outcome for a particular position? ;-) No need to worry though, double ko is a coexistence under all regular superko variants, even the natural situational superko! >> . # O . O . O . O # # # # O O O O O O # . # O O # # # # # # # O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . C) Here black is dead. Presumably this is the same under superko? << There is no problem whatsoever. O just fills a ko mouth as soon as possible and we get (D). > D+B, E, E+B I will comment after the prohibition rule. > My limited understanding of Japanese rules is that one should try to > treat different groups separately. O please, I need some rest after a month of J89 discussion;( (J89 are too imprecise to be very sure about local or global consideration.) -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (10) Date: Thu, 16 Dec 1999 21:14:31 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 Dear Bill, > Neither these rules nor the Ing rules allow players in superko > sequences to choose where to accept the ko ban and play a threat. True, but it is not the primary aim of rules to get the last bit of extra variation. After all the function of repetition rules is to restrict variation. It may be ironical that Ing apparently wanted maximal finite variation or that a fixed ko rule would probably serve best for that purpose. However, for practical repetition rules we need a handily wordable rule more than Conway-like mathematical simplicity. Superko may be well enough, however, whether or not it provides us with certain special extra yose possibilities is more a nice side-effect than an ultimate aim. OC, for theoretical studies one can choose whichever repetition rule deems best. -- robert jasiek http://www.snafu.de/jasiek/rules.html Fix the ko rule problem - choose the fixed ko rule! Thereby you fix all ko yose trouble at once;) Subject: Ing-ko Lecture (11) Date: Thu, 16 Dec 1999 17:30:25 +0100 From: Robert Jasiek Newsgroups: rec.games.go Within each disturbing ko the disturber rule is applied. Ing91 is very imprecise here and hardly says anything about allowed move-sequences. Newko fills the gap. The following is the crucial part: "During a ko stone move-sequence in a disturbing ko a player becomes the disturber of the disturbing ko by disturbing it first. As soon as the ko position of the disturbing ko with the stones on it has been repeated since the moment just before the occurrance of the current disturber, he is not allowed to disturb the disturbing ko. With the end of the ko stone move-sequence in the disturbing ko a disturber ceases to exist." The rule talks about "the disturber", i.e. at any one time only one player can be a disturber. Thus it is necessary to end a disturber's existence so that the other player can also become the disturber later. The Ing91 booklet makes it clear that some ko positions can have either player as the disturber. In stable disturbing lifes this is especially fair considering the symmetries of typical shape patterns. Since the disturber rule shall restrict play within a disturbing ko, it must restrict sequences of plays within some. Any non-ko-stone move within a disturbing ko would dissolve the ko position or at least one ko mouth of it. So it is only ko stone moves that need be restricted. Thus the disturber rule considers ko stone move-sequences that take place within one disturbing ko. Since a disturber's existence must end at some time, it is quite natural to use the end of the ko stone move-sequence within the disturber's disturbing ko for this purpose. I.e. as soon as a non-ko-stone move is played or a ko stone move in a different ko position is played or a pass play is played the disturber of the disturbing ko in question ceases to exist. Disturbing in a disturbing ko, as explained earlier, can be a cause for becoming the disturber of the ko position. However, only one player can be the disturber, so as soon as there is already the disturber of the particular disturbing ko the opponent does not become a disturber even if he plays a disturbing move within in that disturbing ko. Also the disturber does not become a double disturber or so of one ko position. The first player to disturb in a particular disturbing ko when there currently is no disturber for it becomes the disturber. Only a disturbing move can create the disturber. A ko stone move in a disturbing move that does not disturb it does not yet create the disturber. So from an initial instable state to the first stable state the move in such a prologue does not disturb or create the disturber yet. The disturber rule does not restrict the number of times a particular player may become the disturber of a particular disturing ko. Rather this problem is handled by the prohibition rule to come. The disturber rule restricts play within the disturbing ko of its disturber for him and only for him. From the moment just before the occurrance of the current disturber on the disturbing ko is watched for restriction. A possible prologue does not count towards the observation. Only when a disturbing ko was already in a stable state and from it the one player disturbed and thereby became the disturber positions of the disturbing ko are controlled. This includes the position just before the occurrance of the disturber. Watching those positions becomes relevant as soon as since the first watched position a repetition is arrived at, a repetition of the ko position's ko stones. So That repetition may still arise. Then and only then the disturber's play is restricted: He is then prohibited to disturb in the disturbing ko, i.e. to play another ko stone from the then stable ko position's state. Simply speaking, after one cycle of disturbing immediately continued recycling is prohibited for the time being. The disturber must first cease to exist by playing tenuki, a non-ko-stone move, or a pass play before possibly being allowed to disturb in the same disturbing ko again. However, there will also be the prohibition rule. *** Examples: t = tenuki p = pass a, b, ... = ko mouths # # # # # (1) . # a # O O O # O b O O O O O disturbing life a-b Newko: O[a] O becomes the distuber of a-b O[ab] legal O[abb] prohibited by basic ko rule O[abp] O ceases to be the disturber of a-b O[abpppp] a legal game end O[pppp] a legal game end O[abpa] # becomes the disturber of a-b O[abpabp] # ceases to be the disturber of a-b Positional Superko (PSK): O[pp] perfect play O[abpap] O loses all O[abpab] prohibited by PSK # # # # # (2) . # O # O O O a O b O O O O O disturbing life a-b Newko: #[a] legal prologue #[aa] prohibited by basic ko rule #[ap] legal #[apppp] a legal game end #[appa] O becomes the disturber of a-b #[appaa] prohibited by basic ko rule #[appabp] O ceases to be the disturber of a-b For an analysis of all possibilities the prohibition rule is required as well. It will leave the double ko on the board, whatever happens. PSK: #[app] perfect play #[apbapp] perfect play # # # # # # # (3) a # b # O # O # O # O c O d O O O O O O O disturbing life a-b-c-d Newko: #[c] # becomes the disturber of a-b-c-d #[cbdcbd] first repetition since just before disturber's birth #[cbdcbdc] prohibited by disturber rule #[cbdcbdd] prohibited by disturber rule and by basic ko rule #[cbdcbdp] # ceases to be the disturber of a-b-c-d #[cbdcbdt] # ceases to be the disturber of a-b-c-d #[cbdcbdpa] O becomes the disturber of a-b-c-d #[cbdcbdta] O becomes the disturber of a-b-c-d In a quadruple ko longer cycles might occur until a repetition arises. Always the basic ko rule is intact, too. Superko: Excercise left for BS, JR, or JD :) Japanese before game stop: void game Japanese after game end: One player ought to be demoted;) (Actual analysis is beyond this thread.) # # # # # # # (4) a # O # O # O # O b O c O d O O O O O O O disturbing life a-b-c-d Newko: #[b] prologue #[ba] O becomes the disturber of a-b-c-d #[bacbacb] first repetition since just before disturber's birth #[bacbacba] prohibited by disturber rule #[bacbacbp] O ceases to be the disturber of a-b-c-d #[bacbacbtc] # becomes the disturber of a-b-c-d Japanese before game stop: # to play gives void game. . # O b O # . . . . (5) # # # O O # . . . . a # O O # # . . . . # O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . disturbing death a-b Newko: #[b] dissolves all O[a] O becomes the disturber of a-b O[abp] O ceases to be the disturber of a-b O[abpa] dissolves all PSK: O[abpa] dissolves all . # O . O . O # . . (6) # # # O O O O # . . . # O O # # # # . . # O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . disturbing life Analysis for Newko or PSK as for (1). . # O . O . O . O # (7) # # # O O O O O O # . # O O # # # # # # # O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . Ing91: disturbing death Newko: Formally it is still a stable disturbing life but O will reduce it to a disturbing death. In practice there is hardly a difference to Ing91. x # O O O . O . O # (8) # # # O O O O O O # a # O O # # # # # # # O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . single disturbing death Newko: O[a] NOT a disturbing move O[apx] Superko: O[apx] x # O O O . O . O # (8) # # # O O O O O O # O # O O # # # # # # a O O # # . . . . , O O # # . . . . . . # # # . . . . . . . . . . . . . . . . . single disturbing death Newko: #[a] # becomes the disturber of a #[aa] prohibted by basic ko rule #[ap] # ceases to be the disturber of a #[appa] NOT a disturbing move #[appapxpppp] a legal game end Superko: #[atpapxpp] t saves O the possible removal under 2-pass rules . # # . O (9) # O O O O a O O O O O O O O O single disturbing life #[a] # becomes the disturber After another capture and recapture # ceases to be the disturber with the following pass play. -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (11) Date: Fri, 17 Dec 1999 11:04:27 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 ±èÇöÁø wrote: > how many word'disturb'wer e usded in his post? In general, Newko uses disturb disturber disturber rule disturbing ko disturbing life disturbing death stable disturbing life instable disturbing life single disturbing life stable disturbing death instable disturbing death single disturbing death Disturbing:) -- robert jasiek Subject: Ing-ko Lecture (12) Date: Fri, 17 Dec 1999 13:43:50 +0100 From: Robert Jasiek Newsgroups: rec.games.go Ing91 says virtually nothing about play through several ko positions. Newko uses the prohibtion rule for this purpose, which is slightly simplified as follows: "From all board positions with the same set of board points of all ko positions and with the same configuration of stones on the board without the set each player may only once play a ko stone in each fighting ko or become disturber of each disturbing ko. As an exception for all such board positions the first played ko stone in a fighting ko is ignored." The rule classifies all board positions. Herefore each board position is partitioned into two sets: 1) The set of all board points of all ko positions. 2) The rest of the board. Board positions are in the same class if A) The set (1) is the same. B) The set (2) shows the same configuration of stones. Thus board positions in the same class are different only within their ko positions. Example: . . . . . . . . # O . . . # . # O . . . # O O # . . . O . O (X) . . . . . . . . # O . . . # . # O . . . # O O # . . . O # . (Y) Both X and Y belong to the same class of board positions because in both the set of ko positions @ is the same: . . . . . . . . # O . . . # @ @ O . . . # O O # . . . O @ @ X and Y differ only within their ko positions. The rest of the board position is the same. . . . . . . . . # O . . . # . # O . . . # O O # # . . O . O (Z) Z does not belong to the class of X and Y because of the additional stone on the rest of the board. *** For each class of board positions play is restricted. Each player has his own restrictions. They permit him one and only one own action in each ko position (with the class of board positions being the same). An action can be I) in a fighting ko: playing a ko stone in it, II) in a disturbing ko: becoming its disturber. *** The prohibition rule has its exception: For each class of board positions the first played ko stone in a fighting ko is ignored. Thereby a fighting ko retains its fighting character even if play goes through several ko positions. *** Examples: . . . . . . . . . (1) # # # # # # # # # O O O O # O O O O # O . O # O . O # . # O O # O O # . # O . O # O . O # . . . . . . . . . two stable disturbing deaths # # # # # # # # # O O O O # O O O O all board points @ in ko positions @ O . O # O . O @ @ # O O # O O # @ playing only ko stones will leave the set @ # @ @ O # O @ @ # the same . . . . . . . . . # # # # # # # # # O O O O # O O O O # O . O # O . O # involved ko mouths a,b,c,d b # O O # O O # d # O a O # O c O # #[a] # becomes the disturber of a-b #[abc] # becomes the disturber of c-d and ceases to be the disturber of a-b #[abcd] legal #[abcdb] prohibited by prohibition rule because # may not again become the disturber of a-b #[abcdd] prohibited by basic ko rule Thus # loses all his corner stones. . # # . O (2) # O O O O . O O O O O O O O O single disturbing life b # # . O c O O O O a O O O O involved board points O O O O O #[a] # becomes the disturber #[abcp] # ceases to be the disturber #[abcpa] prohibited by prohibtion rule because # may not again become the disturber of the same disturbing ko #[abcpppp] a legal game end # # # # # # (3) . # . # O # O O # O . O O O O O O O stable disturbing life and single disturbing death # # # # # # . # . # # # # # # O O O # # O . O . # # # # # # . # b # O # O O # O c O involved ko mouths a,b,c O O O O O O # # # # # # . # . # # # # # # O O O # # O a O . #[ab] O becomes the disturber of b-c #[abca] O becomes the disturber of a and O ceases to be the disturber of b-c #[abcappa] legal #[abcappac] prohibited by the prohibition rule because O may not again become the disturber of b-c # # # O O O (4) # # # O O O . # a # O . # # # O O O O O O # # # . O O # # . O . O # . # three fighting kos a,b,c O O O # # # # # # O O O . # b # O . # # # O O O O O O # # # . O O # # . O . O # . # O O O # # # # # # O O O . # O c O . # # # O O O O[attatt..] perfect play; legal because tt change the rest of the board and thus leaves a prior class of board positions O[acbacb] legal because so far each player has played a ko stone in each fighting ko only once O[acbacba] legal because the prohibition rule ignores O[a] as the first ko stone in a fighting ko O[acbacbatt..] legal; a ko fight may proceed with such nasty long ko stone move-sequences, however, the first sequence is possible as well and concentrates on the most valuable ko position # # # # # # # # # (5) # # b # c # O # O O O # O # O d O e O O O O O O O O O # # # # # # # # # . . # O O . . . . . # a # O . O . . . . # O O . . . . . . . . . . . . . O[adbadb] legal; so far each player has only once played a ko stone in the fighting ko a and become the disturber of b-c-d-e O[adbadba] legal due to the exception of the prohibtion rule O[adbadbat] finally a regular ko threat t is required O[at] this would have the same effect as using symmetrical ko stone ko threats in b-c-d-e -- robert jasiek http://www.snafu.de/jasiek/rules.html Subject: Re: Ing-ko Lecture (12) Date: Sat, 18 Dec 1999 15:11:24 +0100 From: Robert Jasiek Newsgroups: rec.games.go References: 1 , 2 Les Fables wrote: > Applied to the circular ko > # O . > . # O 3x3 version > # O . > this means that black's starting a fighting ko leads to two roundabouts > [...] until the initial position is repeated for the second time. Then > black [...] has to find a ko threat in order to be allowed to continue > the ko. Correct. (This is a strange case.) *** Tim wrote: >> . # . O # . . . . . # O O O # . . . . . . O # # # . . . . . O O # . . . . . . . # # # . . . . . . . . . . . . . . . . . . . . . . . . . . . << a d e O # . . . . . involved board points b O O O # . . . . . c O # # # . . . . . a-b-c and a-d-e are two single disturbing lifes O O # . . . . . . . # # # . . . . . . . there is no ko coupling because the O string and . . . . . . . . . . the string d are not also both adjacent to the . . . . . . . . . . ko mouth a-d-e ;) #[cabpeadpapc..] is a legal sequence because # may become the disturber in each ko position once. -- robert jasiek http://www.snafu.de/jasiek/rules.html