Gérard TAILLE wrote:Bill Spight wrote:
I am a bit surprised that you had nothing more to say, considering that you constructed a whole board that included what you want to call a double sente. If I had asked about, say, {3|-4}, I thought that you would call it a gote, but now I'm not sure.
Your question is now a little controversial Bill.
It is not only what is my understanding of {6|-1||-5} but it is far beyond it is how I characterise such area, gote or sente or ...
Well, double sente has been controversial for at least 45 years. I don't think that gote and sente are particularly controversial.
One thing is that these terms have been around for a long time, and like words in general, have acquired different senses. For instance, if I have made a play
with sente, my opponent has answered it, and we may say that my play was sente. But suppose that he should not have answered it. Then we may say that it was really a gote. It was a sente because he answered it, but it was a gote because he should not have.
We also talk about taking sente or taking gote, which use still different senses of the terms. Players pick up these different senses without consulting a dictionary.
But when we talk about estimating the value of a position or of a play, we use the terms, sente and gote, in a technical sense, and they need to be precise and clearly defined. Unfortunately, in the texts that I and many others learned from, these technical terms were not precisely and clearly defined. Instead we were taught that if a position was a double gote, we estimated the territory as the average of the two resulting positions, but if the position was a sente, we estimated the territory as that of the result of the sente sequence. OC, this does not make sense for double sente, since there are two different sente sequences with different results. In any event, no estimate for the territory of a double sente position was ever offered.
It is a controversial question because anybody may define gote and sente as she wants.
Informally, sure. Like Lewis Carroll's Humpty Dumpty:
A word means just what I choose it to mean.
Because it is basically a question of definition it is impossible to say which definition is the best.
That doesn't work for technical terms. They are not a question of definition. If two technical definitions do not produce the same results, then at least one of them is wrong, or they belong to different theories.
I am completely open to any definition proposed and I can of course discuss with anybody, taking any definition. In the other hand I expect also that we can discuss taking another definition.
I understand Robert's definition of "double sente" and I agree at 100% that with such definition "double sente" does not exist. What can I say more? I agree, I agree, I agree.
I expect that you understand the theory behind his definition.
Now I proposed for discussion another definition of "double sente" and the answer is : No, no no, this is not the good definition, the good definition is ... and with this definition "double sente" does not exist. How can we discuss?
What is the theory behind your definition?
AFAICT, there is no theory behind the common usage of
double sente. Some rather horrible examples exist in textbooks.
I was kibitzing a pro game with Jiang Jujo, 9 dan, several years ago, and one of the players, in the endgame, did not answer one of the textbook examples of a double sente. I nudged Jujo and pointed out, as a joke, that the player did not answer that play, thinking that Jujo knew that I knew that he knew that I knew . . . that the play was not really a sente. Jujo just looked at me like I was crazy.
Yes Bill I see you have open the door and you even use yourself double sente to characterise certain situations. Fine.
Nobody I know disputes that there are double sente plays, depending on the global situation. The question is whether there are double sente positions, independent of the rest of the board. The textbooks showed us such positions, independent of the rest of the board, and claimed that they were double sente. Were they right, that's the question. Kano was plainly wrong, dead wrong. O Meien doesn't do that. Yes, as John Fairbairn points out, he mentions double sente in reviewing a game, in the global context of the whole board. No problem.
He doesn't isolate the plays and claim that they are double sente.
First of all yes Bill I can easily call {3|-4} a gote. In absence of follow-up by definition it is for me a gote.
It has to be local gote because the second player has no local reply.
But do not forget it is only a definition. We all know that a couple of miai gote may act as ko threat and you may feel such gote point being sente.
Sure, but that's a different game, such as {3|-4} + {3|-4}, which is equal to -1. But unlike a simple point of White territory, either player may play in one of the miai pair
with sente. Or, in a ko fight, if the opponent ignores the threat,
with gote. Because of such possibilities, I classify the miai pair as
ambiguous. Technically, you can apply any of the terms, gote, sente, or ambiguous to the situation and get the same result. That's OK. They all fit into the same theory.
You can even call the miai pair a double sente. I don't know anyone else but me who does that, however.
What is the point? If you say that a gote may be a sente that only means you have not clearly defined these terms. As soon as you propose a definition for sente and gote you have no contradiction providing you do not change the definition between two sentences.
So let's call gote an area {x|y} with x > y. It is a definition and nothing else.
Well, it can't be sente by itself. I.e., it cannot be intrinsically sente. It may be played with sente in certain situations, but that's different. That's accidental, in a philosophical sense.
Gote points have very important characteristics: they are comparable, the evaluation (x+y)/2 allows us to tell which gote is the best one and you can proof that by playing the gote in the order given by this evaluation you are always correct.
You can also prove it as Nogami/Shimamura did by showing that {x|y} + {x|y} = x + y, when x > y.
What about sente. My definition is : an area {x|y||z} is sente for black if x>y>z and (x+y)/2 > y-z
Here again it is only a definition and nothing else.
BTW, I think that you mean (x-y)/2 > y-z. It's a definition that fits the theory.
(And, FWIW, it is the same one that I derived way back when. Some friends thought that I should publish it, but I thought that it was fairly obvious. In retrospect, maybe I should have tried.) For kos, there are different theories, any of which may or may not apply to any particular situation. But for these types of plays and positions CGT evaluates non-ko positions in line with traditional go theory up to and including O Meien's approach, if in some cases more precisely. Traditional go theory did not evaluate double sente positions, and CGT does not need the term.
In particular I do not claim that white has to answer immediately to a black move. It may be the case in a lot of practical cases but it is not part of the definition.
Nor is it a part of traditional go theory or CGT. I remember as a shodan explaining to people why a certain play was sente. I pointed out that the threat of the sente was larger than the reverse sente, so that
if there were other plays on the board that were smaller than the threat and greater than the reverse sente then the player whose sente it was would be able to play the sente before the other player could play the reverse sente. (I was aware that there were exceptions.
) That is why, I continued, we assume, as a rule, that the sente will be played. Traditional go theory calls that phenomenon the
privilege of a sente. In thermographic terms, the privilege is indicated by a colored mast.
Now what about your {6|-1||-5} example. According to my definition it is neither a gote nor a sente.
Right. Your definitions are too narrow. The theory encompasses more general definitions. In terms of them the average value of this position, m = (6 - 1 - 2*5)/4 if and only if -1 ≥ m ≥ -5. m = -1¼, so that condition is met. Inequality holds, so that also means that the position is technically, intrinsically gote.
However, given the other positions on the board, correct play is for Black to play in this position
with sente. We may say that it is globally, accidentally, sente. Even better, perhaps, is to say that
the position is gote, but
on this board the move is sente.
In fact depending of the circumstances it may have the behaviour of a gote or a sente.
Accidentally.
In any case such area do not have the characteristic of gote area: you can calculate an evaluation like ((6+(-1))/2) + (-5))/2 of course but this time this area can be incomparable to true gote, and you cannot be sure that the calculated evaluation will allow you to play your yose in the best way.
Sure it has gote characteristics. It does not raise the local temperature when played with gote. In fact, it lowers it. That is an essential characteristic of gote, in the theory. Furthermore, {6|-1||-5} + {6|-1||-5} + {6|-1||-5} + {6|-1||-5} = -5. Only gote and some ambiguous positions have the characteristic that a finite number of them equals a number.
Concerning the area {5|0||-2} it is a sente according to my definition. If, taking my definition, the theory tells me that black will very often (against an ideal environment?) be able to play in this area with an immediate answer by white it is fine. It is a good characteristic even I know it is only an "average" behaviour.
Because this is a technical, intrinsic sente, four of them will not equal 0, which is its mean value. If Black plays first in the foursome, the correct result will indeed be 0. But if White plays first the correct result will be -2. The average value, m, will satisfy 0 ≥ m ≥ -2/4 = -½. This result if White plays first approaches m in the limit as the number of instances goes to infinity. That is so for all intrinsic, local sente. (I used to think that that characteristic defined sente, but it doesn't.)
Gérard TAILLE wrote:Now I can propose a definition of a "double sente", oh sorry a "double blabla"
a "double blabla" is an area {a|b||c|d} with a>b>c>d and (a-b)/2 > b-c and (c-d)/2 > b-c
Surely if the theory analyses a "double blabla" it will prove it is a quite hot point with interesting characteristics etc. etc. and it may even help to choose the correct order for playing them with good chance to find the best one etc.
Here again nobody knows if a play in this area will be answer immediately by the opponent. It is not in the definition is it?is a difficult one isn't it?
It appears that you are making a definition based upon analogy with your definition of sente for {x|y||z} as (x-y)/2 > (y-z)/1. (I supplied the divisor of 1.) Is there a theory behind that definition? In
The Endgame by Ogawa/Davies, Davies points out that (according to traditional go theory) the divisor comes from the net number of plays between the two results. Between x and y there are two net plays, and between y and z there is one net play. However, for {a|b||c|d} the net number of plays between b and c is 0. Davies points out that the divisor for (b-c) should therefore be 0.
Edit: This is a long note. Let me highlight a major point.
Now I proposed for discussion another definition of "double sente" and the answer is : No, no no, this is not the good definition, the good definition is ... and with this definition "double sente" does not exist. How can we discuss?
It's not you.
It's the old textbooks.
I repeat, adding emphasis:
Nobody I know disputes that there are double sente plays, depending on the global situation.
The question is whether there are double sente positions, independent of the rest of the board. The textbooks showed us such positions, independent of the rest of the board, and claimed that they were double sente. Were they right, that's the question. Kano was plainly wrong, dead wrong. O Meien doesn't do that. . . . He doesn't isolate plays and claim that they are double sente {in themselves}.
