Hi Bill,Bill Spight wrote:OK, Jika, let's get down to basics.
How much territory does White have?
Thank you for your efforts, Bill. The writings about endgame values are very confusing, and it is very helpful to see that the obvious inconsistencies are not just a problem with our understanding.Bill Spight wrote:Ever since getting on rec.games.go around 25 years ago I have tried to dispel this confusion. To use gain to explain swing values, as you and Ed both have done, without noting that for gote there are two gains involved, one gain made by Black when she moves first, and one gain made by White when he moves first, makes it difficult for a beginner to avoid confusing swing values with gains, it seems to me. That is why I have asked you both not to do so.
(and maybe drive you all nuts) a new example from 321go:...in other situations I don't understand how big a territory is, probably because I don't know what will be territory in future.
But you forgot that the eye that Black is getting is a false eye "by the second line". This intersection will eventually be filled. So the capture just gets you one point thanks to the prisoner, and no point of territory.Jika wrote:They say, the value of playing at a is 2.
I would have said so intuitively, because black can capture a stone.
Right, except that this is now 1-(-1) = 2.Jika wrote:But... then I remembered that I must keep in mind that by doing so, he prevents white from getting 1 point.
So, should the answer not be 2-(-1)=3?
The intersections that change something must be counted. If you don't know if the triangled intersections will be affected, then you can't count the value of the endgame yet.Jika wrote:And I have no clue how far to count (see triangled intersections)
But would not white win a prisoner too when beating black?Pio2001 wrote:But you forgot that the eye that Black is getting is a false eye "by the second line". This intersection will eventually be filled. So the capture just gets you one point thanks to the prisoner, and no point of territory.Jika wrote:They say, the value of playing at a is 2.
I would have said so intuitively, because black can capture a stone.
So, if white could beat back, white looses mathematically nothing, 1 eye point for 1 prisoner, 1-0 =1 ?Right, except that this is now 1-(-1) = 2.Jika wrote:But... then I remembered that I must keep in mind that by doing so, he prevents white from getting 1 point.
So, should the answer not be 2-(-1)=3?
Yes, they are stable, but I do not know if they have to be counted now, or only when they are closed, or... counted as half a point or... whatnot...The intersections that change something must be counted. If you don't know if the triangled intersections will be affected, then you can't count the value of the endgame yet.Jika wrote:And I have no clue how far to count (see triangled intersections)
If the status of these intersections is unclear to you, you can try a variant of the game of go against the artificial intelligence in this page : https://www.crazy-sensei.com/location=rules
In this variant, instead of counting empty intersection plus prisoners, you just count the number of stones on the board. As long as you can add more stones, you go on, until you can't add anymore stone without being killed.
This will clearly show you why the triangled intersections are stable.
Well, that's how much secure territory that White has that cannot be taken away, even if White passes all the time. (Unless White fills his own eye, OC.Jika wrote:Hi Bill,Bill Spight wrote:OK, Jika, let's get down to basics.
How much territory does White have?
I would have said "2".
) So that's a good answer. 
Another, practical answer is how much territory can White defend if Black plays first? Since you get Ed's answer, you see that White can defend 3 pts. of territory.I get Ed's answer.
But I never know how many possible future moves to think ahead.
In this case, there are not many possible moves, and I understand why there is miai.
Join the club.But in other situations I don't understand how big a territory is, probably because I don't know what will be territory in future.
De nada.Pio2001 wrote:Thank you for your efforts, Bill. The writings about endgame values are very confusing, and it is very helpful to see that the obvious inconsistencies are not just a problem with our understanding.Bill Spight wrote:Ever since getting on rec.games.go around 25 years ago I have tried to dispel this confusion.
The traditional "rule" is incorrect. In terms of gains, it is the reverse sente that gains something. Playing a local sente with sente gains nothing, in terms of points. That is, making the threat and answering it gain the same amount, so their gains cancel out. All we know about how much they gain is that it is more than how much the reverse sente gains. That is why we may assume that the sente will normally be played, since the sente player can normally play it with sente before her opponent can afford to play the reverse sente.Even today, I feel confused, even about the modern "miai values". For example, when I compare a sente move with a value of 1 point and a gote one with a value of 20 points.
If I apply the traditional rule that says that a sente move is worth twice a gote move, and even if I use the modern "tally" calculation, all the theory says that the gote move is worth much more that the sente move, while it is obvious to anyone that it is better to play the sente first, in order to get both.
For those who are confused about this detail - have you ever noticed that when pros or top amateurs estimate the score, they always pre-subtract out the sente pushes and other sente intrusions, even if the opponent hasn't played them yet? (See any number of online Go commentary videos, or if you've been to a Go club or Congress you may have seen strong dan players or pros do this in-person). Pretty much every strong amateur or pro that I've ever encountered counts this way.Bill Spight wrote:Playing a local sente with sente gains nothing, in terms of points.
The simple answer is that you might be mistaken about what "modern" counting theory saysPio2001 wrote:Even today, I feel confused, even about the modern "miai values". For example, when I compare a sente move with a value of 1 point and a gote one with a value of 20 points.
If I apply the traditional rule that says that a sente move is worth twice a gote move, and even if I use the modern "tally" calculation, all the theory says that the gote move is worth much more that the sente move, while it is obvious to anyone that it is better to play the sente first, in order to get both.
. There are TWO important values for locally-sente moves, not one - the reverse sente value (also known as the miai value, or the per-move gain), and the threat value - the gain of the followup move if unanswered.Do you mean those?Follow-up question:
How much territory does White have, on average, at each of the 1-3 points in the top left corner?
Now after a night's sleep I'm getting it!But you forgot that the eye that Black is getting is a false eye "by the second line". This intersection will eventually be filled. So the capture just gets you one point thanks to the prisoner, and no point of territory.
...
Right, except that this is now 1-(-1) = 2.
Very good.Jika wrote:@Bill, post 53:
Do you mean those?Follow-up question:
How much territory does White have, on average, at each of the 1-3 points in the top left corner?
a: 1
b: 1/2
c: 1/2
Correct??
That's fine.(PS Sorry that I'm using he and she opposite to you - I'm reading "How to play Go", and they call white "she"; it would be confusing to use the opposite of what I'm reading.)
We are all female under the skin. 
(A little biological joke. Sorry! I did not understand "1-3" was go-speak, but thought it meant "the value of the points 1 to 3" (therefore my question: which three points)I meant b and c, which may be called 1-3 points or 3-1 points. a is a 1-1 point.
