Thermography

Bill Spight · Post by **Bill Spight** » Tue Dec 01, 2020 5:33 am

Gérard TAILLE wrote:
Bill Spight wrote: It is not that this Three Points without Capturing position is worth -2 by territory scoring, it is worth -2 by the Japanese 1989 rules. By other territory rules it is worth -3, as advertised.
Yes Bill. Assume I am using J89 rules. How I translate to the CGT game the fact that the Three Points without Capturing position is worth -2 rather than -3?

Wing it.

Gérard TAILLE wrote:Concerning seki let's take a more general position with eyes and dead stones:

$$B Seki
$$ -----------------
$$ | C B W C W W C |
$$ | B C W W X W B |
$$ | W W W X X X W |
$$ | X X X . X W W |
$$ | . . . X X W C |
$$ | . . . . X W B |
$$ | . . . . X W C |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B Seki $$ ----------------- $$ | C B W C W W C | $$ | B C W W X W B | $$ | W W W X X X W | $$ | X X X . X W W | $$ | . . . X X W C | $$ | . . . . X W B | $$ | . . . . X W C | $$ -----------------[/go]
Area counting : if I understand your normalization I count 15 white stones against 4 black stones and I normalized the position to -11.

Well, you subtract the stones, so to normalize you add 11 to the local area score.

Gérard TAILLE wrote:Thus the area looks like a game of the form -11 + G with G being a seki counted 0.

Why count it as 0? It's not 0 in CGT. We're not normalizing to J89.

Gérard TAILLE wrote:As a go player I evaluate the position 3 points for black and 21 points for white for a total of -18.

BTW in territory scoring I count this position +4 (at the very end of the gam, after the last dame point, white can force the capture of 4 black stones).

How translate in CGT game such information concerning the rules used?

Except for the group tax. area scoring normalizes well into CGT.

Let's play this out in CGT at temperature -1.

Click Here To Show Diagram Code: [go]$$Wc White first. $$ ----------------- $$ | C B W 7 W W 1 | $$ | B C W W X W B | $$ | W W W X X X W | $$ | X X X . X W W | $$ | . . . X X W 3 | $$ | 8 . . . X W B | $$ | 6 . . . X W 2 | $$ -----------------[/go]

at

captures

connects

Result: -6 (counting

and

but not the Black territory.

Click Here To Show Diagram Code: [go]$$Bc Black first $$ ----------------- $$ | C X W 8 W W 4 | $$ | X C W W X W X | $$ | W W W X X X W | $$ | X X X . X W W | $$ | . . . X X W 2 | $$ | . . . . X W B | $$ | 5 7 9 . X W 1 | $$ -----------------[/go]

at

captures

connects

Result: -7

So by CGT the seki is worth -7.

Now let's count it by area scoring. It should be -7 - 11 = -18.

Click Here To Show Diagram Code: [go]$$B Seki $$ ----------------- $$ | C B W W W W W | $$ | B C W W X W W | $$ | W W W X X X W | $$ | X X X . X W W | $$ | . . . X X W W | $$ | . . . . X W W | $$ | . . . . X W C | $$ -----------------[/go]

Each player has one point of territory. So let's just count the stones. 2

stones vs. 20

stones. Area score = -18, as advertised.

Gérard TAILLE · Post by **Gérard TAILLE** » Tue Dec 01, 2020 7:56 am

Bill Spight wrote: Except for the group tax. area scoring normalizes well into CGT.

I am a little worried about seeing you mentionned group tax.
Does that mean that in seki with eyes in only one side you have to adapt your count?

Click Here To Show Diagram Code: [go]$$B $$ ----------------- $$ | . O . X . O . | $$ | O O O X O O O | $$ | X X O O X X X | $$ | X X X X X . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]

How do you proceed?

Bill Spight · Post by **Bill Spight** » Tue Dec 01, 2020 9:33 am

Gérard TAILLE wrote:
Bill Spight wrote: Except for the group tax. area scoring normalizes well into CGT.
I am a little worried about seeing you mentionned group tax.
Does that mean that in seki with eyes in only one side you have to adapt your count?

Yes. To eliminate the group tax from his rules Berlekamp had to go to a bit of trouble. IIRC, at some point he had an encore with all remaining stones immortalized.

Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . O . X . O . |
$$ | O O O X O O O |
$$ | X X O O X X X |
$$ | X X X X X . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . O . X . O . | $$ | O O O X O O O | $$ | X X O O X X X | $$ | X X X X X . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]
How do you proceed?

In the seki neither side wishes to play, so the CGT score is 0. But if the stones are immortalized:

Click Here To Show Diagram Code: [go]$$B $$ ----------------- $$ | 6 W 2 B 1 W 4 | $$ | W W W B W W W | $$ | X X W W X X X | $$ | X X X X X . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]

The local score is -2, as with the marked stones immortalized, White can afford to fill the two eyes.

Gérard TAILLE · Post by **Gérard TAILLE** » Tue Dec 01, 2020 10:49 am

Bill Spight wrote:
Gérard TAILLE wrote:
Bill Spight wrote: Except for the group tax. area scoring normalizes well into CGT.
I am a little worried about seeing you mentionned group tax.
Does that mean that in seki with eyes in only one side you have to adapt your count?
Yes. To eliminate the group tax from his rules Berlekamp had to go to a bit of trouble. IIRC, at some point he had an encore with all remaining stones immortalized.
Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . O . X . O . |
$$ | O O O X O O O |
$$ | X X O O X X X |
$$ | X X X X X . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . O . X . O . | $$ | O O O X O O O | $$ | X X O O X X X | $$ | X X X X X . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]
How do you proceed?
In the seki neither side wishes to play, so the CGT score is 0. But if the stones are immortalized:

$$B
$$ -----------------
$$ | 6 W 2 B 1 W 4 |
$$ | W W W B W W W |
$$ | X X W W X X X |
$$ | X X X X X . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ | . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | 6 W 2 B 1 W 4 | $$ | W W W B W W W | $$ | X X W W X X X | $$ | X X X X X . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]
The local score is -2, as with the marked stones immortalized, White can afford to fill the two eyes.

Oops it looks like an adaptation of CGT to specific rules with adding complexity. CGT can be a success on go player community only if it is efficient (OC) but not too complex.
My view is that it is completly unuseful to try to resolve in CGT zugswang positions (I mean positions like seki or three points without capture, which will be definitly resolved after the last dame had been played).
In one part such resolution is highly dependent on the specific rules and will add difficulties on CGT, and in other part these positions are very known to any go player and easily resolved by them.

Do you see a problem if a go player, using CGT, replace systematically any zugswang positions encountered during analasis by the number corresponding to the resolution of this position by the go player in accordance to the rules used?
That way CGT stays not too complex while go players could not be upset by the resolution of these zugswang positions which are, most of the time, quite obvious.

Bill Spight · Post by **Bill Spight** » Tue Dec 01, 2020 11:51 am

Gérard TAILLE wrote:Oops it looks like an adaptation of CGT to specific rules with adding complexity. CGT can be a success on go player community only if it is efficient (OC) but not too complex.
My view is that it is completly unuseful to try to resolve in CGT zugswang positions (I mean positions like seki or three points without capture, which will be definitly resolved after the last dame had been played).
In one part such resolution is highly dependent on the specific rules and will add difficulties on CGT, and in other part these positions are very known to any go player and easily resolved by them.

Do you see a problem if a go player, using CGT, replace systematically any zugswang positions encountered during analasis by the number corresponding to the resolution of this position by the go player in accordance to the rules used?
That way CGT stays not too complex while go players could not be upset by the resolution of these zugswang positions which are, most of the time, quite obvious.

Well, I have a certain fondness for a group tax.

I am also not bothered by variations in the rules of games.

My preference going forward is for Button Go, which can easily be implemented by a small change in the AGA rules, and which has already been used in international play.

All that aside, I think that one major value of CGT is the demonstration that territory scoring is not illogical, as has been claimed and widely believed.

And as practical matters, chilled go infinitesimals, difference games, and ko evaluation are significant advances in go theory.

Edit: And, OC, thermography agrees with traditional go evaluation, but adds potentially important information.

Gérard TAILLE · Post by **Gérard TAILLE** » Tue Dec 01, 2020 12:11 pm

Bill Spight wrote: And as practical matters, chilled go infinitesimals, difference games, and ko evaluation are significant advances in go theory.

Yes Bill. In addition I like very much thermography and its mean value calculation.

Gérard TAILLE · Post by **Gérard TAILLE** » Fri Dec 11, 2020 8:16 am

When I studied infinitesimals I never encounterd ko situation. Are they excluded from the definition?

Here is an example:

Click Here To Show Diagram Code: [go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]

Each move to win the ko is worth 1 point as any infinitesimal.

My question is the following : is this position an infinitesimal?
In any case does the play of infinitesimals take into account such position?

Bill Spight · Post by **Bill Spight** » Fri Dec 11, 2020 9:47 am

Gérard TAILLE wrote:When I studied infinitesimals I never encounterd ko situation. Are they excluded from the definition?

Here is an example:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Each move to win the ko is worth 1 point as any infinitesimal.

Very nice.

My question is the following : is this position an infinitesimal?

There are kos that gain 1 point. David Wolfe told me that there are no ko infinitesimals. I never checked with Berlekamp.

In any case does the play of infinitesimals take into account such position?

Except for the possibility that such a ko might not be settled at temperature 1 in territory scoring, but later in the game, you take it in to account in Chilled Go.

Gérard TAILLE · Post by **Gérard TAILLE** » Fri Dec 11, 2020 10:37 am

Bill Spight wrote:
Gérard TAILLE wrote:When I studied infinitesimals I never encounterd ko situation. Are they excluded from the definition?

Here is an example:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Each move to win the ko is worth 1 point as any infinitesimal.
Very nice.

My question is the following : is this position an infinitesimal?
There are kos that gain 1 point. David Wolfe told me that there are no ko infinitesimals. I never checked with Berlekamp.

OK the game above G is not an infinitesimal.
More generally due to the ko you cannot compare, add or substract (=> no difference game) the game G to any other game.
Because of that, in one way we can say the game G is completly outside the CGT scope but in another way we are able to give a score and a miai value to G and we can even draw a thermograph and that means that such position is taken into account by CGT.
Yes you already told me that ko may be a mess for CGT. The apparent condradiction above proves this fact

Bill Spight · Post by **Bill Spight** » Fri Dec 11, 2020 10:50 am

Gérard TAILLE wrote:
Bill Spight wrote:
Gérard TAILLE wrote:When I studied infinitesimals I never encounterd ko situation. Are they excluded from the definition?

Here is an example:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Each move to win the ko is worth 1 point as any infinitesimal.
Very nice.

My question is the following : is this position an infinitesimal?
There are kos that gain 1 point. David Wolfe told me that there are no ko infinitesimals. I never checked with Berlekamp.
OK the game above G is not an infinitesimal.
More generally due to the ko you cannot compare, add or substract (=> no difference game) the game G to any other game.
Because of that, in one way we can say the game G is completly outside the CGT scope but in another way we are able to give a score and a miai value to G and we can even draw a thermograph and that means that such position is taken into account by CGT.
Yes you already told me that ko may be a mess for CGT. The apparent condradiction above proves this fact

I actually think that the main thing that shows that kos are not combinatorial games is the fact that a ko threat may be equal to 0 as a combinatorial game, and therefore we cannot say that K + 0 = K, where K is a ko.

Gérard TAILLE · Post by **Gérard TAILLE** » Fri Dec 11, 2020 2:40 pm

Click Here To Show Diagram Code: [go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]

Well, assume now we can tell black (or white) is komaster.
With this new information can the position be considered as a CGT position that can be compared, added or substracted to other CGT positions?
BTW in this hypothesis (I mean it exists a komaster) how do you calculate the score and the miai value of the position?

Bill Spight · Post by **Bill Spight** » Fri Dec 11, 2020 4:02 pm

Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Well, assume now we can tell black (or white) is komaster.
With this new information can the position be considered as a CGT position that can be compared, added or substracted to other CGT positions?

Komaster does not do it. There is still a ko fight. We just know who wins it.

Komonster probably does it. For instance, when Black is komonster this position is like 2*. When White is komonster it acts like {2|-1}.

Gérard TAILLE wrote:BTW in this hypothesis (I mean it exists a komaster) how do you calculate the score and the miai value of the position?

This is a placid ko, so the mast value and temperature are the same regardless of who is komaster. The difference lies in the thermograph below temperature 1. When White is komaster the right wall in that range is the line, v = -1 + 2t. When Black is komaster the right wall in that range is v = 1 - t, the same as the left wall. That's because when Black is komaster the local plays when White plays first in that range are White takes ko, Black takes back, Black wins ko.

Gérard TAILLE · Post by **Gérard TAILLE** » Sat Dec 12, 2020 6:29 am

Bill Spight wrote:
Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Well, assume now we can tell black (or white) is komaster.
With this new information can the position be considered as a CGT position that can be compared, added or substracted to other CGT positions?
Komaster does not do it. There is still a ko fight. We just know who wins it.

Komonster probably does it. For instance, when Black is komonster this position is like 2*. When White is komonster it acts like {2|-1}.

if Black is komonster I do not understand why this position is like 2*. Why not simply 2 ?

Bill Spight · Post by **Bill Spight** » Sat Dec 12, 2020 9:33 am

Gérard TAILLE wrote:
Bill Spight wrote:
Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
Well, assume now we can tell black (or white) is komaster.
With this new information can the position be considered as a CGT position that can be compared, added or substracted to other CGT positions?
Komaster does not do it. There is still a ko fight. We just know who wins it.

Komonster probably does it. For instance, when Black is komonster this position is like 2*. When White is komonster it acts like {2|-1}.
if Black is komonster I do not understand why this position is like 2*. Why not simply 2 ?

The * is for filling the ko at temperature 0. And I should have written {2|-1*} for the same reason.

Gérard TAILLE · Post by **Gérard TAILLE** » Sat Dec 12, 2020 10:28 am

Bill Spight wrote:
Gérard TAILLE wrote:
$$B
$$ -----------------
$$ | . X X a X O . . . |
$$ | O X . X O O . . . |
$$ | . O X . X . . . . |
$$ | . O O X X . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ -----------------
Click Here To Show Diagram Code
[go]$$B $$ ----------------- $$ | . X X a X O . . . | $$ | O X . X O O . . . | $$ | . O X . X . . . . | $$ | . O O X X . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ | . . . . . . . . . | $$ -----------------[/go]
if Black is komonster I do not understand why this position is like 2*. Why not simply 2 ?
The * is for filling the ko at temperature 0. And I should have written {2|-1*} for the same reason.

That was my interpretation but it is not clear : because * = {0|0} and because black is komonster the resulting position do not allow white to play towards 0. That is why I do not see a "*" here.

Let's take another example:

Click Here To Show Diagram Code: [go]$$B $$ ----------------- $$ | a O X . O . O | $$ | X X X X O O O | $$ | O O O O X X X | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ | . . . . . . . | $$ -----------------[/go]

Here again black will wait temperature drops to 0 before playing at "a" to gain a prisoner.
But it is not a {0|0}=* because white cannot play.

Life In 19x19

Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography

Re: Thermography