It is currently Thu Mar 28, 2024 8:13 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 10 posts ] 
Author Message
Offline
 Post subject: Not Counting Stones for Judgement / Evaluation
Post #1 Posted: Tue Jun 25, 2019 11:14 pm 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
There is a misconception that using area scoring rules would make it necessary to count all stones on the board when making a positional judgement during the early or middle game or global endgame evaluation. This is not so. In my books Positional Judgement 1 [PJ1] and Endgame 2 [E2], I give general principles (often truths), which rely on theorems proven in a later volume.


"Usually, when making a positional judgement during the opening, middle game or early endgame, regardless of the scoring rules, calculate the territory count." [PJ1, E2]

"Usually" because asymmetrical sekis provide an exception.


Advice for area scoring:


"If there have been no passes, the stone difference equals the prisoner difference or, if it is White's turn, the prisoner difference plus 1." [E2]

That is, you need not count all the stones on the board. Instead, it is sufficient to recall the difference of white and black prisoner stones.


"Modify an area count of the whole board by subtracting the number of black passes and adding the number of white passes." [E2]

That is, even if passes have already occurred, it is still sufficient to recall the prisoner difference instead of counting all stones on the board.


"Only count the stones in a specified locale". [E2]

This is used for studying local sequences. It is sufficient to count the local stones in the most suitable chosen locale instead of counting all local stones or all stones on the board. We would choose a locale of intersections on which changes occur due to the local sequences. We would ignore static local surrounding stones.


Then there are several principles determining the winner by parities instead of counting all stones on the board. [E2]


"Ignoring microendgame means ignoring two-sided dames." [E2]


There are further principles for endgame evaluation, so that evaluation under area scoring is like evaluation under territory scoring. [E2]


The "local stone difference" is the number of black stones minus the number of white stones in the locale. [E2]


"Usually, for a positional judgement of an initial position in a locale, the area count minus the local stone difference is its territory count." [E2]

That is, even for accurate local endgame evaluation during the late endgame, we need not count all stones on the board. It is sufficient to count the stones of the most suitably chosen locale.

"Usually" because asymmetrical sekis provide an exception.


Then, there are the principles of strategic differences: fill dames, sekis, dame ko fights. [E2]

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #2 Posted: Thu Jun 27, 2019 11:54 pm 
Judan

Posts: 6725
Location: Cambridge, UK
Liked others: 436
Was liked: 3719
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
RobertJasiek wrote:
There is a misconception that using area scoring rules would make it necessary to count all stones on the board when making a positional judgement during the early or middle game or global endgame evaluation.


Is there? All the Chinese players I've spoken to estimate the score by counting territory even if they do Chinese counting at the ends of the game. Or is the misconception by people who don't count at the end with area scoring thinking that those that do estimate the score area-style?

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #3 Posted: Fri Jun 28, 2019 1:32 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
Not everybody has the misconception and possibly Chinese have it less often. Those Chinese counting territory for judgement might or might not be aware of the correct relations to counting area. Most of the relations are new or new in their accurate description in my book. Parity-winner relations have been known since the 1990s but my proofs came later.

When I was first thrown into an area scoring tournament (EGC 1994), I had no idea of the relations. Even in 1997, when I counted all the stones on the board and in the Ing boxes to verify my upset 1 point victory during the late endgame, in which I lost my last 4 points due to a dame ko fight not knowing it at all. Heck, I did not even know the ordinary 2 points value of a basic endgame ko then.

Just counting territory for judgement is no evidence of not having misconceptions!

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #4 Posted: Fri Jun 28, 2019 3:51 am 
Lives in gote

Posts: 418
Liked others: 9
Was liked: 83
Rank: kgs 5 kyu
KGS: Pio2001
A loose way of counting during a game played with area scoring rules, I suppose, is to count the territory, and keep doing so as long as the difference between Black and White's score is more than 2 points.

Then, if we need to have an accurate estimation of the score, with an accuracy better than 2 points, we can convert the territory points into area points, knowing the relationship between the two.
Or, if in doubt, we can then count the stones of just one player. It is not necessary to count the stones of both players, as the absolute score of one player is enough to know the result of the game. But it is necessary to know exactly how to divide the komi and add it properly to the right player. We can learn by heart the threshold for an even game with 7.5 komi (used with area rules). Black 184 - Black looses. Black 185 - Black wins.
It is always possible to count all the stones of both players, especially early in the game, when it is difficult to estimate the points in empty zones. It is not very long. And, as Robert recalls above, it is not necessary to recount for every move. Once the score difference is known, we can keep it in mind and update it as the game progresses.

I've already done so in some of my games. It is very easy and if, like me, you have trouble remembering numbers, it is a safe method, because there is no addition with prisoners, nor empty intersections that are worth 1 point, and intersections with dead stones that are worth 2 points.
I know I am not good at these calculations during a game, and I have found that I make less mistakes with the numbers when I just count the intersections on my side and the intersections on my opponent's side (area).

It is also easier for me to evaluate the difference between two sequences by just counting the "number of intersections whose colour is changing" (area) between the two positions, instead of having to mentally count the territory and the number of dead stones at the same time. The dead stones are not even the same for the two sequences. That's too many things to keep in mind for me.

I'm not saying that this method is good. I'm just telling why I use it myself.
It's like the Rubik's Cube. The pro's method is to use the most efficient algorithms to solve it as fast as possible.
I'm not a pro of the Rubik's Cube, and all I use is the very slow method for beginners (LL edge orientation, LL edge swap, LL corner swap, and LL corner orientation). Like area counting, it is not efficient, but it is the easiest to learn.

Also, we have to keep in mind that all these calculations are thrown away as soon as a tesuji comes into play, and then you realize that what you thought was alive is in fact dead, that the group with 6 approach liberties has in fact 2 approach liberties because of the throw-in, that the endgame move that you thought was gote is in fact reverse sente because you misread the sequence etc.

Also, more often than not, once the endgame has begun, having an accurate count doesn't change anything : there is only one best move, and it is the same whether you are ahead or not in the game.

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #5 Posted: Mon Jul 01, 2019 12:08 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
Citation reference: https://lifein19x19.com/viewtopic.php?p=246380#p246380

Yakago wrote: "Chinese rules with 7.5 komi is equivalent to Japanese/Korean with either 6.5 or 7.5 komi"

You are wrong. Endgame 2 Values provides the theory on pp. 204ff:

"Standard area komi is ...5.5, 7.5, 9.5... Assume standard area komi."

"If the board parity (odd on 19x19) and seki parity (even in most positions with zero or two not scored intersections in sekis) are unequal, the possible area scores are ...-0.5 (smallest white win), 1.5 (smallest black win)... Increasing the komi by 1 does not change the winner."

Therefore, in most games, for the standard area komi 7.5 and the smallest black win 1.5, increasing the komi by 1 to 8.5 does not change the winner with the smallest black win 0.5. For the standard area komi 5.5 and the smallest black win 1.5, increasing the komi by 1 to 6.5 does not change the winner with the smallest black win 0.5. For the standard area komi 7.5 and the smallest white win -0.5, DEcreasing the komi by 1 to 6.5 CHANGES the winner with the smallest black win 0.5.

Hence, it is correct that, in most game scoring positions (with the mentioned parities), area scoring rules (such as Chinese) with 7.5 komi have the same winner as territory scoring rules (such as Japanese / Korean) with 7.5 or 8.5 komi. It is wrong that, in most game scoring positions (with the mentioned parities) and smallest scores, area scoring rules (such as Chinese) with 7.5 komi have the same winner as territory scoring rules (such as Japanese / Korean) with 6.5 komi.

See the book for further related truths of scoring and winner. I speak of truths and not just principles because I have proved the theory mathematically. The proofs fill several pages.

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #6 Posted: Mon Jul 01, 2019 2:23 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3374
While there are differences between rule sets that can produce large differences between net area and territory scores for even games before komi, typically the two are either the same (when the number of stones played by each player are equal) or the net area score is one point greater than the net territory score (when Black has played one more stone than White). Each result occurs roughly half the time. It is also the case that on boards with odd parity, such as the 19x19, 13x13, and 9x9, net area scores with even parity are rare.

On odd parity boards a territory komi of 7.5 will produce almost the same results as an area komi of 7.5. Net area board scores of 7 (net territory scores of 6 or 7) will be a loss for Black, while net area board scores of 9 (net territory scores of 8 or 9) will be wins for Black. The rare differences will occur with net area scores of 8 which are net territory scores of 7.

With a territory komi of 6.5 the differences will occur with net area board scores of 7 which are net territory scores of 7. These results are not rare on odd parity boards.

If we ask what territory komi is closest to an area komi of 7.5, it is 7.5,

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.


This post by Bill Spight was liked by: Waylon
Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #7 Posted: Tue Aug 27, 2019 12:42 am 
Dies in gote

Posts: 52
Liked others: 2
Was liked: 12
I recently came to think of this again.

What I meant when I said that 'chinese 7.5 is equivalent to japanese 6.5 or 7.5' Is that there is a difference based on who gets to play the last move.

Take for instance this SGF. In the final position, we have a territory difference of 6 points in blacks favor, while the area difference is 7.
The reason is that black has played one more stone.
In this game white wins with just 6.5 komi in japanese rules, while 7.5 is needed wrt. area.
Had black omitted the last move and let the game end there, the area and territory rules would agree.
This situation is of course quite normal.

So apart from possible edge cases with weird sekis, I'm not sure if it was something else you wanted to correct with respect to the statement?


Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #8 Posted: Tue Aug 27, 2019 3:09 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3374
Yakago wrote:
I recently came to think of this again.

What I meant when I said that 'chinese 7.5 is equivalent to japanese 6.5 or 7.5' Is that there is a difference based on who gets to play the last move.

Take for instance this SGF. In the final position, we have a territory difference of 6 points in blacks favor, while the area difference is 7.
The reason is that black has played one more stone.
In this game white wins with just 6.5 komi in japanese rules, while 7.5 is needed wrt. area.
Had black omitted the last move and let the game end there, the area and territory rules would agree.
This situation is of course quite normal.

So apart from possible edge cases with weird sekis, I'm not sure if it was something else you wanted to correct with respect to the statement?


If you are using area scoring with 7.5 komi, which is how the bots play, then territory scoring with 6.5 komi can give you a different winner. That happens when the area score is 7 for Black and White gets the last dame (and there are no other differences because of ko or seki). In that case the territory score is also 7. To get the same winner you have to use the same komi, i.e., 7.5. So an area komi of 7.5 is not equivalent to a territory komi of 6.5 or 7.5, as a rule. A territory komi of 6.5 will relatively often (but only a few percent of the time) give a different winner than an area komi of 7.5.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #9 Posted: Tue Aug 27, 2019 4:27 am 
Gosei
User avatar

Posts: 1753
Liked others: 177
Was liked: 491
I think there is a misunderstanding on Yakago's statement. Consider:

A: area scoring with komi 7.5
B: territory scoring with komi 7.5
C: territory scoring with komi 6.5

Yakago said that, for a given game without weird sekis,

(A and B determine the same winner) or (A and C determine the same winner).

Yakago didn't say that (A and B determine the same winner for all games), nor that (A and C determine the same winner for all games).

Top
 Profile  
 
Offline
 Post subject: Re: Not Counting Stones for Judgement / Evaluation
Post #10 Posted: Tue Aug 27, 2019 7:57 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3374
Disregarding the edge cases, as Yakago puts it, there is an ambiguity in his statement that area 7.5 komi is equivalent to territory 6.5 komi or 7.5 komi. It could mean that area 7.5 komi is equivalent to territory 6.5 komi or area 7.5 komi is equivalent to territory 7.5 komi. That statement is correct, in the sense that area 7.5 komi and territory 7.5 komi produce the same winner (disregarding edge cases). Or it could mean that area 7.5 komi is equivalent to (territory 6.5 komi or 7.5 komi). That stament is incorrect in the same way as the statement that area 7.5 komi is equivalent to (territory 4.5 komi or 7.5 komi).

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 10 posts ] 

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group