Hello there =)

As you can deduce by my 9k KGS rank, I'm pretty much a newbie, and I'm just starting the "studying" part of Go, having only played by instinct until now.

Since I'm starting to rationalize and understand the game from a logical perspective, I've ran into many dilemmas and questions. Most of those seem to get an answer in one book or another, but there are a couple that I haven't solved yet and keep bugging me. Both regard difficulties that may arise between the players during the counting phase. I hope you can help me!

First question: Regarding bent four in the corner.

I recently read a chapter about the bent four in the corner in a life and death problem. Specifically, my problem arised after the explanation of a position, in which something that looked like a Seki evolved into a bent four in the corner in case white pushed, putting itself into atari. Naturally, this made it possible for white to start a direct Ko.

In the book I read, it said that in the original, seki-like shape, black is dead and can be removed as such in the end. The explanation is what made me raise an eyebrow: apparently, since White could get rid of all possible ko threats at the end of the game and THEN assess the situation, it was fine to take out the black stones directly. But unless my math is wrong, that doesn't add up. Taking the stones right away because you can hypothetically get rid of all ko threats isn't the same as actually getting rid of them, as that'd mean playing stones in your own territory, in some extreme cases even outweighting the gain of the kill. So how is this managed? Am I in my right, if I was black in that example, to request white to fix his open ko threats before declaring my shape dead?

Second question: Regarding general life and death.

This comes from something that happened at a game I played at KGS when I was a DDK over there. I think I was blatantly cheated on, but back then I didn't have the knowledge to suspect it and I didn't know how to react.

We got to the counting stage and I declared one of his groups dead. He revived them, I declared them dead and he revived. I told him in the chat that I considered the stones dead and he disagreed, inmediatly asking me for proof.

I wasn't sure that the burden of proof was laying on me but I shrugged and resumed the game. I played towards the kill and he passed. I told him he shouldn't be passing and I got something like "I don't think I need to play to keep it alive" as a response. Obviously he was wrong and the shape ended up captured, but the whole sequence cost me many points. It struck me at the moment how useful it had been for him simply not to acknowledge the status of a group. I assume this is not an issue in a public game in which some sort of judging is involved, but what in a 1v1 situation? How do you handle it if someone doesn't agree about the status of a group and requests you to kill it, therefore wasting points?


Thank you in advance for your answers =)

First question: Regarding bent four in the corner.

Under Japanese rules the only valid ko threat in the after-game phase is a pass. So under Japanese rules it is simply dead.

Under Chinese style rules the situation is more complicated. You can remove ko threats at the end of the game by simply playing in your own territory. This doesn't cost you any points once all dame have been filled since area-scoring rules are used. If there are irremovable ko threats at the end of the game you might save your bent 4 group, usually sacrificing something else.

http://senseis.xmp.net/?BentFourInTheCornerIsDead

Second question: Regarding general life and death.

Under Japanese style rules/territory scoring you either talk your opponent into submission or call a KGS mod.

In theory the situation is solved using hypothetical play. So you execute virtual play from the end position to determine the status of the groups, but the scored position isn't changed by that. There are many subtleties involved in this and the details differ between different territory-scoring rules. But it rarely matters in practice.

Under area scoring you simply play it out. Once again it doesn't cost you points.

_________________
Sanity is for the weak.

As you may be aware, there are two scoring systems (area and territory) which always give nearly the same result (within a point or two). They give the same result because there are, in effect, three classes of move: (i) moves that change the balance of territory by X pts., (ii) moves that don't change the balance of territory, but that are in neither player's territory, (iii) moves that don't change the balance of territory, and are inside one player's territory.

These moves are worth the same amount in both systems: the only difference is that each of them is worth one extra point under area rules (so X+1, 1, 0) and one fewer under territory rules (X, 0, -1). It always adds up the same... until you get to the point in the game where there are no more moves of type -ii- left. Then if one player keeps playing, he gets -1, -1, -1 under territory rules, but 0, 0, 0 under area rules.

This is all a roundabout way of saying that rulesets which use territory scoring normally have specific rules about what shapes are alive and dead at the end of the game, to keep the scoring consistent with area rules. Bent four is a special case of that.




Yes, he was just being stupid. Don't worry too much about it. If someone resumes and you think it's dead, just pass until he starts playing stones to defend his "living" shape. If you think it's dead and he thinks its seki, just pass and get a mod.

On preview:

As others have mentioned, there are two main ways of scoring go today, which are almost equivalent. The questions you pose arise for Japanese/Korean scoring, aka territory scoring.

Click Here To Show Diagram Code

[go]$$ How many points for Black?
$$ --------------
$$ | . O X . X O .
$$ | X X X X X O .
$$ | O O O O O O .
$$ | . . . , . . .[/go]

At the end of play, assuming that the outside White stones are alive, how many points does Black have in the corner under territory scoring?

Black has 4 points, 3 points of territory and 1 point for the dead stone.

Now, some beginners quite reasonably assume that you do not get points for dead stones, only for captured stones, and so there are only 3 points for Black in the corner, since Black has to fill 1 point of territory to capture the White stone. That would be a variant of go, but one that has never become popular.

Once you accept the principle that each dead stone counts 1 point, then there is no problem with counting the bent four stones without having to fill the opponent's ko threats first, just as it is not necessary to capture ordinary dead stones.

There is a problem if the opponent has a ko threat that cannot be filled first. Such unremovable ko threats exist, but are rare. Under Japanese rules, bent four is dead, anyway. The current Japanese rules have special rules of hypothetical play to determine life or death after play has ended.

Basically, a stone on the board at the end of play is by default assumed to be alive. The burden of proof that it is dead is on the player who would count it as a point. But proving death by capture does not mean that the player actually has to capture the dead stone at a cost of one or more points.

These potential problems with scoring have only rarely caused disputes between experienced players, which is why go went for centuries, if not millennia, without codified rules.

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

Visualize whirled peas.

Everything with love. Stay safe.

Your opponent sounds like one of those ignorant beginners who thinks that dead stones must be captured.

Here is one approach, if you do not want to wait to get a ruling. If he claims that dead stones are alive, ask him to fill the dame between his live group and yours, while you pass (unless a move is necessary). If he refuses, claiming seki, that claim may be easy to refute.

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

Visualize whirled peas.

Everything with love. Stay safe.

Both of your questions basically boil down to the fact that Japanese rules are not terribly logical. Chinese and AGA rules don't have either issue.

Question #1 is a consequence of the way Japanese rules avoid question #2, which is:

a) Set up a second board identically to the first board.
b) Play it out on that board (only valid ko threat is pass).
c) Apply the outcome of that play to the first board.

If this sounds like a hack--well, in my opinion, it is. The Japanese wanted to count territory + prisoners, not stones, and that leaves your problem #2 to deal with. In my opinion, AGA rules solve this much more elegantly by adding pass stones and requiring white to pass last.

_________________
That which can be destroyed by the truth should be.
--
My (sadly neglected, but not forgotten) project: http://dailyjoseki.com

Pass stones are probably better, but so far my experience is that most tournament goers don't know why they're doing it.

_________________
Occupy Babel!

@jts: elegance is a matter of taste, I guess.

AGA rules pro/cons:
+ count area or territory, result will be the same.
+ no special rulings necessary (bent four, etc).
+ unambiguous pass-- no shared language required...
- ...except to explain the pass rule.
- no one understands why they're doing the passes that way.

@hyperpape: In defense of the average tournament-goer, typically it is not explained well or at all. The focus is on what you should do at the end of the game, not why you should do it, and I've never heard it explained that the procedure makes the scoring equivalent to area (Chinese) counting, even if you count Japanese-style. So, I totally agree with you.

_________________
That which can be destroyed by the truth should be.
--
My (sadly neglected, but not forgotten) project: http://dailyjoseki.com

See, there you go. I thought the point of the pass stones was to make the score come out the same as territory scoring, even if you use area scoring. (Isn't the point of the pass stones to make sure that B doesn't get an extra dame for moving first?)

Either way, in my book they're a kludge.

An example.


komi: 0

Click Here To Show Diagram Code

[go]$$B
$$ -----------
$$ | . . . . . |
$$ | . O O O O |
$$ | O O X X X |
$$ | X X X . . |
$$ | . . . . . |
$$ -----------[/go]

Territory: W: 6, B: 7
Prisoners: W: 0, B: 0
Stones: W: 6, B: 6
Pass stones: B: 1, W: 1 (black passed first)

Japanese scoring: B+1 (territory + prisoners)
Chinese scoring: B+1 (territory + stones)
AGA scoring: B+1 (territory + prisoners + pass stones OR territory + stones)

But:

Click Here To Show Diagram Code

[go]$$B
$$ ------------
$$ | . . . . . |
$$ | . O O O O |
$$ | O O X X X |
$$ | X X X X . |
$$ | . . . . . |
$$ ------------[/go]

Territory: W: 6, B: 6
Prisoners: W: 0, B: 0
Stones: W: 6, B: 7
Pass stones: B: 2, W: 1 (white passed first and last, so black got two stones instead of one)

Japanese scoring: tie (territory + prisoners)
Chinese scoring: B+1 (territory + stones)
AGA scoring: B+1 (territory + prisoners + pass stones OR territory + stones)

Note that black "getting an extra dame" is equivalent to black having played inside his own territory.


If Mr. Jasiek sees this I'm sure he could write several pages of explanation/clarification, but I think an example might be the best way to explain.

_________________
That which can be destroyed by the truth should be.
--
My (sadly neglected, but not forgotten) project: http://dailyjoseki.com

Its just the opposite -- to make the score come out the same as area scoring, even if you use territory scoring.

(because area scoring doesn't change when adding to your prisoner count, but territory scoring does)

As The Middleman's numerous villans would say, its sheer elegance in its simplicity

I expect that you are right. But it is easy to explain them. With pass stones (White passing last) when the board is arranged for counting each player has the same number of stones on the board. That means that the net territory score equals the net area score, so you can simply count territory. That is more familiar to most US players.

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

Visualize whirled peas.

Everything with love. Stay safe.

Yes. The better way of think about it is that the AGA rules are an area scoring ruleset, with all the strategic implications that are implied from that (in terms of dames, endgame kos, etc.) If you have agreed with your opponent ahead of time to use area scoring, there is of course no need for pass stones. AGA TDs of course mention the pass stones because something like 99% of all players in AGA tournaments will try to score by territory and if you do that, you need to keep track of both captures and pass stones.

I think it's more elegant than the alternatives I've seen, like the World Mind Sports Games rules, which require you to keep track of who made the first pass and make a 0 or 1 komi adjustment based on who passes first.

With pass stones, you can say, "both players get an equal number of moves, and it is this fact that allows both area and territory counting and allows play out to resolve L&D disputes."

Here's another reason AGA rules are cool. If you do the math and adjust for the 7.5 komi, the score is always B+1.5, B+3.5, etc. or W+0.5, W+2.5, etc. Other parities are not possible unless there are an odd number of points in seki (rare) or maybe some other pathological weirdness. So if you're scoring an AGA game and come up with a result like B+0.5 or W+1.5, it might be a good idea to have a recount, because it's likely someone made a mistake. I actually see miscounts fairly often even with experienced players. People are tired at the end of a game, and doing arithmetic that involves both multiplication and addition is often unreliable. Also, prisoners fall of the board, etc. Stuff happens. I wish people would just accept area scoring, but so many have never been taught it even though it's simpler. People complain that area scoring is slower. If I were cynical, I'd think they'd rather get the wrong answer in a minute and a half rather than the right one in two minutes. More likely it's just that amateurs don't care about the result that much if it's close because they know if it's that close the result is pretty much random. Or maybe they just think they are better at scoring than they really are, the same way that 95% of drivers think they are above average.

I think it's just as easy to make mistakes in area scoring. Some of your groups of 10 stones could accidentally become groups of nine stones, stones can fall off the board unnoticed while trying to sort them into groups of ten, you can miscount the number of groups of 10 you have, etc.

_________________
We don't know who we are; we don't know where we are.
Each of us woke up one moment and here we were in the darkness.
We're nameless things with no memory; no knowledge of what went before,
No understanding of what is now, no knowledge of what will be.

Certainly these mistakes can happen, but the potential set of territory scoring mistakes is larger. And one problem with territory scoring (lost or extra prisoners) could occur at any time during the game, not just during scoring. In area scoring, there is also the factor that usually only one color is counted, which allows both players to pay more attention to what is going on.

Area scoring is not perfect, but I do think it's better.

Now you make me curious. How do you actually do area counting, so it's more accurate than territory counting?

In territory counting, you just rearrange the territory in 10s, 20s, if necessary 5s, and so on. Since you can fill in prisoners, the territory gets less, reducing the work you have to do when counting, and as additional bonus, increasing the leeway you have in rearranging. Rearranging in multiples of 10 makes counting mistakes improbable. And rearranging is easy, since you have lots of stones that are not counted, so you can move them without shifting any borders. The shape of the stones at the end is irrelevant, only the territory needs to be in convenient groups.

How would you do the equivalent in area counting? When you have to count both territory and stones, how can you rearrange anything without shifting the borders? And without rearranging to convenient groups, you're just asking for mistakes.

see sensei's

it looks somewhat tricky, but actually it is quite logical. you count empty intersections like during territory scoring (while taking advantage in possible 'rounding' the count by adding/removing stones), remember the number and rearrange stones into piles of ten to conveniently count them. then add the two numbers

(not that i would ever tried it)

_________________
Spilling gasoline feels good.

I might be wrong, but probably not.