What do J89 rules say about a direct ko ?

[Pio2001]

Hi,
I have studied the french and japanese rules and I have written an article about the differences between them in french.

But there is a case that I don't uderstand in the japanese rule. A direct ko remaining at the end of the game (figure attached below).

If fighting the ko was the last thing that both players did (Black played the last move in E3, recapturing one white stone), and White has no more ko threat, White passes.
What happens if Black passes too instead of connecting the ko ?

My two references for the japanese rules are Robert Jasiek's page about the World Amateur Championship rules, that are the closest I could find for the japanese 1949 professionnal rules, and James Davies' translation, on Wilfred Hansens' page, for the 1989 professional rules.

As far as I understand, under the 1949 rules, the relevant part is "A defensive move is required for a direct ko shape, when immediate means are available." Which means that black should add a stone in D3 before the score is counted.

But if I carefully follow the 1989 rules, I find the opposite. White can't immediately recapture the ko, and if White plays elsewhere, Black can connect. Thus, the E3 stone can't be captured. Therefore it is alive.
D3 is surrounded only by black stones that are alive, thus it is an eyespace.
The black strings has no dame, thus they are not in seki.
D3 is an eyespace surrounded by no strings that are in seki, therefore it is Black's territory.

This is in contradiction with the 1949 rules, but normally, both rulesets should agree, shouldn't they ?

[Mike Novack]

First you have to understand what the term "direct ko" refers to.


If it were a direct ko (and the player whose stones were in atari passed) I don't see why the player who made the last move would need to fill the ko. Simply capture the group (would always be a different move than filling the ko)

[Pio2001]

First you have to understand what the term "direct ko" refers to.

In fact, I was speaking about the case that I illustrated in the picture. I don't know if it is called a direct ko. It looked similar to the second picture in the 1980 rule under the "direct ko shape" paragraph.

If it were a direct ko (and the player whose stones were in atari passed) I don't see why the player who made the last move would need to fill the ko. Simply capture the group (would always be a different move than filling the ko)

It is the same : under 1989 rules, black would be alive and not in seki, therefore he wouldn't need to capture, and would get one more point.

[Bill Spight]

Hi,
I have studied the french and japanese rules and I have written an article about the differences between them in french.

But there is a case that I don't uderstand in the japanese rule. A direct ko remaining at the end of the game (figure attached below).

May I recommend that you consult Denis Feldmann? ( http://denisfeldmann.fr/ )

If fighting the ko was the last thing that both players did (Black played the last move in E3, recapturing one white stone), and White has no more ko threat, White passes.
What happens if Black passes too instead of connecting the ko ?

First, play stops. Play does not end until the players agree on the life and death of stones and territory.

Second, the players resolve the question of the Black stone in the unfilled ko. They do this through hypothetical play, which has its own rules. White to play can capture the stone, then Black must pass, which is her only ko threat in hypothetical play, and then White can fill the ko. Since the Black stone can be irrevocably captured in hypothetical play without giving rise to a new living Black stone, it is dead. The fact that Black to play could fill the ko and save the stone does not matter.

At this point there are different possibilities. One is that the players simply end the game. In that case the dead Black stone remains on the board, because it is not inside White territory. At the same time, the empty point in the ko is not Black territory, because it is not surrounded by live stones; it is a dame. (This is the normal case in a double ko seki, BTW.) However, since the live stones adjacent to that point have a dame, they are in seki, and any points that they surround are not territory (No points in seki). Obviously, Black does not like that result.

Another possibility is that Black does not agree to end the game and requests resumption of play. White must grant the resumption, and plays first. White still has a ko ban and passes. Then Black can fill the ko.

If all that seems bizarre, I think so, too. The Japanese '89 rules makers, responding to charges that the '49 rules were ad hoc and illogical, with special rulings, devised rules that almost always gave the same results, assuming correct play. The rules are not terribly clear, and they potentially give rise to strange results.

As far as I understand, under the 1949 rules, the relevant part is "A defensive move is required for a direct ko shape, when immediate means are available." Which means that black should add a stone in D3 before the score is counted.

Right.

But if I carefully follow the 1989 rules, I find the opposite. White can't immediately recapture the ko, and if White plays elsewhere, Black can connect. Thus, the E3 stone can't be captured. Therefore it is alive.

That is not how the Japanese '89 rules determine life and death. See above.

[RobertJasiek]

Pio2001, if you ask about J1989, do not use J1949 or WAGC for clarification. Read

http://home.snafu.de/jasiek/j1989c.html
http://home.snafu.de/jasiek/j2003.html

For the practical application, see Bill's message.

Concerning your thread title, J1989 say nothing about direct ko, although they should. Direct ko was more popular a rules concept in J1949 and WAGC.

[tiger314]

Generally speaking, Japanese rules assume both players play perfectly. The authors probably saw the fact that this is not the case as irrelevant

[Pio2001]

Thanks to everyone who answered.

Second, the players resolve the question of the Black stone in the unfilled ko. They do this through hypothetical play, which has its own rules. White to play can capture the stone,

If White to play can capture the stone, then everything is clear to me. I didn't take this possibility into account because Black's last move was to capture the ko.
Do you mean that switching from actual game into hypothetical play lifts all ko bans ?

[Bill Spight]

Thanks to everyone who answered.

Second, the players resolve the question of the Black stone in the unfilled ko. They do this through hypothetical play, which has its own rules. White to play can capture the stone,

If White to play can capture the stone, then everything is clear to me. I didn't take this possibility into account because Black's last move was to capture the ko.
Do you mean that switching from actual game into hypothetical play lifts all ko bans ?

Hypothetical play is distinct from regular play. It starts with no ko bans, and, if necessary, considers starting with each player. That is theoretically consistent with the idea that each go position has a single value. There are other theories, but that is the one that won out for the 1949 rules.

[Pio2001]

Hypothetical play is distinct from regular play. It starts with no ko bans, and, if necessary, considers starting with each player.

Aaah, that was the info that I was missing.
Thank you.

[Mike Novack]

In fact, I was speaking about the case that I illustrated in the picture. I don't know if it is called a direct ko. It looked similar to the second picture in the 1980 rule under the "direct ko shape" paragraph.

Understood. But you were quoting a rule where the translation you gave was "direct ko". Now I don't know Japanese, so can't check myself. But if it was "direct ko" that the rule was talking about then you example diagram doesn't reflect the situation because your diagram isn't a case of "direct ko".

[John Fairbairn]

But you were quoting a rule where the translation you gave was "direct ko". Now I don't know Japanese, so can't check myself. But if it was "direct ko" that the rule was talking about then you example diagram doesn't reflect the situation because your diagram isn't a case of "direct ko".

AFAIK direct ko always refers to honko in Japanese, and the diagram appears to show a honko (i.e. a genuine ko, not an approach-move ko). So what is there in the position to make you see it differently?

[Cassandra]

As a matter of course, the Ko shown is a "direct" one.

However, as Bill pointed out, it does not make any sense for Black -- under J1989 rules -- to leave this Ko unconnected.

[xed_over]

Generally speaking, Japanese rules assume both players play perfectly. The authors probably saw the fact that this is not the case as irrelevant

hahahaha... that is seriously, probably the best explanation of Japanese rules that I've ever read!