I've been thinking about rigorous computer-emulation of nearly-Japanese rules a bit more. Spight Japanese-style rules described here - https://senseis.xmp.net/?SpightJapaneseStyleRules - are very close to what I'd want.

However, I noticed an interesting detail about ko situations at the end of the game.

Click Here To Show Diagram Code

[go]$$c Does white need to protect at a or b?
$$ -----------------------------
$$ | . O . O . X a . X O . . . |
$$ | O O O X X X X X X O . . . |
$$ | X X X O O O O O O O . . . |
$$ | . . X X . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . X X X X |
$$ | . . . . . . . . . X O O O |
$$ | . . . . . . . . X X O b X |
$$ | . . . . . . . . X O O X O |
$$ | . . . . . . . . X O . O . |
$$ -----------------------------[/go]

(the two positions on this board are each intended to be considered separately).

I believe it's the case that in each of the situations shown the Japanese rules as applied in practice would require white to make an extra defensive play, thereby "costing" one point, even if white had sufficiently more ko threats at the end of the game than black such that black would not actually be able to force white to protect during normal play.

But my reading of Spight Japanese-style rules at suggests that in such situations, white should be able to defer the defensive move until the encore and thereby gain 1 point if white can get the first move in the encore. And if black gets the first move in the encore, then white will die. So that means that during the normal game, if black makes the first pass, white might engage in a pass fight (https://senseis.xmp.net/?PassFight) so as to be the first to play in the encore, and potentially gain 1 point more than in normal Japanese rules.

Is this right?

Position B (bottom right) is a two stage ko. During the first round of play (the play) White must normally win the ko if he can. In fact, each play in the ko gains 6 pts. on average, so it would be quite unusual for this position to remain on the board at the end of the play. That might happen, however, when White is komaster. In that case, if White plays first in the encore he can capture the ko and then Black must eventually hand over a pass stone, so that White gets one more point than he would if he had won the ko during the play.

It is up to Black, then, to force White to win the ko during the play. At the end of the play she may do so by passing first, since her pass allows her to take the ko on her next turn. Being komaster, White can play a ko threat and take the ko back. Black passes again. White has a finite number of ko threats, or else there is a superko. Once White has run out of ko threats, and Black passes, White must win the ko to prevent Black from taking and winning the ko, whether in the play or in the encore.

If White passes first in the play, then Black must take the ko to force White (as komaster) to win it before the encore.

If you wish to call this a pass fight, you may, OC, but it is simply using a pass like a dame, unlike other pass fights.

Position A (top left) is a position where Black to play can make an approach ko, while White to play can make a seki. If Black is komaster, White should make seki during play, if possible, to gain 4 pts. on average. If White makes seki during play, and Black takes the seki stone, Black gets one point.

During the encore, if Black plays first and takes the seki stone, White will eventually hand over a pass stone. But then Black will hand over a stone in compensation for the stone played in seki. So Black will get one point. If White plays first and hands over a pass stone, Black will take the seki stone, and eventually hand over a compensation stone. So the result will be the same regardless of who plays first in the encore.

But suppose that White is komaster during the play. Black becomes komaster in the encore. So White must either make seki during the play or as the first play in the encore. If White makes seki during the play, Black gets one point, as indicated above. But if White makes seki as the first play in the encore, and then Black captures the seki stone, followed by two passes or the equivalent, then Black will get nothing for the seki stone, as she will have to hand over a compensation stone.

Suppose that White is komaster and Black passes, there being no dame left. If White passes, Black gets to start the encore as komaster. Disaster! If White makes seki now, Black gets one point in the seki. So let White play a ko threat, and after Black answers it White passes. As long as White is still komaster, Black cannot afford to make the ko. But if Black by passing can force White to use up his last komaster ko threat, then Black should pass instead of filling a dame, and if White fills a dame, then pass again. This is a proper pass fight.

N. B. Compensation stones in seki are weird.

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

Visualize whirled peas.

Everything with love. Stay safe.

Thanks Bill!

I think you're right, I was wrong in my analysis of the lower-right, if I understand you correctly indeed only the top situation actually features a real pass-fight for an extra point.

So I have a followup question. It seems that the pass fight and possible extra point on the top arises from the fact that, conceptually, the encore does two things simultaneously:

1. Imposes a "zero-ko-threat-like" environment everywhere to try to satisfy the Japanese rules' concept of "wanting" to score every position locally and independently as if there were no ko threats anywhere else.
2. Changes the global temperature by 1 point, reducing the cost of moves enough to allow cleanup moves to resolve positions via play without loss.

It is A that destabilizes the position, because while white is komaster the position is stable, but as soon as we enter the ko-threat-environment of the encore, white needs an extra play. The "problem" comes from the fact that A and B happen at the same time, so that (if white can win the pass fight) black may never get an opportunity to take advantage of A occurring, before B also happens and white gets to make the defensive play when moves have dropped in cost by 1 point, securing white a 1-point improvement.

Would this pass fight go away if we imposed A and B separately? First, have an encore round where only A is imposed - enforce the new ko rules but leave the global temperature the same, with neither any handing of prisoners for passing nor any restriction on both players taking the same number of turns. Now black should have the chance to force white to defend at the original temperature no matter what, so white can never gain an extra point here and there is no incentive to delay the defense or have a pass fight. Then have the second (and all subsequent) encore rounds additionally impose B as well to allow cleanup of any further positions.

Are there any other unexpected pathologies that would be introduced by splitting A and B to take effect separately to remove this pass fight?

This sounds like it may lose points (compared to Japanese, if L&D doesn't depend on who plays first) when the opponent can force some cleanup using the extra ko rule, as now that still happens in an expensive phase.

Yeah, that's actually the intention, because the examples I currently know of are all ones where you *want* the opponent to be able to force such cleanup and have the other player lose points, because in those positions the regular Japanese rules as far as I know would normally require the defensive move and lose those points as well as opposed to allowing the player to potentially gain a point. The position I posted above (that one that I didn't mis-analyze) is precisely one such example which I think I originally found from the 1989 commentary - although I'm not entirely sure and would be open to hearing if this position would be handled differently in practice in a real game. Reproduced again here for convenience:

Click Here To Show Diagram Code

[go]$$c Should white need to protect at a at the end of the game?
$$ -----------------------------
$$ | . O . O . X a . X O . . . |
$$ | O O O X X X X X X O . . . |
$$ | X X X O O O O O O O . . . |
$$ | . . X X . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]

Is there a position where that's not true or where it leads to other bad behavior? I'm very curious if my idea for removing this pass fight works.

I didn't mean in this example (where it removes the difference compared to Japanese, not creates one). Here L&D DOES depend on who plays first with extra ko rules (unlike my requirement above).

I thought about cases where the additional moves that are only necessary under the extra ko rules would normally happen in hypothetical phase (with Japanese rules), thus intentionally free.

Actually this seems tricky, I'm not sure such positions are possible in practice, it may also depend on how the rules handle stops and resumptions. But something like this may be interesting:

Click Here To Show Diagram Code

[go]$$c Ready to score?
$$ -------------------
$$ | X X X . O . O . |
$$ | O O O O O O X O |
$$ | X X X X X O X X |
$$ | . . . . X X O X |
$$ | . . . . O O O X |
$$ | . . . O . O . X |
$$ | . . . . . O . X |
$$ -------------------[/go]

No dame, outer stones alive, W has lots of threats and says B edges dead as is.

That's an interesting position.