I like these principles, and I agree that they seem like desirable properties for board games. But I also think they are not absolute, and that there may be other principles than these, and that some games might violate these but still be good games.


Certainly super-ko violates your "Principle 1". However, most (all?) of the alternatives to super-ko seem to violate your "Principle 3" (the super-ko rule is very simple, while the alternatives end up more complex).

I'd also point out that chess violates your "Principle 1" in several ways:

• Castling is only allowed if the king and rook involved have not previously moved
• The three-fold repetition and five-fold repetition rules allow for a draw based on the entire past history of board positions. Note that this is somewhat like super-ko in go.
• The seventy-five-move rule allows for a draw based on the characteristics of the past 75 moves

(I suppose you might also argue that chess violates "Principle 3" in the sense that the rules and game pieces are complicated.)

You simply do not understand that you must not pick out an example of bad rule design to support your discussion position on cycles.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Well, you should know that you are wrong about the Japanese rules. The Japanese rules definitely DO NOT allow a player to "attempt to gain points by passing rather than reinforcing territory." The Japanese rules only allow for (A) both players to pass, thereby stopping the game or (B) play stones one after the other. The Japanese rules do not allow a player to play a stone after the other player has passed. The only exception is for kos in confirming life and death. So there is no possibility of gaining a point (dead stone) by passing.

Passing a captured stone when "passing" is a "game mechanic" because a game-piece is being worked on. This is especially important as it relates to scoring. Simply agreeing to end the game and move on to scoring is not a "game mechanic." Nothing is happening within the game. Passing may be allowed by a "rule", but not a "mechanic."


The score value is part of the game and so changing the score value is a "mechanic" within the game (for whatever its worth to you). I do not know whether Area III or Button Go (people seem fond of this one) include any game-piece to track the score for these mechanics. Maybe the mental bookkeeping is not burdensome if black is expected to pass and not passing would be an exceptional case that the players could not be expected to forget.


Is my understanding of game design wrong? Did anyone present evidence otherwise? Did some people clearly never consider whether a rule was good game design or not?

And back up for a second. Where did I ever argue that these "principles" were mandates? I didn't. I called them "principles" not mandates. I have said that game design involves tradeoffs. This should be obvious as every rule adds some burden to the players. The principles I mentioned are fundamental principles of game design, not because I said so, but because this is how board games are designed. If you think that games are not designed with these goals then please try to find examples.

Of course there are many rules or game mechanics violate of these principles. I have already discussed the design tradeoffs for the basic Ko rule. Having a rule to identify a ko is a burden on the players, but it leads to such a depth of strategy and interesting gameplay that it is an amazing design decision.

I haven't seen a good game that violates the principles often. Even bad games do not require additional mental bookkeeping. But when a game does create a burden, it is done with the intention of improving gameplay (depth of strategy, complexity of tactics, or just plain fun of the game). These improves must at least offset the burden of the rule or else the rule/mechanic is a bad design decision.

Even putting aside the mental bookkeeping required by superko, preventing cycles actually limits game-play options, thereby reducing tactical considerations, thereby making the game worse. A player can no longer avoid a loss by creating a difficult to achieve perpetual cycle. And, aren't such cycles are rare and exciting. Don't people like bragging about their molasses? Infinite cycles are fun. It should be a wonder to achieve it, not something to be prevented. But some people have the goal to prevent a feature of the game from happening. Not only does superko take away exciting and complex gameplay (limiting gameplay possibilities), but all it does in return is provide the "gotcha!" of losing the game because you failed to remember that a play would recreate a board position. An artificial victory condition is not interesting by itself. It is already bad game design not even considering the cumbersome mental bookkeeping.

That's not what I was doing. I was just pointing out that you misunderstand J89. Very weird that you would call me out without checking first.

And for being such a complainer maybe you should try to explain why you think Article 2 is bad rule design.

Also, Article 2 has nothing to do with cycles. And I was not using it to support my position.

Right?! Every attempt to avoid perpetual cycles ends up being overly cumbersome. Now, I know this might sound crazy. But it's almost as if the designers of Go deliberately intended for there to be the possibility of perpetual cycles.
----------

Sure. These are just principles of design. Most games follow them. However, many games, especially modern games, will violate the principles in an attempt to provide new and exciting (hopefully improved) game-play. Every design decision is a trade off. The question is whether the improvement to the game (depth of strategy, complexity of tactics, and fun) are outweighs the complexity of the rule/difficulty in implementing the mechanic.

Certainly the depth of strategy, complexity of tactics, and fun of the various chess pieces and their mechanics vastly outweigh the complexity of the rules. Overall, great board game design.

As for castling, it adds depth of strategy and excitement, but with the burden of additional mental bookkeeping. It is worth it? I don't know. Maybe it's a wash. At least it's not rule that causes a player to accidentally lose the game.

As a side note, I have noticed that when games (intentionally) require additional mental bookkeeping, it usually happens with a situation that is so important/exceptional (eg moving a king) that the players are expected to not be able to forget. Though most of the time, additional mental bookkeeping seems to be an unintentional mistake that was overlooked in design and would have been changed if noticed.

Chess may have a higher likelihood for repetition to occur compared to Go. Given the likelihood for chess, maybe checking for repetition is the best version of the game.

At least for Go, the game is better with the possibility of cycles.

I did not claim Article 2 being bad rules design.

Article 9 "End of the Game", Clause 1 is what matters here in combination.
1. When a player passes his move and his opponent passes in succession, the game stops.

This clause is nothing more than an attempt to prevent CYCLES that are perceived as inappropriate.

ONLY TWO successive "pass" are an unsatisfactory means to declare the "stop" of the game, as the "pass" carry different non-verbal messages each.

I waiver my right to place a stone on the board. I think that I am ahead, so I do not want to humilate you by occupying a dame point or one point of my own territory. Probably my assessment was not correct, and you see some profitable place left for you. It's your turn again.
Thank you very much indeed, but I do not see any profitable place on the board left for me placing a stone. If that also was the reason for you waivering your right to place a stone on the board, I would like to suggest stopping the game.
Yes, indeed, it seems that we agree that there are only dame points left on the board, so let's stop the game and enter the next stage.

THREE successive "pass" would have been much better, if the aim of "passing" was ONLY inducing the stop of the game.
But the J89's authors used ONLY TWO "pass", so they had yet another, but hidden, aim.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Three passes simply wouldn't work for that with double ko seki.

What? No. Nothing in Article 9 attempts to prevent cycles. Article 9 is simply the rule to begin confirmation of life and death status. Section 1 defines how this confirmation process is initiated. "Cycles" are not a concept within the rules of Go. You can tell because a cycle is never defined in the rules. While cycles of a feature of Go game-play, they are not an element of the Go rules or Go game design.

The explanation to Article 9 explicitly states that a pass is a declaration to stop the game (ie proceed with Article 9). It is nothing more than that. During normal gameplay (not after resumption or during confirmation) the ONLY intent conveyed by a pass according to the Japanese rules is a declaration to proceed with Article 9.

The third pass is superfluous. When both players pass, they have already agreed to proceed with Article 9. Requiring a 3rd pass that has no meaning within the game would be bad game design.

Quite apparently you do not want to understand that each of the "pass" carries a different non-verbal message.

To compensate, however, you are phantastic in cherry-picking.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Why not? Because the third pass could be replaced by a capture into one of the double-ko shapes?

This is why I capitalised "ONLY".

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Yes, but three passes wouldn't be good for inducing stop simply because of this: With the basic rules the opponent can normally be forced to eventually pass, but cannot necessarily be forced to pass in response to a pass.

I figured it out! You just don't understand the Japanese rule. The normal "pass" in Go when speaking English is not a pass. You are not passing the turn to your opponent. Read the Japanese rules. The action is abandoning your play to instead declare that you are ready to stop the game and proceed with Article 9. There is no ability to "pass" your turn over to your opponent as the term would commonly mean in other games. The Japanese rules do state that this is often referred to as a pass but they specifically state that the abandonment of your play is a declaration to begin Article 9. There is no ability to pass your turn to your opponent.

That's wrong. The opponent cannot be forced to pass under the Japanese rules. If black plays and then white passes, but black chooses not to pass, then the only action allowed by the rules is a white play or a white pass. But black does not have to pass.

It's funny, I thought I was making an argument in favor of a super-ko rule. I think allowing perpetual cycles ends up requiring somewhat similar mental bookkeeping, because otherwise you don't even know if you have a cycle. At the same time, it complicates the decision about who wins the game, since it is no longer just "who has the largest score". Is a perpetual cycle automatically a draw? A "no-result"? A win for the player who is already ahead on the board? A win for the player who captures more stones during the cycle?


I think that's unlikely, but who knows?


I'd argue that for go, the depth of strategy, complexity of tactics, and rarity of needing to deal with super-ko vastly outweighs the drawback of needing super-ko on those rare occasions.


I disagree. But it's OK that we disagree, it doesn't matter in most go games anyway. In fact, it has never mattered in any go game that I've played in my entire life.

I can see where you are coming from, but this doesn't seem to square with logic or real life.

I haven't read much of the foregoing thread so may have missed a crucial point, but it seems to me that if you say Pass, that is shorthand for "I think the game is over", and that is what you also are saying. But that is not the same as saying "The game is over." The opponent has the right to ignore you and to make a move on the board. And this is precisely what has happened more than once in pro play (e.g. Game 5 of the 2008 Kisei, White passes on move 292, Black plays 293, play continues up to move 310). A 1998 game had five passes.

Black's right to make that move 293 in the Kisei game (i.e. the turn to play is being passed to him) seems so blindingly obvious, and is implicit in your own words) that I struggle to understand what you are saying.

However, possible enlightenment came to me from the murky past. When I used to play, the common scenario at the end of the game (using always what we called Japanese rules) was that one player (call him White) would say Pass or something like, "I think it's over, OK?" But very often, Black might say, "No, I don't think so - it looks like you have play something over there" or "Are you sure? - I can still play here." Whereupon White might say, "Oops, you're right" and play a move after all. In which case he has not passed on the turn to play. Or he might say, "No, I don't think I have to fill in - but go ahead and try something if you want." In which case he has specifically passed on the turn to play to Black.

Now I have no idea if this is still how western amateurs play (I suspect it is), but it is certainly not how the professionals play. If a pro says Pass, he is saying, "It's your move now."

And remember the Preamble to the rules (the one by rules committee chairman Yoshikuni Ichiro) stresses that they are designed to reflect the way Japanese go has traditionally been played, so it's the pro way that counts.

There are a few other observations I will make, not specific to you, but they may interest you and others.

1. Article 2's commentary 1 says "The making of moves in alternation is a right." That's it. It seems like a blunt, comes-out-of-nowhere statement. In fact, it reflects a huge debate in Japan going back to 1928 as to whether making a move was a right or an obligation. For a Japanese this comment 2.1 does not come out of nowhere - it has a huge context.

2. Article 2's other comment includes the statement "Relinquishing [the right - as per 2.1] to make a move (a pass) is a declaration by the person relinquishing his right that the game is over."

No problems there, but relinquishing = pass is not really transitive. Other texts contain, for example 休止（パス）, i.e. pass = pausing the game. A subtle difference, but a difference nonetheless.

3. What is now called the confirmation phase has long had other names. One is 仮終局　(provisional end of the game), another is 停止　(cessation; used in the J89 rules), yet another the above 休止 (pause), and the most outlandish is Yasunaga Hajime's 第二の世界 (second world), which came from the habit of some researchers calling the actual game the primary world. (Fashions in metaphors do change.) Other pertinent phrases that were once in fashion were 本来の終局 and 窮局による終局. I mention these merely to illustrate that the rules field had been well and truly ploughed by Japanese rules mavens. They even discussed ideas like pass stones, though they called them 棄権賃. Yet still, the committee decided essentially to stick mostly with tradition.

4. 停止 (cessation) is the term used in current texts, and everyone here seems to take it to mean cessation of play, the game is over etc. Which is correct up to a point, Lord Cropper, but I have seen 停止 being used in older texts by rules researchers in the context of cessation of the right to play. Again a subtle difference but a difference, and one that would make the out-of-nowhere phrase "the making of moves in alternation is a right" have some relevance to the rest of the text.

Click Here To Show Diagram Code

[go]$$W
$$ -----------------------
$$ | O a X . O . O X . . .
$$ | X X X X O O O X . . .
$$ | O O O O X X X X . . .
$$ | . . . . . . . . . . .
$$ | . . . . . . . . . . .[/go]

We are still in normal play, it is white to play and I assume the last interesting move on all board is a black move at "a" to get a captured stone.
White pass, if black is not allowed to play at "a" then : black pass, black resumes the game, white pass, black plays at "a", white pass, black pass
What is the difference with directly (without game stops and resumes):
White pass, black plays at" a", white pass, black pass.

I don't think that go needs additional game-pieces (although KGS uses a special mark to show ko recapture prohibition).

However, I can't imagine that territory scoring is a good game design, as it requires an expert analyse in order to define the score. All territory scoring rules so far are extremely complicated, and understood only by a few. Area scoring rules are way better designed.
However, tracking cycles is mandatory in area scoring, because of situations like this :

Click Here To Show Diagram Code

[go]$$W
$$ ---------------
$$ | a O . X O . .
$$ | X O X X O . .
$$ | . X X O O . .
$$ | X X O O . . .
$$ | O O O . . . .
$$ | . . . . . . .[/go]

This is seki. However, under area scoring, if Black looses the game as it stands, Black could play "a" forever, starting and endless cycle.
The game is better if this is forbidden.

Click Here To Show Diagram Code

[go]$$B
$$ ---------------
$$ | . X O . . |
$$ | . X O O . |
$$ | X . 1 O . |
$$ | . X O O . |
$$ | . X O . . |
$$ ---------------[/go]

You're trying to tell us that after Black 1, White pass, filling the ko is an illegal move...

You mentioned these 2 games in the other thread (The "no ban in double ko" rule to solve pass-for-ko issue). As I explained over there, these 2 games (GoGoD files 1999-06-30b and 2008-02-27B) seem to involve game resumption. The 1998 game actually shows the 2 passes to stop the game and begin Article 9 "confirmation". The 2008 Kisei game does not show the passes to stop the game, but it does show the pass after the game is resumed. And even if games didn't involve resumption, these games are not cases involving one player passing so that the other can play several moves in a row. This is the game-activity that I think is clearly not allowed because it is not playing one after the other.

The play after a pass in the 2008 Kisei match was allowed because the Japanese rules explicitly allow for a pass when the game is resumed. What is interesting is that resumption was not needed because the black stones were clearly dead 死に石 based on the definition of 活き石.

====================
I read that but not having not seen the previous arguments, my understanding is that 着手 is an "allowance" or "right," but not an "obligation" because the player may choose 着手の放棄 instead. I appreciate your comments because I think there is a difference between 着手の放棄 and "passing your turn to your opponent." 着手の放棄 is not パス, even if it might commonly be called that.

Black does not need to play at "a". The one white stone to the left of "a" is not a living stone 活き石? It does not meet the definition because it can be captured by a black play at "a", and white cannot then play a new stone that cannot be captured.

I sort of agree. There are tradeoffs. Territory scoring is less cumbersome in that the stones do not have to be placed, but it is more complex in that life and death are not as obvious. On the whole, I think Area Scoring is a better design. But Stone Scoring is better design than Territory or Area because the scoring is directly related to the play mechanic. Area scoring needs additional rules for scoring.

I disagree. White could decide to let black capture and give up the seki. If the game is so close that this would decide the result, then allowing for black to create this complex situation in a close game adds to the depth of strategy (gameplay is improved at the cost of no winner). I think that gameplay should have more weight than merely deciding the winner. Though I understand why this is not the case for tournament play.

Yes. You cannot fill it during the game. It may be filled after the game is stopped by both players passing. Dame are filled. If the player that lost the ko disagrees, they can request to resume the game and their opponent will then have the opportunity to fill. The ko rule prevents the loser of the ko from retaking the ko even if they were allowed to play first by the resumption rules. A ko can only be recaptured after a pass when confirming life and death, not when resuming a game.