Go chinese rules should pass count


Back to forum

Thibault de Vassal    (2010-01-26)
Go chinese rules : should pass count?

A player just told me that in chinese rules for the game of Go, consecutive pass change the score of the game (thus in some cases the winner as well), I just visited a few websites that do not even mention this rule that is taken in account by some Go programs.

In my opinion it does not have any sense to link this to the score of the game, but I may be wrong... Does anyone have an opinion on this and clear examples that show it should be taken (or not) here at FICGS ?

I'll add a word on this in the rules after it is clearer to me.

Thanks in advance!

Don Groves    (2010-01-27 06:01:17)
Go chinese rules : should pass count?

Since Chinese rules count occupied points as well as surrounded points, passing is a serious mistake unless that player is certain of victory (at least 184-1/2 confirmed points on FICGS).

Consecutive passes end the game, so how could the score change?

Svante Carl von Erichsen    (2010-01-27 21:06:17)
Go chinese rules: should pass count?

The question is not posed very clearly, and it seems that more explanation is required. Passes do not count in "chinese rules" per se. I think that maybe some confusion arose because there are some rule sets that put some value on all or specific passes.

Anyway, you should read e.g. this very nice write-up of the matters to consider, "On the rules of Go" by Ikeda Toshio, which you can find at gobase:


Svante Carl von Erichsen    (2010-01-27 21:09:41)
Go chinese rules : should pass count?

I'd like to add that, if a rule set puts some value on a pass, there is a good reason for it, not just "hey, let's make passes count."

Thibault de Vassal    (2010-01-27 23:23:10)
Go chinese rules : should pass count?

I must say that the case encounted was not clear at all to me as I had very few informations on how this program scores Go games, anyway I should have written "passing MAY change the score"... The document is very interesting, thanks !

I think I'll specify in the rules that "The score is not influenced by which player passes first or last", which looks like more conventional & logical, unless anyone can explain why this decision should be avoided.

Thanks again.

Don Groves    (2010-01-28 00:33:32)
Go chinese rules : should pass count?

That rule would only apply when both players pass consecutively, right?

If one player passes and the other places a stone, obviously the score changes.

Ulrich Imbeck    (2010-01-28 23:21:53)
Go chinese rules : should pass count?

If one player passes and the other places a stone, the game goes on.

Svante Carl von Erichsen    (2010-01-29 00:48:50)
Go chinese rules : should pass count?

As far as I see, you do not have your rules written up anywhere yet, just a reference to "chinese rules".

I think that referring to rules that are not in place has high potential of confusion.

If you want to explicitly put down the rules that shall be used on this server (game rules, not tournament or server rules), I can offer you some assistance.

Lazaro Munoz    (2010-01-29 06:21:52)
Go chinese rules : should pass count?

AGA rules use passing stones, that is when you pass your opponent gets a stone. It does not matter too much at the end of the game when you get two passes in a row by alternate players.

Actually in AGA, The white player must always end the game so he may have to due a third pass (costing a real stone).

It sort of penalizes the player that chooses to try to end the game before all points have been taken.

Svante Carl von Erichsen    (2010-01-29 23:03:44)
Go chinese rules: should pass count?

Lazaro Munoz, the pass stones do not "penalize" anyone, they just make the score of AGA territory scoring always equal to the score of AGA area scoring. Under area scoring, failing to occupy even a "neutral" point instead of passing costs a point (unless no neutrals or an even count of neutrals is left), and this has to be reflected in the territory scoring. AGA rules thus have two exactly equivalent scoring methods.

A better explanation might be this: Under area scoring, each move is worth exactly one point more than under territory scoring, viz. the point that it occupies. Under the assumption that both players make the same number of moves (that is the reason for the rule that White always has to move last), this precisely cancels out. Since a pass is worth 0 points under area scoring, it has to be -1 points under territory scoring, which is represented by the pass stone.

Thibault de Vassal    (2010-01-30 01:58:20)
Go chinese rules : should pass count?

Extract from the tournament pages : "Scoring method is area scoring with chinese counting. Positional superko rule applies, it's impossible to repeat a previous board position."

I must admit that I have some difficulties to perfectly understand all this in english... so does it mean that passing never changes the score in FICGS games, or the contrary ? (anyway thanks again for helping)

Don Groves    (2010-01-30 05:04:58)
Go chinese rules : should pass count?

Thanks, Svante Carl, for the excellent interpretation. I now understand it for the first time ;-)

Svante Carl von Erichsen    (2010-01-31 14:38:58)
Go chinese rules: should pass count?

There is an official french ruleset that is, as far as I recall, almost exactly the same as the AGA rules. Look here, and with regard to the pass stones, also note the section "Méthode de décompte rapide":


This explanation uses area scoring throughout, except for the faster scoring method explained in the last section mentioned above. Pass stones are only needed for this faster scoring method. It is mostly irrelevant for scoring games on a computer.

The french rules are almost exactly what you seem to intend with your reference to chinese rules, except that they use situational instead of positional superko (see section "répétition").

Thibault de Vassal    (2010-02-03 16:59:57)
Go chinese rules : should pass count?

Thanks again Svante Carl, these pages are really interesting to read. There are so many rulesets for Go... I'm not sure to understand everything though.

I can't understand why, in a deterministic game such as Go where Black always plays the first move, there should be an extra half-point or full-point for White if he plays last or not, the same for any pass stone, so in my opinion the way games are scored here shouldn't change, but I'm not sure if something should be added in the rules. Do you think that something like "pass stones are not counted" would be useful & clear enough ?

Any opinion ?

Svante Carl von Erichsen    (2010-02-03 18:25:37)
Go chinese rules: should pass count?

Pass stones are irrelevant for area counting, so I think that you should not even mention them. It would only cause confusion.

Pass stones and last move compensation are methods to reconcile area and territory scoring. Last move compensation has another merit: in area scoring, the usually possible results always differ by two points, because when a point changes ownership, it is a loss of one point for one and a gain of one point for the other player. Last move compensation "sharpens" the possible results, and makes scoring very similar to territory scoring. However, this also is not necessary, so, at least as long as you don't fully understand this yourself, I would advise to keep simple area scoring and not mention anything else.

Thibault de Vassal    (2010-02-03 19:56:46)
Go chinese rules : should pass count?

I think I see how this last move compensation "sharpens" the results, but if to make the game sharper is the main interest in it, I don't think that Go needs that :)

Anyway, obviously there's no need to change the rules. Thanks again for your lights (sorry for my ignorance).