In this techno age it's always nice when human feelings and common sense take over. I was reminded of this today by the news that a triple ko took place in the China Weiqi League jamboree in Pingxinag, between Gu Li and Li Zhe. Even though, as I understand it, superko is still in the Chinese rule book, they ignored that and called the game void, then had a replay (Gu Li won).

So far as I know, every case of unresolved triple ko in China has been dealt with as void, either counted as a half point each or replayed. In quite a few cases, the games have been fast games, where arguably it would be hard to apply superko even if you wanted to. But the China Weiqi League is relatively slow (2.45 hours) and still they replayed.

Warms the cockles of your heart, don't it? Emotion aside, this decision is of course based on impeccable logic: they don't like superko so they don't use it.

This makes now I think 16 void triples kos in the GoGoD database, out of 68,800 games. That's 1 in 4300 games. Alternatively we have 5 in 15.000 games since 2006, or 1 in 3000. But since triple kos are eminently reportable even if they are by nonentity players, I'd guess - using an estimate of the ordinary prelim games that don't get published against the higher-round games that do - that the true figure is more like 1 in 10~15000 games.

I believe the game was between Gu Li and Li Jie, not Li Zhe.

What does the rule book say about triple ko? The 2002 version of Chinese Rules, translated into English by Guo Juan, is a mixed bag of game rules, competition rules, referee rules, and rules for tournament format and organization. In Chapter 1 (game rules, or "main rules" as translated by Guo), item 6, it says "It is forbidden to create a whole board shape which has appeared before." So, superko is in use.

But in Chapter 3 ("rules for referee" by Guo), item 20, it says "Triple Ko, quadruple Ko, eternal life Ko, two-stone Ko, etc., are rare repetitions of the same shape on the whole board which are in principle forbidden". Note the words "in principle" are added.

Then there is this entry at the end of item 20: "Depending on the type of tournament, it is possible to make alternative rules, for example: no result, tie or additional competition". So, superko depends on what the refs agree to do in any given tournament.

In all major tournaments in China so far, as John said, they nullified the game when triple ko occurred.

As was said by the Chinese during the International Go Rules Forum, the superko rule in the Chinese rules was meant only for sending-2-returning-1. So the referee exception rules with all their variety and ambiguity apply to cycles with 4+ plays.

There are different factions in China; otherwise the WMSG2008 positional superko could not be explained. Therefore saying "they don't like superko" is too simplistic.

Where is Guo's translation of the 2002 version available, please?

Fighting superko like a basic ko is easier than voiding the game because the former requires only 1 play in the triple ko while the latter requires at least 6.

Well, the tom.com site gives the name as 李喆 i.e. Li Zhe (it sometimes appears also as 李哲). Lie Jie is 李劼.

Maybe my Chinese is bad. ^_^ I of course meant 李喆, but I thought it is pronounced as Li Jie... He is a rising star and has great potential in winning a few international titles soon. So we need to get his name pronounced right. Can some native speaker help us here?

About the availability of the Chinese 2002 rules, you can find the Chinese version here: <http://games.sports.cn/datebase/encyclopaedia/wq/2010-07-19/2046610.html>. I don't know where to find the English version online. Ask Guo Juan. Hard copies were distributed at the 2010 Asian Games and at the 2011 WAGC.

Who said that? "The Chinese" is pretty vague. What you said above about "exception rules" (kind of an oxymoron ^^) are not true. Also, how do you know what the rules apply to when you have not even read them?

I find the arbitrary association of "does not use superko" with "warm, fuzzy, human feeling" by John quite irritating. I also do not agree with the idea that following your bias is "rational".

_________________
A good system naturally covers all corner cases without further effort.

Ko rules were discussed during the 6th meeting of the International Go Rules Forum, so it must have been during it. Present Chinese delegates were Jin Tong Shi and Wu 9p. I do not recall which of the two said it. Anyway, these two I mean when above I have said "the Chinese".

So one of the two Chinese explained "the superko rule in the Chinese rules was meant only for sending-2-returning-1" and that the referee ko rules would apply and would have been applied for all other long cycles. He read in Chinese the referee ko rules from the 2002 Rules. Yang translated and the translation was good enough to understand that there is little difference between the 1988 and 2002 versions WRT to the referee ko rules.

I have no reason to disbelieve what the two Chinese said and read. So, IMO, my earlier statements must be true, unless you can prove something else.

I do know how the rules are meant to be interpreted in particular because those two Chinese said and read the mentioned facts and also because earlier rough translations of the referee ko rules on rec.games.go, GoDiscussions and here support the contents of the above mentioned reading and translation during the meeting. Besides a purely legalistic interpretation of 1988/2002 Rules suggests the same interpretation because in Asian professional rulesets precedental rules have had a higher priority than basic rules in every such ruleset. John's report on how in practice long cycles have been treated is another confirmation that superko is always overridden when applying the (post-WW II) Chinese pro rules.

I will assume this is because of your faulty memory. I am sure you cannot back it up with the minutes of the meetings. I know Jin Tongshi and Wu Songsheng both very well (Mr. Jin was a chief referee in the 2010 Asian Games, where I was an arbiter); there is no way they, as very competent referees and rigorous rules experts, would make such an erroneous statement.

You want direct proof? Just read both rules, 1998 and 2002; I challenge you to find what you cited in there. Do not hide behind quotations by memory that are nearly impossible to verify. Another proof: the term "exception rule" was made up by you, just as you made up the term "referee ko rule". These terms are completely deviant from the context of the Chinese rules. Your citation of their saying "referee ko rules would apply.." therefore cannot be true.

I do not see what is so erroneous about the statement? The answer to a question in how the rules are practically applied and/or interpreted can only be erroneous if the answer is contradicted by the evidence. And since the evidence seems to show that superko, as we understand it, is not actually used in Chinese professional games, the explanation as given by mr. Jin and mr. Wu looks like an entirely reasonable explanation of how the rule is applied in practice.

Is there any example at all of a professional game where superko was used for a long cycle when Chinese rules were used?

Using the Chinese ko rule only to prevent s2r1 or similar situations is, IMO, a valid interpretation that solves a very immediate problem, since s2r1 is a problem for area scoring (but not for territory scoring). At the same time, they seem not to be bothering with extremely rare situations, and are happy to apply a simple solution of voiding the game.




I do not think reading of the rules can provide much proof at all about how they are interpreted in practice.

I have a copy of the translation, but I won't pass it on, since I don't know if it should be payed for, or if it should be distributed gratis, quote below are the relevant parts for 'superko'.

From past experience, I know Robert's notes are not always accepted as accurate.




6. Forbidden repetition of the same shape on the whole board

It is forbidden to create a whole board shape which has appeared before.

20. Forbidden repetitions of the same shape on the whole board

The only technical reason that obstructs the completion of a game is the repetition
of the same shape on the whole board. In principle it should be prohibited.

20.1 It is forbidden to immediately play a Ko capture without first playing a Ko
threat.

20.2 It is forbidden to play a false multiple rotational Ko (black move A in Diagram
3 is a false Ko).

20.3 Triple Ko, quadruple Ko, eternal life Ko, two-stone Ko, etc., are rare
repetitions of the same shape on the whole board which are in principle forbidden
(see Diagram 4 for examples). Depending on the type of tournament, it is possible
to make alternative rules, for example: no result, tie or additional competition.

_________________
North Lecale

s2r1 is discussed in both 1998 and 2002 rules along with superko, but never a hint that it is considered a special case. I have been to training sessions for the (Chinese) referees, again nary a word about a special case was said. You would think they would bring it up if it is indeed the only case that superko is applied.

Maybe someone can make a search and see if s2r1 has occurred in any Chinese pro games after the 1998 rules were in place. Then we will really see.

My memory concerning an S2R1 statement is correct. It is very likely that it must have been made during the 6th meeting because before ko was not a topic during the meetings while it was during the 6th. I do not recall whether Jin or Wu made the statement but I am absolutely sure that (if it was during the 6th meeting) either of them made it while the other was listening and not objecting. I looked in my report on the 6th meeting but could not find an S2R1 statement there, so I have to rely on my memory only.

You should not for no good reason just assume though that my memory would be faulty!

I agree, of course, that the rules (1988 according to English translation, 2002 according to what I have heard about them so far) do not have such an S2R1 statement in the text.



Actually I am not sure because it could be that Terry Benson or Chris Kirschner also have made some notes. (There were no official minutes. Whoever wanted to take notes did so and possibly made a report out of it. I did.)



Why erroneous?



I have not claimed the S2R1 statement to be explicitly in the rules. I have claimed that one of the Chinese made that statement VERBALLY. The statement is one of intention of what the purpose of the superko rule had been meant to be. One cannot find such intention directly by reading the rules text. Therefore yours is not anything like a direct proof.



Do not start unnecessary meta-discussion about my credibility! I also do not do that for yours.



It is not a matter of truth but of not knowing the Chinese 2002 headline for the related rules section, if the 2002 version has similar sections as the 1988 English version at all.

Why do you start another meta-discussion? (In which it would have to be discussed which kind of notes, where, when, who thought they were not accurate etc. A potentially endless meta-discussion.)

***

Concerning the 2002 citation: thank you!

I would guess that there are probably hundreds of games where s2r1 was on the board as a potential play. I would also guess that there are exactly zero games where any professional played it. To play s2r1, you have pretty much have to be a complete and utter idiot, and I do not think any complete and utter idiots have managed to become professional go players as of yet, so trying to find one in pro play is meaningless. Teaching it in training sessions for referees is roughly equally meaningless. It is a special case, you technically have to deal with it in area scoring rules, but beyond that is has no application at all in the real world.

S2R1 can be interesting when it starts from an instable state, when PSK and SSK are confused or when parts of an S2R1 cycle are played amidts parts of other cycles elsewhere:) All not particularly relevant in practice but fun for theorists.

In the absence of a native contribution, let me assure readers that Chinese-Chinese reference sources all give Li Zhe.

Not remotely relevant to this thread: Anecdotally, I can say Zh sounds a lot like a J. Go find the nearest Chinese person and ask them to say Chinese in Chinese. The first sound is a Zh.

NB: I do not endorse Wikipedia as a source for anything. With that said: A wiki link with someone saying China in Chinese.
http://en.wikipedia.org/wiki/File:Zh-zhongguo.ogg

_________________
Revisiting Go - Study Journal
My Programming Blog - About the evolution of my go bot.

To Chinese person it sounds different though. (Well, depending on where you're from in China, but...)

In the same way the first sound in "thin" and in "that" sound different to English speakers. Foreigners probably think we're batty.

Incidentally the "zh" sound, along with the "ch" and "sh" sounds, requires your tongue to be fairly far back on the roof of your mouth, so a word with an -i- vowel like -ie can't start with zh-, ch-, or sh- : try it and see! So the first sounds in Zhe and Jie can't even hypothetically be switched.