Sub-page of Axd

This page is work in progress. It is my attempt to get to grips with rules, and to try to find a decent ruleset. See also Conclusion.


IMO, the driving force behind Go is the beauty of the game (eg the Japanese notion of emptiness[1], then its relation between wood, lines, black and white): it should be reflected in the rules.

I always try to convince novices around me by referring to the simplicity of the game (although even then women won't take the bait - maybe someday I will have to admit they don't have the mental capabilities... Still, even if that would be true, then there is still intrinsic limits. Or maybe it's just me. ). I usually introduce the game using stone counting. Sometimes I will switch to area scoring, to end with territory scoring. But the intermediate step of area scoring is often too subtle.

In reality, the surface below is [ext] darker. What happened? Pro players bicker about one point. Of course their tournament result depend on it; but why don't they fix such situations, rather than allow such situations to continue to exist? The weak points of Go rules are

  • the result of humans wanting to shortcut the mechanical part of the game (eg dame filling)
  • disputed status, game end
  • board repetition
  • and various attempts at crafting rules (especially by the maths-gifted, who fail to realise that they should serve the community rather than isolate themselves in their ivory tower) that introduce complicated words that scare away lot's of would-be players.

Table of contents

The transition stone->area->terr feels natural (note: Ing writes that JPN rules fork off an ancient form of CHN, so CHN might - from an evolutive standpoint - be more recent)

Sunjang Baduk (without the initial setup) looks like a logical next extension to Go due to this new phase added: removing stones. But the actual rules need a cleanup first.

Dark areas

Then appear complex rulesets (most of this is hidden from the casual player), as well as various questions such as

  • why is there a fixed handicap system?
  • why is there a need to proclaim specific shapes dead?
  • why are there so many discussions around rules?

As a result, I get the feeling that the game rules might not be as simple as I try to explain to beginners; I end up living with a lie (ok - a small one), yet I'm faced with this dilemma:

  • explain simple rules and hide the reality (of course this approach works!)
  • scare potential players away by complex rules

maybe we should go back to the roots.


The handicap system is one of the attractive features of Go. At the same time, it opens the way to questions.

Fixed setup

Fixed handicap placement could reduce the actual value of handicap stones because White knows her way around it. Free handicap placement gets rid of the rules of where to put the stones. OTOH, to Fan Hui the hoshi would represent optimal playing. The hoshi will be very difficult to beat: their attraction and call for attention will last for quite a while, I guess...

the classic argument against free handicap is precisely to forbid black to make a shimari with handicap stones, [...] this was traditionally seen as too great an advantage for black. ([ext] game)

Tournament value

Playing tournaments with handicaps is odd, as one no longer compares strength between players, but rank agility. IMO, the main reason for running tournaments is to determine who's the strongest player -> even-games tournaments.

Historical value

Another big drag factor is that historical games have been played with handicap.

Rule value

Handicap issues should not impact rule issues. Handicaps have their value in informal games, where formal rules might be less essential.

Status of groups - disputes

[thumbnail diagram]

One of the problems seems to come from the counting phase, where groups are to be accepted as dead (or alive) - but it might not always be clear. See also Alternating Play.

IMHO the confusion stems from wanting to define L&D based on the concept of eyes, rather than "liberties".

Rulesets that list a series of positions and their status/point value, are suspect: there is no guarantee that a future problematic position might pop up that should have been covered. But this happened several times in the past, of course during important games in tournaments...

See also


Allowing suicide opens the door to a few extra possibilities. What's wrong with that? The suicide rule (either allowing or forbidding it) is ballast: instead, the rules should state that at the end of the move, no dead group should remain on the board (which is the case). Period. And suicide is often a source of confusion for beginners.

IMO, suicide is to be left away altogether: it must be allowed to kill your own stone(s) (and only an idiot will kill a single own stone, but there is no reason to describe such situation in the rules).

Removing a suicide rule does not reduce the amount of rules: eg Tromp-Taylor rules:

7. A move consists of coloring an empty point one's own color; then clearing the opponent color, and then clearing one's own color.

Interestingly, suicide of a single stone could be forbidden by some ko rule (because the board prior the move would be reached after removal of the dead stone).

See also

Repetition - ko

A HUGE weak spot in Go, that has traditionally been solved by the "basic" ko rule, plus some odd rules (eg triple ko leading to annulation).

The only way forward is superko: one of ssk, psk, nssk. It is the widest expression of "no repeating position".

The fundamental problem that remains is that the human player might not accurately detect a superko situation - although in reality it is unlikely to having to remember a whole board. (And what happens if a player falsely claims a move may not be played because of an actually inexistent superko?) Software should have no issue: a board can be coded into approx 71 bytes per move.

Ing seems to have made the first (yet difficult) steps in a possible solution: reduce superko to a set of ko situations.

Associated with repetition is how to handle passes: when, how much? AGA seems to offer a good answer.

Repetition and signalling end of game might very well be covered by the same superko rule.

Interestingly, repetition is a fundamental issue due to the limited size of a goban.

Another obs: "If you believe they really are different, try retaking a ko 'illegally', then rotating the board 90 degrees and claiming that its a new board position. " (anonymous, Small board Go)


Playing in own territory

A penalty under territory scoring; dismissing this feature would limit the game (a little bit). It is related to "efficient play", so someone who plays efficiently should be rewarded. Also, dame filling is (often) a mechanical and boring process (but sometimes it can force to play in own territory), so rendering dame useless is an interesting idea. Top players however will insist that a single point can change the outcome of the game... Equivalence scoring seems to have this inconvenience.

Maybe (to confirm) this feature is just the dual of dame filling under Area scoring, in which case we are happy. (It is dual until an even number of dame remain).

See also

My personal taste tells me that this is a positive side of the Japanese rules.


For some reason the Japanese rules (?) exempted seki groups from scoring. Why? Maybe seki is the result of insisting that groups must have two eyes to be alive; therefore, seki is perceived as something odd. But Go is about territory, not eyes.

OTOH, if a player manages to reduce the value of a big group through "seki neutralisation", that player deserves a bonus.

Maybe seki made defining 'life' more complicated.


Tournament rules

Unfortunately, because it is tournaments that triggered ruleset research, tournament rules must also be part of the rules. Doh!

And rules of conduct are yet another chapter.

Seems to be part of Ing rules.


IMHO, what we need:


This is not final...

Notes on various rulesets

General notes

  • There is a systematic tendency to adopt "define-before-use" to define rules, the obvious approach... in maths. This approach makes rules so difficult to read. Humans can very well live with Prolog-like declarations (rather than C-like), where the definition of what constitutes a "grid" or a "stone" is mentioned somewhere near the bottom of the page, and the core rules are mentioned high up the list. This doesn't mean that rules should be written in any order.

One can wonder if it is necessary to define the rules as if they were to be telephoned over or sent to some alien civilisation. (I can imagine the objections of mathematicians here...) Another example from Ikeda:"A certain rules theorist has proposed a rule that would permit repeating cycles that do not become endless. This is exactly the sort of language that we must keep out of the rules. It is nonsense to talk about a repeating cycle that does not become endless."

area scoring

  • filling in dame is mechanical (even though it can get rid of other issues)

Ing rules

  • Ing is one way to worm into the ruleset world. aspects I like about Ing: handicap stones (free), suicide (partially - I prefer to see suicide without any restriction), life&death (no special positions),
  • converting lost time in points is a needless complication (Ing Timing: why 2 pts, why not 1,3,10?) (and Ing rules seem to be full of complications)
  • specifying rules of conduct is ballast (see e.g. Ing rules, Art. 10). as a result, such rules make it harder for players to read them
  • just a side note about Ing rules: the counting of stones smells like they came from someone who lost too many stones that disappeared under mysterious circumstances...
  • Ing, possibly: although not convincing, the proposed komi system;
  • What I don't like about Ing: extra hardware (frames, special timer), counting requiring an exact stone count, rules conduct,tournament etc, the whole ko stuff

Ikeda's Rule ([ext]

We have seen that territory rules have merits, and these can be perceived as helping go to advance from its primitive rules. But the advance was carried too far, and created difficulties in codifying the rules.

  • does not allow suicide
  • [ext] taiwan rule smells very close to White-plays-last (AGA). "A concise but complete definition would be highly significant; it would put the finishing touch on area rules. " AR/III? (see also button go.)

Aga rules

French Rules

New Zealand Rules


RJ is a big hurdle one has to take before claiming significant knowledge on rulesets...

Japanese rules are difficult to understand for various reasons. One of them is exceptions. It is only a slight exaggeration to say that professional Japanese rules consist of exceptions. Japanese rules are not as elegant as Territory Scoring rules could be but as ugly as tradition and careless professionals have made them. ([ext] Commentary on the Japanese 2003 Rules)

  • suicide: (possibly obsolete points of view)
  • RJ deserves his own wiki, or should initially publish under his home page; I think that flame wars on SL are not far away, unless RJ changes his approach. It's a difficult exercise for someone who wants to share ideas with the world. IMO the trouble is that much of RJ's material gives the impression of being common knowledge, while it is the fruit of own, painful research that at least gives the impression that big chunks of it still need to be accepted/understood. It reads difficult and will therefore have difficulties in being accepted by the public (if that were a criterion...); as an example, here's a random quote "A player can force something if he, moving first in imagined alternating play, has at least one strategy that achieves it regardless of the opponent's replies." (Mathematical Term Force). Not the kind of phrases that I expect in the end product (where rules will be clear and universally accepted... but maybe this is not RJ's target). A similar issue exists with the "Rules FAQ" @rgg: in that document, "beginners" should read "beginning rules researchers" and I would be happy; now, as an example, "beginners" are suggested a way of teaching go that is not universally accepted (and therefore does not belong in a public FAQ). See also [ext] this post. I suggested RJ to start his own wiki (because this has some clear advantages over classic websites, such as ability to update from anywhere, and change tracking), to create more uniform output (rather than the current bunch of text, HTML and PDF, that is mixed with material unrelated to go and that leaves the impression that part of the material is - unintentionally - only reachable via some link on some page). See also p:7002. Another criticism is that RJ's own pages do not always make clear that it are his own interpretations of existing rulesets or proposals for alternative rulesets, see e.g. [ext] New Amateur-Japanese Rules : in this page, the "unwary" might think (s)he is reading a text on Japanese rules, and probably end up rather confused. As a side issue, on the [ext] commentaries page there is a reference to someone named "Winfried Borchardt": who's that? Another example: [ext] Japanese 2003 Rules looks like official rule text, but "axioms" must be RJ's. One has to read the commentaries to find "# are not official rules but invented by Robert Jasiek,...". Invented? So they're not even attempts to express J2003 in a different language then? There's also my p:6586.

Stone counting

Japanese Rules

Notice that Go Seigen already drew the attention to defects in the rules (see Rule Disputes Involving Go Seigen).

Lasker-Maas Rules

Spight Rules

Button go

  • A refinement of pass stone concept?

Two Button Go

Kee Rules Of Go

  • "Retrospective prohibition" is giving headaches

HermanHiddema/Reverse Button Go

  • problems linked to pass fights

Go associations





  • Switched in Aug. 2004


  • Switched in 2006

French Go Federation



Shimada Takuji
+allows suicide +superko +pass stone equivalent? -seki -T counting
BenjaminTeuber / Guide To Become Strong
where the introduction of Japanese rules could have created this split
And then there are beasts like Disturbing Ko, Superko Anomalies...
See also HermanHiddema/Local Global Relations In Rules Disputes
Anti Seki
HermanHiddema/Go and Chess, a comparison
Thus go requires that a player understands some very basic tactics before really understanding the goal of the game. A sort of catch 22 of the rules of go.
Shimada Takuji
No Result
how odd can Go be, that games might even end without result... And what about [ext] this game?


As a result, I think that stone counting is so close to territory scoring, that the latter will become the better way to introduce the game to newcomers. Also, to me it is obvious that the rules will become more unified in a nearby future: many texts seem to go into the same direction (away from Japanese rules!). The biggest resistance coming from Asia, because of tradition. And from the old player, not finding it necessary to change anything: it's the eternal struggle of life, where change is rarely welcomed (meanwhile people forget that life==change).

I have been musing over these issues for a while now, and I'm increasingly overwhelmed by the feeling that Go rules are FAR from beautiful, and there is no solution in sight.

[1] the "beauty of omission", (a criterion possibly more appropriate to Noh opera than to a worldwide game of strategy), [ext] Bill Taylor, John's Go page. (Personally, I interpret boo as territory scoring, and getting a penalty for playing in own territory. "Yes, but in what phase?", yes, I know...)

Axd/Rules last edited by on October 15, 2015 - 20:52
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library