Pondering Upon The Rules Of Go

I was musing about the topic of Go rules with my wife one time. I had tried researching to out what the original rules were. I was surprised to find that the best facts we have are the partial and potentially incomplete records that have survived history. But nobody knows for certain who originated it or when. All we’ve got are traces of physical evidence scattered over time, and numerous anecdotes spread throughout various written records. But nothing truly definitive.

As Go moves throughout history it is continually modified by people, in an attempt to address perceived issues. Plus there is the way that different cultures and time periods play the game. That is, Go is still the game of Go, but there will be cultural influences to how people play (like Go’s history in Japan) or people simply aren’t very good at it (like in the Wu Diagrams).

The top modern rulesets today are all pretty similar and each sets out to fix something that is wrong with the game. Most use Komi, but that is a rather modern notion, coming into existence in the 1930’s. But how do we know that we are even playing Go?

There are a lot of rulesets that we know about in the modern day, but most of them are also of modern origin. Who knows how many have existed that have not survived history. Each one changes the game in small ways and some change it in big ways. Some turn the game completely on it’s head. Others add players, some add new colors, and several change the design of the game board. Each game offers new strategies, depth, vectors for mastery, and new shapes and patterns to interpret and memorize.

The 9x9 itself isn’t even considered Go. It is a Go variant, just as much as Double Board Go or Atari Go. What we accept today as the true Go, is nothing more than what was has remained popular over the many years Go has existed. Had those in power favored a different kind of Go at any point along the way, we all could easily be playing TriGo, scoffing at those who favor square boards.

I’ve actually struggled with this realization a bit. Seeing little tweaks that people have come up with to modify Go, has left me wondering if spending a great deal of time “mastering” one version is in my best interest. Folks barely adjust the wording of rules, which in turn render substantial implications on the board to tactics and strategy. Perhaps I would be better served exploring their creations, or could myself try playing with the rules of Go to see how that affected the general experience.

I find that prospect far more desirable than to continue memorizing a single way for the rest of my life. I can see the value and the cool factor in a lot of the ideas I’ve come across. In the same way that I find value in so many different abstract games. I’m not walking away from Go by any means, but expanding my horizons surely couldn’t hurt :hugs:. I don’t have a point really. I just think this particular angle on the world of Go rules doesn’t get much attention, but is worth thinking about :blush:


Check out Ancient Chinese Rules and Philosophy if you haven’t already.

True enough, Go’s invention is lost to history and we don’t know for certain how it was played at the time of invention. But the earliest reference points (e.g. the Wu Diagrams [~AD 230]) show a game not so different from modern Go. The principles of (1) placing stones (2) alternating with the opponent (3) on the points (4) of a square grid of intersecting lines (5) with the object to surround territory, have remained through the centuries unchanged (the Wu Diagrams feature a 19x19 board, too). The known differences from modern Go were set stones, group tax, and komi. Set stones and group tax are additives. The basic concept of Go doesn’t require them, and Go is perhaps more elegant without them. Komi, meanwhile, offsets the first move advantage in an unobtrusive way (just grant sufficient extra points at scoring). These changes seem to me far less than for other strategic board games like chess, which has featured different pieces with different powers (unlike Go), in differing configurations (unlike Go), with “safe squares” where pieces can’t be captured (unlike Go).

Since Go is in my mind relatively unchanged, I feel a connection to the past when I play it, and it gives me a lot of personal satisfaction in addition to being plain fun. Your mileage may vary; this is only how I feel about it. Go variants are definitely a blast, too. We had fun with the shape game at club yesterday.


You say that group tax is an additive, but it makes sense from a scoring point of view.

It’s been suggested that the original form of scoring is stone counting, which I understand as being “put as many stones inside your territories as you can, and then count how many stones you have on the board” – so, similar to area counting. Why, then, should it be presumed that you should be allowed to put so many stones in your territory that you fill your own group’s eyes? To me it is more natural that the two eyes should be respected as necessary, meaning they are not allowed to be filled, meaning that two stones cannot be placed there and thus the player cannot add two points to their score.

Using this logic I would argue that “group tax” is neither, in fact, a tax nor an additive but a wholely natural part of the stone counting paradigm.


ofc, this is in line with my little pet hypothesis that modern go originated as capture go (a simple game like gomoku), but players either found that this was less interesting because fights were much more rare, or they noticed the cool potential in the form of snapbacks.
Then the rules of passing were introduced so that there was now a “game end” condition, and scoring (probably starting out as most caps, but maybe changing to most stones so there didn’t need to be an unnecessary sac move) was used instead as the victory condition.
Then players found that the most interesting game was played with each side having two diagonal hoshi, which led to that being the rule (like how chess evolved the option of a pawn going two spaces in the beginning to lessen time), which the Japanese later removed.
Then the somebody found “hey, why bother spending all that time filling in stones and just count the empty space?” Which then led to the followup “why bother capturing those stones in my territory if we can both agree they’re dead?” and while the Chinese and Koreans used group tax to make the scoring equivalent to the older version for a while, it was dropped because it was just an extra complication at that point.

Ofc, this is all based on the idea that “classic games evolve the unfun parts out over time”


hmm, I can see that basic idea.

This kind of thinking is exactly the style of pondering I’ve had going on in my head. I really enjoyed hearing your take on this subject. And I agree with you entirely. I truly enjoyed your thoughts. Thank you for taking the time to write them up @mekriff :pleading_face::smiling_face_with_three_hearts::hugs:

1 Like

People would have realized this immediately. I’m guessing that (under stone scoring) people would have often taken this shortcut once they realized that the game had reached the stage where all that was left was to fill in territory with stones.

I think stone scoring was preferred for its pure simplicity, which is elegant in that right. I mean, the objective can be simply stated as putting as much stones on the board as possible. There’s no need to formally define and distinguish between life/death/seki, territory/dame, or even talk about (in the rules) removing dead stones before scoring. Those concepts are removed as technical definitions in the rules, and instead exist only as strategic considerations to be determined by the players. If stones are dead, it’s up to the other player to capture and remove them. If territory is firmly controlled, the player can occupy it.

Stone counting requires more effort to explicitly follow it out to actual counting, but many games would have still ended via resignation, since a player would realize that they have lost.

Ultimately, this leads to territory scoring (as done in Japanese and Korean rules). I used to feel that Japanese rules were more elegant since it forces play to end without actually capturing strategically dead stones (in order to avoid ruining one’s score), however, I’ve come to realize that life and death determination can be incredibly complex, which forces Japanese rules to become a complicated and inelegant mess in order to formally define how to determine life/death/seki.

1 Like

Like so many, my first exposure to Go was with Japanese Rules.


Starting with a Japanese Language teacher who didn’t actually know how to play properly and many years later, ‘An Introduction to Go’ by Davies/Bozulich which saved me from Go oblivion. :grin:

Though I would still not claim to know them completely (see Japanese Rules 1989) and yet, after playing for a while on OGS, I have developed a preference for Chinese Rules primarily because I find it easier to resolve games under Chinese Rules without involving a mod and secondly because I find it easier to explain Chinese Rules to a beginner.

And now to pondering… I am going to be something of a hypocrite because on the one hand I would write the Rules of Go slightly differently if I had a mandate to do so and on the other hand, I would prefer less variety existed with Go rules.

My ideal rules would be fairly simple:

  • Area scoring
  • Suicide forbidden
  • Komi by arrangement but typically integer komi (with draws) except for certain types of tournaments.
  • No super-ko rule. Three-fold repetition leading to a draw.

So essentially Chinese rules except with integer komi because I find a draw to be an extremely fair, satisfying and honourable result and no super-ko rule because at my level of skill the super-ko rule is essentially a coin toss anyway so a draw is a fairer result. I daresay that at higher ranks the super-ko rule can be coherently worked into strategy by the players but not by me.

Having said all that; I believe it would be significantly better for the future of the Go playing world if we could simply agree to one set of rules. For me the differences in the standard rulesets are not significant enough to justify the damage done by the confusion. For such a cause, I would agree to ANY of the regular versions. Imagine if the Nihon Ki-in and Hanguk Kiwon agreed to simply use Chinese Rules. Some might say that Japan and Korea would lose face by such an action but one might just as easily say that China would lose face by not having been the first to act so generously and positively, for the benefit of future Go playing generations.

Just my two-bobs worth. Still pondering. :thinking:

1 Like

Unfortunately, the divisions regarding the rules seem to be quite firm.

Here is an interesting statement from the International Go Federation:

The careful diplomatic wording suggests that unifying go rules is a politically intractable and sensitive issue. I think it may go beyond just politics and may be a matter of national pride and cultural heritage.


Hoshi are star points, yes? How are star points a part of any ruleset?


Is this not also the case with Chinese? Under Japanese you count any remaining group without two eyes as prisoners, but you still need to remove them before counting, when you begin shifting stones into clean groups, right? I’ve never played a real life Japanese game, so I have no experience shifting stones. But given dead stones would end up covering what should be empty territory, it seems to make sense to remove them.

In Chinese you can leave dead stones on the board and simply count them. No need to remove them, since they will be counted as stones. When in reality the player is being rewarded a point for each territory a dead stone occupies. I feel like I’m missing something obvious :sweat_smile:


Can you elaborate on your thinking behind this preference?


A position has just occurred on the board. Now if it repeats one more time, is that repetition the first or second occurrence of the pattern according to your three-folder rule?


Can you elaborate on what you feel the damage done is? I find that most people I talk to only know about a single ruleset. The rest literally are shrouded in mystery. I too feel this is damaging, but I have trouble quantifying it in the face that so few seem to understand the difference between rulesets. So while I would like to cite confusion creating disharmony or turning people off Go somehow, I don’t know that is actually the case :thinking:


As a logophile your use of the word bob here caught my eye. But after looking it up, I couldn’t figure out how you were using it. Any chance this is slang or were you just being playful with language?


This is exactly what I think lies at the root of coming up with agreeing on a single rule set. Which is a profound statement considering how most people view board games (any) as trivial; the play things of children :hugs:

Easy, you look back at really old chinese games (like Huang Longshi old), and you’ll notice that there’s a certain pattern on the board before even the first move – both sides have two 4-4 stones on the board diagonally opposite from each other. Like a mandatory opening.

1 Like

:pleading_face: That is extremely interesting. I had no idea :smiling_face_with_three_hearts:

Not really. The game of Go works perfectly well with or without this rule. For those who don’t know what we’re talking about: Suicide a.k.a. Self-Capture. For me, allowing self-capture is a counter-intuitive complication to the rules that doesn’t add anything to the game.

Okay. Let me revise my ideal rules thusly: " • No super-ko rule. Irreconcilable repetition leading to a draw.

You can define that however you like. Playing IRL most people would struggle to correctly implement a precise procedure for it but nonetheless everyone is capable of recognising it (eventually). So I’m not too concerned with precisely how it’s implemented. Three-fold repetition with ‘who’s turn’ and all ko-status included seems sufficient to identify irreconcilability to me.

[Confusion of multiple rulesets]

  1. It confused the hell out of me and could have ended my Go playing before it really began.

As previously mentioned, I first learned Japanese rules from someone who didn’t understand them and was finally rescued by Davies/Bozulich ‘An Intro… to Go’ BUT in between these two events I found for myself ‘The Game of Wei-Chi’ by Daniele Pecorini and Tong Shu which confused the hell out of me because it was describing Chinese rules (and scoring) and I didn’t even know there was a difference so I was trying to reconcile the irreconcilable.

  1. It needlessly complicates the process of introducing new people to Go.

Taken from AdamR’s excellent new tute:

Different rulesets in Go >

And here comes the curveball but I had to tell you sooner or later. These things I just taught you apply for Japanese scoring… Yeah. Go has several rulesets and Japanese are arguably the most popular in Western Europe. Does that mean you cannot play with someone who was taught Chinese scoring? Luckily, no! All the basic principles are the same and unless you played something very weird the result will be the same as well. So, it almost does not matter what rules you know. If I were you, I would not bother with the differences right now. If you are really, really curious about the differences, you can check them out here.

  1. I would rather time was spent discussing strategy and the game itself. Just look at how much time & energy is wasted explaining the differences and implications of the differences between rules here on the forums.

Monetary slang. cf. ‘two-cents worth’.

Author’s note: While I’m not keen on the multitude of rule-sets for Go, I have no aversion to variants that are clearly identified as variants. eg. Blind Go, One-colour Go, Tibetan Go, Vorlon Go, Capture Go, Thue-Morse Go etc. In fact I often enjoy trying these new angles.


even though i started with J rule and it is more popular in the world, here is how i see the relationship between J and C rule:

C rule is much simpler, there is no confusing ad hoc judgement, no ambiguity.

The ambiguity of J rule, to me, arises out of an attempt to use a different set of rule to TRY TO REPLICATE THE SAME RESULT OF C RULE.

if they are willing to accept that a different set of rules can result in different result, then most of the ambiguity is easily resolved. if one side needs to fill in his own territory to kill the opponent’s group, then let it be.

but because J wants to replicate C result, and the math doesn’t add up when you have to fill in your territory to kill, or some other situation, they set up ad hoc rules to decide which one is alive and which is dead, like the “bend 4 rule”. this is the root of all the problem: TRYING TO REPLICATE THE SAME RESULT OF C RULE, BUT WITH A DIFFERENT SET OF RULE.


That has indeed happened. There’s a famous game from Cho Chikun where he was losing, so he figured out a way to force a triple ko that led to super ko. This forced the game into a tie since it was Japanese rules. He figured a tie was better than a loss!


I agree that area scoring rules are much simpler, since the territory scoring in Japanese rules seems to require a complicated and counter-intuitive procedure to resolve life and death.

However, I don’t think it’s entirely accurate to say that Japanese rules are trying to replicate the same results as Chinese rules. There are some rare cases where Japanese rules produce very different life and death determinations. I think some of those cases might actually be by design, in order to keep in line with historical and traditional precedence.

For example, Japanese and Chinese rules behave very differently in a case that I discuss in this thread:


I found an interesting article about the history of go rules:

Here are some key points that are relevant to the discussion above:

The modern Japanese and Chinese rules (or any go rules for that matter) were not formally written down until the mid 20th century. Before that, they were both passed down through oral tradition and understood conventions among the go community.

Go was introduced to Japan about 1300 years ago (or maybe even earlier). This means go would have a spread into Japan via an ancient form of stone scoring rules.

The modern Japanese rules did not evolve from (and are not an attempt to replicate) the modern Chinese rules. They both evolved (and since diverged) from am ancient common ancestor that employed stone scoring.

The group tax effect of stone scoring may have been the historical influence for the Japanese rule that eyes of groups in seki positions do not count as territory.

Stone scoring persisted in China until the beginning of the 20th century, while the Japanese rules seems to have eliminated the group tax effect for living groups much earlier. Perhaps, that Japanese development even influenced the modern Chinese rules.


WHAAAAT?! :scream::exploding_head: (mind blown)

This is an amazingly cool find @yebellz. Thank you :sob:. Sooooooo coool :nerd_face:

1 Like

And this is why I dislike Japanese. They eliminated the eyes group tax but not the seki tax. It’s half stone counting and half modern. Makes no sense. :frowning:


I wish we had the reasoning behind their design decisions, the way we do about Komi.