Pure Go

[John Fairbairn]

I was discussing ancient Chinese rules yesterday with Chen Zuyuan, the world's premier rules expert, and he startled me by recommending the Japanese "Pure Go" to teach children go in the West.

I had heard of it, but only quite recently when O Meien gave it his thumbs-up, too, in his 2019 book on go rules of the word, along with an extensive write-up. But I had not bothered reading the detail, despite being in my second childhood, partly because it was for kids but mostly because it was about rules.

However, Chen's recommendation made me take a (brief) look, and it does appear to have a long pedigree, extensive support and much success. Which made me wonder: how come I've never heard of it in English? Of course I may have just missed it because I tend to look the other way as soon as I see the word "rules", but I have actually (and grudgingly) been involved with some attempts to teach go to children, not least some events organised by Peter Wendes in England. One such event (attended by Japanese pros Shigeno Yuki and Shinkai Hiroko) was at the Imperial War Museum in London. My involvement was to write a pamphlet for the museum on the Atomic Bomb game - the children played next to a replica of the delivery bomb! As with most things Peter did, it was a big success as regards numbers. I also discussed children's go in the a fair bit with Peter's excellent US counterpart, Ernest Brown, who was also familiar with the Taiwanese method of teaching go as a variety of fishing.

So there were several potential sources for me to hear about Pure Go: Japanese pros and Peter/Ernest themselves. Plus a third, perhaps, if you add that this was in the middle of the Hikaru boom.

Then, of course, there is the fact that I read a lot of Japanese books and magazines. I imagine there must have been something there and I just skated over it. In fact, though, I do have a vague memory of go being taught in a successful way in an old people's home in Japan about 40 years ago, but in my mind that's associated with capture go (another Japanese experiment). However, it seems now that that might have been Pure Go.

Does anyone here know about it?

There is a web site: jungo.go-en.com. You will see there that O Meien has a column and young female pro Hoshiai Shiho is involved (and through her, possibly, the great kids' teacher Kikuchi Yasuro, who was her teacher, may have some input). Even if you can't read the Japanese, the many photos will indicate the apparent level of success. Note that it is advertised as an introductory method, by which you can be up and playing within ten minutes. At some later point you can transfer to Japanese or Chinese rules, as your prefer, and as far as I can see from a quick glance it shares various attributes like that with AGA rules (which O Meien does not mention).

[Javaness2]

This might be because I can't read any japanese, but it does looks like a nice clean presentation to me.

So it's played on a 7x7 board.
The ko rule used is the so called 'basic ko' rather than the supernatural ko.
The winner is the one who has the most stones left on the board at the end of the game.
They show how to capture.
They explain which stones are connected to each other.
Nothing about having to wear a tie.

Certainly worth consideration for teaching juniors. I'd probably still prefer to start with atari go.

[goPlayerJuggler]

It seems to be equivalent to Stone scoring. The SL page mentions that it is "interesting ... for educational purposes. "

What a coincidence, I was reflecting on this topic today too.

In Strasbourg, France it's used extensively with children as the next step after 5-stone capture atari Go.

This video is very good (but in French):
l'enseignment du Go à l'école.

I watched it today it as some children attend my local Go club on Saturdays so we need to improve on how we handle them. I think we'll copy the people in Strasbourg!

[gennan]

I don't really like atari go, so I just start with go right away.
I may start on a 6x6 board with 6-7 year olds and 9x9 with older children.

I first show some "castles" and count the size of the "courtyards" (empty areas inside the castles). These are points.
I tell them that you need to make more points than your opponent to win.

Then I tell that go stones are like living beings, needing air to breathe. And the stones can only breathe through the lines going out from the stones.
Then I show how black takes away all breath from a white stone to capture it (removing it from the board).
I tell them that these prisoners are also worth a point per stone, so you keep them until the end of the game.

Then I show how a stone can extend from an atari to gain new breathing places for the whole (connected) string of stones.
With older children (8+) I may show some more strings and let them count liberties (how many moves it takes to capture).

With a small group of children this introduction takes 5-10 minutes.

Then I let them play, against each other or against children that already know how to play (possibly with handicap).

[tundra]

supernatural ko.

Interesting term you have coined there

(Not that I am criticizing - I would be rather pleased with myself if I had come up with it.)

[NordicGoDojo]

I featured 'pure go' in an NGD post back in May. The concept was quite familiar among Japanese professional players even before Ō Meien made it into a sellable product.

I haven't had many opportunities to teach go to beginners recently, but from what little I have seen it seems like a vastly superior alternative to, e.g., Chinese and Japanese rules. Stone scoring makes the game-end process much more understandable to new players, while still essentially describing the same game as Chinese and Japanese rules (save for the group tax, of course). This my opinion is formed from three university lectures I have held, with roughly 100 participants each.

This brief beginner's guide explains how to play 'pure go' with no price tag for the rule explanation.

[lightvector]

If stone-scoring rules became popular alongside modern rules, I would be fascinated to see what differences it had on high-level human strategic play. As I showed back in this post, it certainly has an effect on the bots, even all the way as early as move 5: viewtopic.php?f=15&t=17674

[Bill Spight]

I featured 'pure go' in an NGD post back in May. The concept was quite familiar among Japanese professional players even before Ō Meien made it into a sellable product.

An interesting post.

I am not sure that the difficulty of understanding the rules is much of a bar to interest in a game, as long as potential players can see players having fun playing the game. Bridge used to be very popular, even though its rules are complicated and often infringed. As for baseball, forget it! Baseball rules are quite complicated.

I was skeptical for a long time about the capture game, but now I think that it is a worthy introduction to territory go, and an interesting game in itself. I don't know what is best for teaching beginners, or even if one size fits all, but I would like to respond to some of your criticisms of the capture game.

One problem is that, presumably, the new players have come to learn how to play go; and then the teacher starts by saying: ‘okay, I’ll first teach you a game which is a bit like go, but different.’ While this teaching method might be effective, it includes an obvious marketing mistake, and probably the students’ motivation to learn is damaged from the start.

My mother was an avid bridge player, and I learned to play bridge at age 6. But none of my friends or classmates learned how to play until middle school. However, we did play a whist variant called Oh Hell that could be played by several players. There were no partnerships, but you bid for yourself, with only one round of bidding. It was a lot of fun.

So particularly with kids, the capture game can be taught as a fun game in its own right. The object of the game is well defined, and it is over in minutes instead of hours. I have an adult friend who learned to play go, but it was like pulling teeth. I taught him the capture game, and he started having much more fun.

Another problem is that the transition from atari go to ‘real go’ will feel abrupt: up until one point you are concentrating on how to capture the opponent’s stones, and then you have to start caring about some difficult new concept called ‘territory.’

Yes, territory can be a tricky concept. For one thing, why are dead stones territory? And how do we know that stones are dead? And why don't you have to capture them?

The capture game, in its no pass variant, is a game of territory. The concept of territory naturally emerges from it. When you have to capture more than one stone to win the game, the equivalence between captured stones and empty points of territory becomes clear. Furthermore, as you increase the number of stones to be captured, the capture game becomes more and more like regular go, so that you can avoid an abrupt transition.

In addition, contrary to my original expectations, the capture game is quite strategic. The reason lies in the ease of making life. All you need for life in capture-1 is a single eye with two empty points. As a result, play on the 7x7 board is not a tactical slugfest where somebody has to die. Tactics is less important than in regular go, strategy is more important. At least on small boards.

Why is the no pass capture game a game of territory? Consider the following position, reached after an improbable line of play.

$$B Equal territory
$$ -----------
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ -----------

Click Here To Show Diagram Code

[go]$$B Equal territory
$$ -----------
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ | . O . X . |
$$ -----------[/go]

Neither player will wish to play inside the opponent's territory, even if they do not have a word for territory, because it is obvious that the opponent could capture it. It would be a dead stone, even if the players do not have a word for a dead stone.

Play continues. It is Black's turn.

$$B Filling the dame
$$ -----------
$$ | . O 3 X . |
$$ | . O 2 X . |
$$ | . O 1 X . |
$$ | . O 4 X . |
$$ | . O 5 X . |
$$ -----------

Click Here To Show Diagram Code

[go]$$B Filling the dame
$$ -----------
$$ | . O 3 X . |
$$ | . O 2 X . |
$$ | . O 1 X . |
$$ | . O 4 X . |
$$ | . O 5 X . |
$$ -----------[/go]

The players fill the dame, even if they have no word for dame. They are forming the concepts of territory, dame, and dead stones.

Play continues, White to play.

$$Wm6 Playing out the game
$$ -----------
$$ | 3 O X X 4 |
$$ | . O O X . |
$$ | 1 O X X 2 |
$$ | . O O X . |
$$ | 5 O X X 6 |
$$ -----------

Click Here To Show Diagram Code

[go]$$Wm6 Playing out the game
$$ -----------
$$ | 3 O X X 4 |
$$ | . O O X . |
$$ | 1 O X X 2 |
$$ | . O O X . |
$$ | 5 O X X 6 |
$$ -----------[/go]

The players take turns filling in their own territory. After White has no move but to play another stone inside his own territory, and then Black will capture all of his stones. So White can resign.

There is a natural place to stop play in the capture game and score the game. In this case it is here, where neither player wants to play inside their own territory, much less their opponent's.

$$B No more dame to fill
$$ -----------
$$ | . O X X . |
$$ | . O O X . |
$$ | . O X X . |
$$ | . O O X . |
$$ | . O X X . |
$$ -----------

Click Here To Show Diagram Code

[go]$$B No more dame to fill
$$ -----------
$$ | . O X X . |
$$ | . O O X . |
$$ | . O X X . |
$$ | . O O X . |
$$ | . O X X . |
$$ -----------[/go]

At this point, the players have to start filling in their own territory. Each player has three moves before they have to play self-atari or possibly play inside the opponent's territory. Since each move occupies a point of territory, we can count a player's territory as the number of moves they can safely play inside it. Each player has three points of territory, i.e., five points minus the group tax. Territory go with a group tax! Who'd a thunk it? (As we know, it was one of the earliest forms of go. )

What about the equivalence between a dead stone and a point of territory? Capture-2 gives us the answer. Now in the position above each player has 4 safe plays instead of 3. Each player can play a dead stone inside the opponent's territory.

$$B Dead stone in Capture-2
$$ -----------
$$ | . O X X . |
$$ | . O O X . |
$$ | B O O X . |
$$ | . O O X . |
$$ | . O X X . |
$$ -----------

Click Here To Show Diagram Code

[go]$$B Dead stone in Capture-2
$$ -----------
$$ | . O X X . |
$$ | . O O X . |
$$ | B O O X . |
$$ | . O O X . |
$$ | . O X X . |
$$ -----------[/go]

In this position Black has only 3 safe moves, whil White has 4. This can be counted as 5 points minus the group tax for Black and 6 points minus the group tax for White. I.e., we count the stone as dead. And we still count it after White captures it, if play continues. So a prisoner is also worth 1 point.

Anybody have any other explanation for why prisoners and dead stones count as territory?

Well, yes, if stone counting came first, but the oldest game records that are counted are counted in terms of territory with a group tax.

[John Fairbairn]

Anybody have any other explanation for why prisoners and dead stones count as territory? Well, yes, if stone counting came first, but the oldest game records that are counted are counted in terms of territory with a group tax.

Yes, the oldest surviving game records are territory with group tax (though probably also with a requirement to play equal number of moves), but the oldest reference to scoring, in the Dunhuang Go Manual, very very strongly implies stone counting was used before then.

Chapter 6 says　碁有停道及两溢者，子多为胜。取局子停，受饶先下者输。纵有多子，理不合计。

So we know both sides fill in their domains to the brimming point (溢) nd that the one with the most stones (子) wins. There is no reference to group tax but they have bothered to make a reference to handicap play. It is conceivable that none was used until, at some later point, someone pointed out that things get messy with sekis, say.

Either way, stone scoring was used.

The Dunhuang manual uses the "Japanese" character 碁. We believe now that go reached Japan well before Tang times, and they would of course have inherited that character then, but we also know that go in Japan in Tang times was almost certainly played in the same way as in Tang China itself. We can infer this from a poem (in Chinese) by the famous Japanese scholar Sugawara no Michizane, who wrote a poem on watching a guest scholar from Tang China play go at the Japanese court ("visiting professors" from China were invited over to teach the Japanese emperor about Confucianism). If there had been any change in the way the two countries played go, it seems highly unlikely that the poem would have been written in the way it was.

Since, at this stage, we have no special reason to assume the Japanese ever used group tax, there is a case to be made that they never had it, because they adopted a pure group-tax-free version from China.

We also know that key term 溢 is referenced in the Thirteen Chapters Classic, in the same compendium that gives the games records mentioned above (the Carefree & Innocent Pastime Collection). It seems more than plausible that the stone-filling basis was still seen as underpinning the territory way of counting. Given that the most famous ancient Chinese book of mathematics mentions go pieces, and there were ancient treatises on computing the number of possible positions on the go board, it is easy to conceive of the possibility that mathematicians even then had "interfered" with the game. They had done the equivalent of the AGA's making the game scorable with "Chinese" or "Japanese" rules - and with the requirement for equal moves, presumably they had also discovered button go?

There is a lot of speculation embedded there, and we could add more, e.g. were there different concurrent versions of go (as seems likely)? But what seems almost unequivocal was that stone scoring was known before the oldest surviving game records.

[lightvector]

If you care about ancient Go rules history, I found this reference to be a fascinating summary/analysis: https://www.usgo.org/sites/default/file ... orules.pdf

At least just reading this analysis, suggests to me that stone scoring was the origin of Go, rather than the capture game. This is partly because of the textual analysis of "overflowing" mentioned in the earlier parts of that article, but also because of the emphasis in early texts on "equal stones" and only counting "moves by each side" in addition to the group tax rule. "Equal stones" seems unnatural or arbitrary in the context of capture Go, whereas it is extremely natural in implementing stone scoring by territory counting. This is because it suggests that the ancients understood the mathematical relationship in stone scoring that if both sides have made an equal number of board plays, then you can determine who will be able to fill more stones on to the board equivalently by counting the empty territories and captures and taxing each group by 2, and comparing that count instead, without actually spending the time to fill all the stones.

Of course, as you see later in the article, if you guess that this was what was going on, you see that the evolution into Tang dynasty rules ended up misunderstanding the "equal stones" principle. The counting method itself became the definition of the score, rather than a shortcut to computing the score. But then if the counting method of territory + captures now becomes the definition of the score, then the "equal stones" principle, in requiring a player to either hand a prisoner or to fill in and lose a point of their own territory if the number of dame is odd, intuitively seems discordant and unnecessary - why should a player lose an extra point? So perhaps now you simply enforce "equal stones" by having a player not fill in the last dame! But as seen in the Tang era with this scoring method, simultaneously with this alteration to "equal stones", of course the equivalence with stone scoring is now broken. And now that the equivalence is broken, and with dame moves not worth anything, the entire "equal stones" concept seems unnecessary, so it gets removed later after that.

So the equal stones concept, and therefore the connection to stone scoring, is lost. And of course, once the connection to stone scoring is lost, the justification for group tax also vanishes.

But anyways, the fact that there was an "equal stones" principle in the first place, even one that got seemingly misinterpreted and applied in a way that broke mathematical equivalence, suggests to me that stone scoring is the more natural origin relative to capture Go. Because if you try to use territory counting to implement stone scoring, you find exactly that equal board plays is a rule that you want in order to make the mathematical equivalence. (And also there are references to filling up the board, and also there are not any overt ancient references to counting and comparing numbers of captures or reaching certain numbers of captures).

(edit: John Fairbairn got there first )

[Bill Spight]

Well, OC, we are speculating.

However, if the group tax lies in stone scoring, where did it come from? OC, it could have come from anywhere, like the rules for scoring Sunjang Baduk allow some stones to be removed from territory but not others, and the Japanese rules do not count territory in seki. But if all you are doing is filling in your territory, why not fill it all in, as would be consistent with modern area rules?

It makes sense, however, if, at least in principle, you are playing stones inside your territory, and not just placing them there. If that is the case, then the group tax applies equally to territory go as to area go. You don't have to assume that either one begot the other. Every form of no pass go that has capture (Edit: as we know it) has a group tax.

Furthermore, you do not need an equal stones provision to transform one form of scoring into the other. Besides which, as we know, the equal stones provision is known to have applied when there were still dame on the board. In which case it does not transform territory scores into area scores. It is philosophically and esthetically pleasing, however.

[pgwq]

Actually, by my research, stones mentioned by Dunhuang Go Manual are not living stone on the board,they are captured stones.

[pgwq]

<Lan Ke Jing>

[RobertJasiek]

if you try to use territory counting to implement stone scoring, you find exactly that equal board plays is a rule that you want in order to make the mathematical equivalence.
(edit: John Fairbairn got there first :) )

But I got there earlier (2004) with prisoner counting:)
https://senseis.xmp.net/?PrisonerCounti ... oneScoring

[Bill Spight]

if you try to use territory counting to implement stone scoring, you find exactly that equal board plays is a rule that you want in order to make the mathematical equivalence.
(edit: John Fairbairn got there first )

But I got there earlier (2004) with prisoner counting:)
https://senseis.xmp.net/?PrisonerCounti ... oneScoring

Well, I got there in 1977. I don't think I was the first, either.