This is a much abbreviated version of an article I wrote under the title Go on The Roof of the World in Go World 59 (Winter 1989). It was the first extensive piece on Tibetan go in English. Anthropologist Peter Shotwell of Boston, USA, kindly took it a step further by visiting Tibet. The result of his researches appeared in Go World 69 (Go in the Snow). It seems as if the rules first reported in Japan and China were somewhat more complex, unless it is a case of local variations. There are still some gaps in our knowledge, but we know enough to have a playable game. My original article had extensive notes and a bibliography, both omitted here. A couple of omissions concerning rules in the original article have been restored.

IN NOVEMBER 1958 the Japanese go magazine Kido announced the discovery of go in Sikkim. Shortly afterwards, in March 1959, the Crown Prince of Sikkim went to Japan to attend a Buddhist conference. The Japanese Go Association took the opportunity to invite him to their headquarters, and there he agreed to play a game under the rules of his kingdom - on a traditional 17x17 board - with the 6-dan professional Iyomoto Momoichi,

It turned out that the Sikkimese rules were the same as Tibetan rules, for the Crown Prince had learnt the game from his wife, a daughter of the Dalai Lama. The discovery of this version of go was important because it seemed likely to shed light on how go used to be played in ancient China.

For reasons of etiquette only a few moves were played in the game between the Crown Prince and Iyomoto. White plays first but the game starts with white and six black stones already in position on the third line.

This feature of fixed starting stones for both sides is also seen in ancient Korean go, but on the fourth line. The old Chinese habit of starting a game with two white and two black stones on the four corner star-points may be a remnant of this.

The Sikkimese board was very large and made of cloth. The stones were of the type once usual in China - a sort of flat cone. No similar boards have been found in China except for the 17x17 stone boards found in Han tombs and therefore dated to the first two centuries AD. No records of games played on a 17x17 board exist.

Although not illustrated by this game, the method of counting at the end was said to be the same as in modern China. You count your own territory plus your own stones; captured stones are not counted. Dame have value and so are played out. There was no two-point group tax as there was in old China. But a major difference was that positions like snapbacks and nakades are treated rather like kos: you are not allowed to play back on any point vacated by captured stones until at least one move has elapsed - which makes killing groups none too easy. Sekis were known.

White played first but it was the stronger player who took White. His advantage was offset by komi, but they did not say by how much - probably there was a variable komi in lieu of a handicap system.

The next major development was a booklet called Data on the History of Physica1 Education published by the Beijing Physical Education Council around 1980 which contained an article entitled "Old Tibetan Go", This was first brought to wide attention by Yasunaga Hajime in a Japanese go magazine in June 1983, although the original research was due to the Chinese player Cheng Xiaoliu.

Cheng's view is that in the area now encompassed by Tibet, Nepal, Sikkim and Bhutan go first appeared around 700 AD. That was when the first regular cultural links between the Tibetan peoples and the Han Chinese were established.

Go in Tibet is called mig mang, which means literally 'many eyes', This does not refer, however, to eyes as in go, but to the points on a board, and the phrase really means no more than 'game board'. As such it is applied to several games, all of which use black and white pieces.

Cheng suggests that go has been unchanged throughout the centuries in Tibet. Perhaps the oldest known board there is also a 17x17 board, carved into rock in Changdu, Qinghai Province, which is reputed to be the site of a game played by the legendary king Gesar (hero of the world's longest epic poem) over 1,000 years ago. Cheng says this is still revered by local shepherd folk. They believe that by walking round the stone several times they will acquire wisdom.

Go has never been formalised as a competitive activity in Tibet in the sense of matches, tournaments or championships. Indeed, games would frequently involve more than two players: four or even six, in two teams, each player taking turns to play for his side but freely consulting with his partners.

The game would appear to be rather uncommon, however (we shall see an explanation for this below). Jen Nai-ch'ang, describing go in Tibet as the 'large encircling game' says:

"It is very difficult and poses great demands, so that only very few play it. Each of the two players has 151 stones, one side black the other white. These include six large stones which never move and remain on the same spot. All the other stones can be moved. The whole method and rules of play are very similar to the large encircling game of China, It is said to be played only in Lhasa…."

This article is also the source of a reference by Hummel and Brewster, but they add a photograph of a game in progress inside a tent in Lhasa which indicates that the fixed starting stones are indeed larger than the rest. It is also apparent that the board is probably larger than we are used to, and most of the stones are quite small so that adjacent stones do not appear to be connected. Since White has 27 pieces to Black's 26, we can usefully infer that White played first.

Cheng describes a go custom in one of the old kingdoms of Tibet known as the 'tongue fight'. As each player plays a stone, he must make responses in a predetermined format designed to show how quick witted he is. Speed is of the essence. For example, if player A plays the first stone he may start the repartee with, "This is a rabbit." B might respond, as he plays his stone, with: "My stone is a fox," continuing with A: "This is a panther, " B: "This is a tiger," and so on. I imagine these lose something in translation - perhaps the lines have to rhyme or start with the same letter in Tibetan.

Cheng says that the game was really only popular among the aristocracy. All the Tibetan kings have enjoyed playing, and there are many stories to this effect in folk tales. He mentions one about a king who was wise and strong, and exceedingly good at go. But he was also extremely proficient with his tongue. As he played go, therefore, he would also discuss great matters of state with his senior ministers in the form of a tongue fight.

He had a certain advantage in that he kept birds that could mimic speech. He allowed them to fly around and eavesdrop, and listened to their chirruping when they returned. In this way he discovered that his queen was planning to be disloyal to him. The king at once called for his queen to play go with him, and in the course of their tongue fight he used extremely clever words that implied a rebuke for the queen should she ever be disloyal. The queen was ashamed and blushed. Halfway through the game she complained of feeling poorly and hurriedly withdrew to her quarters.

Thereupon the king carefully analysed the words she had uttered during the tongue fight. Convinced of her treachery, he repeated the ploy the next day, again with the help of his birds, and eventually trapped the queen into compromising herself.

There is another story that after the Gtsang dynasty had taken over Tibet for the first time in the 17th century and tried to suppress Buddhism, the Dalai Lama made a secret pact in 1641 with the Mongol general Gu-shri Khan, leader of the Khoshut tribe, as a result of which Gu-shri threw out the Gtsang. But the Mongols decided to stay in Tibet under Gu-shri's successor, Lha-bzang Khan, with Lha-bzang as king and the Dalai Lama as spiritual ruler. The arrangement worked, and ushered in a period of peace. But when the Lama died in 1679, his trusted retainer Sangs-rgyas rgya-mtsho hid the fact while he sought the next reincarnation of the Dalai Lama (the dead one's soul was supposed to transmigrate into the body of a child) and had himself appointed as minister-regent in the Lama's name. Relations with Lha-bzang became strained, but with both men surrounded by strong military forces they eventually decided to play a match of three games of go, under public gaze, for hegemony of Tibet. Lha-bzang Khan won and eventually killed Sangs-rgyas rgya-mtsho in 1705.

To the highly superstitious Tibetans go then became unlucky, and the game declined among the populace, although it remained popular with the aristocracy, It was their custom, during the festival of 'Playing in the Garden' on the 15th day of the 4th month of the Tibetan calendar to invite their friends to their flower gardens, there to dine, sing, dance and play games - with go regarded as the most important.

As to the fact that go was known to the Mongols, it is interesting to compare the description of their version of the game given by Assia Popova. She describes more or less the same game as in Tibet. The name indeed indicates it reached the Mongols from there. The board is of grey or yellow cloth on which is drawn a grid of 17x17 lines. Six points are marked by black points and six points by white points. There are two kinds of pieces: (a) six black "bulls" and six white; (b) 144 black "dogs" and 144 white, smaller than the bulls.

The game is in two phases, The first is a fixed initial arrangement as a tabia (obviously the Arabic term for the starting arrangements used in Muslim chess) in which the two players alternately put the black bulls on the black points and the white bulls on the white points. Each player then in turn surrounds his bulls with his own pieces. This arrangement divides the board into 12 zones of equipotential influence and greatly simplifies the game.

Then she specifically adds that, in the second stage, the players alternately place their "dogs" on vacant points, trying to form eyes. Once pieces are on the board they cannot be moved. The winner is the one who surrounds the greater number of empty points.

In a follow-up note in rec.games.go, Shotwell said that plays must be within one space of an existing stone along a line (knight's moves and diagonal are not allowed). That rule was not mentioned by the Prince of Sikkim, nor did he observe it in his own plays. Shotwell suggested his game with the Japanese was played with compromise rules. That seems too far-fetched. More likely there are local variations (as he conceded), or the players in the games he observed were weak and just happened to play all their moves close together - that is characteristic of beginners' go anywhere. Regrettably, Shotwell was unable to record any games.