The history and evolution of kyu rank in Go game

The AI analysis finds a lot of errors, but no obvious blunders. I’m 2 dan, and I think that I’d maybe make different mistakes, but be more accurate in some places. I’d put black at about 1 or 2 dan, white about 3 or 4 dan.


I’ve put up for this for long enough, and time to continue our journey.

This is the list of the Hoensha member list in 1894, right after they revert back from pure kyu ranking to the traditional dan ranking on Apr. 1893

Compare this with the 1891 member list, we can find they are mostly the same, except they renamed and group the old 10, 11, and 12 kyu rank into “3 kyu of Shodan” (三級初段), “2 kyu of Shodan”(二級初段), and “1 kyu of Shodan” (初級初段). Effectively make all the 9 kyu to 12 kyu players “Shodan”. And instead of increasing order of “kyu”, it reverted to decreasing order of kyu within the “Shodan”.

The reason behind the decision of reverting back to the old dan system in 1893 is not well documented, but from most sources I can find they pointed to an agreement almost a decade ago at 1886 between the original head of Hoensha Murase Shuho (村瀨秀甫) and Honinbo Shuei (本因坊秀榮 the head of Honinbo house at the time). After the 9th game on Nov 15, 1885, of their ten-game matches to determine who would dominate the Go community (Murase Shuho already won 5 of the 9 games in josen 定先, always played white), Honinbo Shuei proposed to give up his Honinbo house title to Shuho, recognized the rank and diploma issued by Hoensha, and granted the title of 準名人(Jun Meijin, effectively 8d) in the name of Honinbo house to Shuho. However, Shuho also had to recognize and gave Shuei 7d diploma in the name of Honinbo house (Shuei was only 5d at the time of the match), and all the future diploma issued had to be signed by both parties (and split the fee gathered), and Hoensha had to return to the old dan ranking system (they have to sync their rank somehow).

Murase Shuho finally agreed to the term on July 30, 1886, and became the 18th head of Honinbo house - Honinbo Shuho (本因坊秀甫), and he wanted the head of Hoensha to also be the head of Honinbo house in the future, tried to cover everyone into one banner (the first attempt to a nation-wide association). But Shuho died very suddenly on Oct 14, 1886, just 2 and half months after (the shortest term as the head of Honinbo), without leaving any clear instructions as to who will be the next in line for the Honinbo house and the Hoensha.

Since it was so sudden, Shuei was again tried to push for his return as the head of the Honinbo house. But the 2nd hand of the Hoensha 中川亀三郎 (Nakagawa Kamesaburo, the first, not to be confused with his adopted son Ishii Senji 石井千治, who later on also called himself Nakagawa Kamesaburo) disagreed, and already assumed to be the head of Hoensha on Nov 1, 1886. Everything seemed to revert back to the way before the agreement (except Shuei still got his 7d recognition), and Shuei became the de facto head of the Honinbo house again. Although when Shuei issued an open challenge to everyone for his Honinbo title, Nakagawa Kamesaburo didn’t accept and just kept expanding Hoensha for more branches all over Japan and start recruiting more players from other houses and more young players through their development programs (their Insei system). The Osaka branch of the Hoensha was established in 1887, and 青年研究会 (young study group) of the Hoensha was established in 1889 (the sort of official Insei system for Hoensha).

On Mar 26, 1893, the headquarter of the Hoensha moved to a brand new building, and a ceremonial game was played between Iwasaki Kenzo (巖崎健造 the vice president and the 2nd hand of the Hoensha) and Shuei (at the time, he proclaimed to be Honinbo Shuei, but not everyone recognized it). Iwasaki Kenzo was famous for being a “thoughtful” player (he played very slow), and the ceremonial game only went for 21 moves and suspended. When the Hoensha wanted to print this game in their newspaper/magazine for the next month, they wanted to list Shuei as a Hoensha 3 kyu, but Shuei wanted to use Honinbo house ranking, and made a big deal about his old agreement with Shohu a decade earlier in 1886, and demanded they honor it.

At this point, for some reason, perhaps related to the deteriorated health of the executive Kobayashi Tetsujiro (小林鉄次郎 who was one actually running the day-to-day task), or perhaps the year earlier in 1892, Shuei started 囲碁奨励会 - the Go promoting group (became 四象会 later on), and it acted like a monthly tournament with prize money (which Yasuhisa Tamura 田村保寿 the later famous Honinbo Shusai won the 1st round, and this is like proto pro games with sponsors), and Honinbo house started to regain their fame where many strong players didn’t just play games within Hoensha, but between the Hoensha and the Honinbo house members. In any case, Hoensha seemed to finally agree on this term, and on Apr 1, 1893, the next issue of the Hoensha newspaper started to list games using the old dan ranking, with a reverted dan ranking member list shown above the next year.

I personally believe that the death of Kobayashi Tetsujir (小林鉄次郎) on Nov 7, 1893, played a major role in this. The concurrent ranking system worked fine for more than a decade, and Hoensha seemed to be at its height at this point. But the original member of Hoensha started to age and died, and Nakagawa Kamesaburo was more of a diplomat. With Kobayashi Tetsujir gone, Nakagawa Kamesaburo seemed to want to lead the Hoensha more like a traditional house.


Interesting! I never knew about this tension between Shuei and Nakagawa.

Or that “Nakagawa Kamesaburo II” was adopted.

Since you’re on the topic, could you clarify the initial relationship between Shuei and Shuho, please? Sensei’s Library uses a strange phrase, saying that Shuei was Shuho’s “semi-younger brother”.

1 Like

Shuho has no blood relationship with Shuei. Shuho (born 1838, originally named 村瀨弥吉 Murase Yakichi) lived next door to the Honinbo house, and by chance being discovered by Honinbo Josaku (本因坊丈策 the 13th head of Honinbo house) when he was 8 in 1846. And one year later when Jusaku died, Honinbo Shuwa (本因坊秀和, originally named 土屋俊平 Tsuchiya Shunpei the 14th head of Hoininbo house) included Shuho as one of his pupils. Shuho was contemporary with Shusaku and was the 2nd or 3rd in line for the head of Honinbo during the 1860s, right after Shusaku, and on par with the 3rd son of Honinbo Jowa (本因坊丈和, the 12th head of Honinbo house) - Nakagawa Kamesaburo (中川亀三郎 born 1837 he was adopted by one of his uncle’s family, all of Jowa’s sons were given to other families).

The Honinbo (本因坊) name is not technically a “family” or “clan” name, but originally just the name of the house where the first Honinbo Sansa as a monk lived in the temple. 坊 indicates it is a yard and a place. Historically until Shuwa, the Honinbo house had never been a heretical hereditary clan, but the next heir would be selected based on merit, the strongest pupils would be “adopted” as the next titleholder (a cross between brand name and a corporate name, but using feudal laws). However, after Shusaku died unexpectedly, Shuwa for the first time, selected his own eldest son - Honinbo Shuetsu (本因坊秀悦) as the heir (who was only 14 years old and 3 dan at the time).

Shuei (born 1852, originally named Tsuchiya Heijiro 土屋平次郎) was the 2nd son of Shuwa, and was adopted by the house of Hayashi (林家) when the original heir died the same year as Shusaku in 1862, and given the name Hayashi Shuei. Although mostly he still considered his blood family over his “adopted” family house. And he was also the reason why the Hayashi house was later merged into the Honinbo house. As for his trip with Shuho started from 1872 to 1875, it is hard to tell. There were few records as to what they were doing in those three years, just that Shuho met several important people that later on help to establish the Hoensha. From their own accounts, they seemed to be pretty close with each other.


heretical clan

I think you mean hereditary.

Great post, btw! Thank you.

1 Like

I cannot edit my first post to add a link to my next reply, should I make a new post as part II and link this post for a new index? Or just keep the content in one post?

Thread management decisions are surely up to you.

I remember I asked before on how to edit and there is a trick to change the post to a wiki (maybe involving an admin to do it?) do your changes and then change it back to a normal post.

1 Like

I just simply couldn’t edit it anymore. Maybe a post can only be edited within certain time limit after it was posted?

Yes, exactly. Let see ( @Vsotvep ) if a mod can change it into wiki?

1 Like


1 Like

@claire_yang you should be able to edit it now, and when finished our wonderful mods can lock it back (as you want).

Before I finish digging into the transition between the 1890s and 1920s, I found someone made a relationship graph of all the key people involved in the creation of Hoensha

The red lines mean blood relationship, the blue lines indicate teacher-pupils, the purple lines are adoption, the green lines are marriages, different colors in the blocks indicate different houses.

I feel it is definitely interesting enough, if someone wants to write a novel, or make a tv series, or movies out of this love-and-hate mess where modern Go born out of it.


Perfect picture to hang in a go school headmaster office.

1 Like

While I was researching the early 1900s Go players all across Japan, I found this very interesting list

Notice the people circled in red

Inukai Tsuyoshi (犬養 毅) and Toyama Mitsuru (頭山 満), the first one Inukai Tsuyoshi was a cabinet minister at the time, and later Prime Minister of Japan (where his death by assassination indirectly lead to the progress of War World II in East Asia), and the second one Toyama Mitsuru was a key player in East Asia “spy network”, where every event regarding Korea, China, and Japan between the 1890s to WW2 had something to do with him.

Although it was not uncommon for nobles, businessmen, and industrialists to sponsor Go since the 1870s (after the Shogunate failed, and Go communities looking for new financiers), and list them as guests or honorary members, while these sponsors themselves are usually novices. But I actually found a game played by Toyama Mitsuru against 吉田俊男 (Yoshida Toshio, son of 吉田半十郎 5 dan - a founding member of Hoensha, and Toshio said to have the strength of at least Shodan)

I transcribed it below to correct an error

From the look of it, Toyama Mitsuru was a pretty strong player himself, not just a simple sponsor. There is even a picture of him playing a 13x13 game with Inukai Tsuyoshi

From all the information I found, there was something fundamentally changed during the 1890s to 1910s in the Go community in Japan, and most likely the real root of amateur Go community started to emerge.


I found these pdfs at Evolution of Go - #5 by Oni

The History of Go Rules, Chen Zuyuan 2011

The Game of Go: Speculations on its Origins and Symbolism in Ancient China, Peter Shotwell 1994–2008

In the book Masterpieces of Handicap Go, Volume 2 (thanks to TPR for my receiving it), there is a review by Kobayashi Koichi of Inukai’s
four-stone victory against Iwasa Kei 6p “at the end of the Taisho Era (1912–1926)”.

Months ago, when I started to trace back the most extensive list with almost 5000 players published in 1908 (明治41年)

I successfully traced it back to older documents like this, published a year before it at 1907 (明治40年)

Or this, published 2 years before at 1906 (明治39年)

囲碁名鑑 西南部 (兵庫県、中国・四国・九州地方) 明治39年 pg2
囲碁名鑑 西南部 (兵庫県、中国・四国・九州地方) 明治39年 pg1

With even older documents like this at 1901 (明治34年)

Or decades before at 1892 (明治25年)

Or like this with no date information, but based on known players on the list already passed away before the 1890s, should be around the 1860s to 1880s

When I dug deeper and further in time in the amateur Go world in the late 19th century, I realized that it is simply impossible to trace thousands and thousands of people who mostly didn’t leave records in history. I need a different approach to get an overall view of who they were, how they lived, and how in general people learned to play Go at the time.

Some time during the summer vacation, I chatted and talked with one of my friends who know economic history that deal with activities and groups of people in the past, and they gave me the idea of analyzing quantitative data to better understand the context of the era instead of investigating individual person’s activities. So I began to ask some very fundamental questions:

  1. What was the population of Go players in the late 19th century to early 20th century and how did they change over the decades

  2. How did amateur players learn Go at the time, and who could afford to play? Can someone be self-taught without a teacher?

  3. What were the local activities like Go clubs and conventions during that time, and who would participate in them, and what were their interactions with other organizations. And how did they participate in economic activities (how did they earn their income, their cost of livings, how easy for them to travel, and to buy materials, such as a set of Go stones, gobans, magazines, and newspapers, etc.)

My grand journal of trying to dig up the origin of kyu rank, now became a journey of understanding how Go players lived, learned, paid, and played in the late 19th to the early 20th century.


The first clue about how amateur players interacted with local communities is an account from a poem/writer/composer who lived in this time period - Omachi Keigetsu (大町桂月 1869~1925). He published a book called - 筆 (pen, or to write) in 1908 (明治41年), and there is a paragraph - 囲碁の感 “my thoughts/feelings about Go”. It described how he played handicap games in a Go instructor’s house (a mixture of Go cafe/club/school) near where he lived.

This is what he described - recently, many new buildings opened for business near the village where he lived (somewhere near and just outside of Tokyo). Of all the roadside billboards, many are doctor’s billboard for clinics(醫師), hostels/drinking bars(酒屋), some are tea houses(茶の湯指南の家), flower/internal decoration shops(生花指南の家), and singing/performing theaters (謠曲の家). But he stopped in front of a new Go teaching house billboard. And then he described himself as a player who often played bad or very bad moves(拙手 大拙手), and wasn’t enthusiastic about Go if he couldn’t find others to play with. He was a very casual amateur player and had no hope of reaching shodan (初段 1-dan) or played with people of such strength before.

The owner of the Go teaching house he found was an old man over 50 years old, and said to had the strength of a shodan player. He was welcomed into the establishment. At first, the old man gave him 9 handicaps, and the author won. The next day, they removed 2 handicap stones, and played 7-handicaps games, and then 6-handicaps games, the author still won them. Afterward, he advanced to 5-handicaps games, and he couldn’t win easily. They played 6 to 7 games with 5 handicaps before the old man promoted him to 4 handicaps.

This account is interesting because the player lists in the previous reply I dug up (dated further back around the 1880s to 1890s) had these descriptions :

They basically said that in order to organize (local) Go conventions/competitions, they listed players who had been judged to have the strength of being able to play 2 to 3 handicaps against a player above shodan strength, where some of them marked with circles and triangles were of strength of 3 to 4 handicaps. The implication of these lists, combined with the description from the author Omachi Keigetsu above, is that the compiling of local amateur players lists must have been a bottom up gradual process. Players who were members of the four Great Houses, or the more modern Hoensha, would open local Go teaching houses in places where they lived or travelled to. They would then meet and recruit amateur players in the area as their “customers” as their way of earning their livings by teaching (not that different from many high rank amateurs and pros today). These membership lists would form the basis of the higher up digest lists of strong amateur players in the large area (like a prefecture or city). Ant then, someone finally collected many regional digest lists to compile and boil them down to a giant 5000 people list supposedly covered all of Japan (possibly for the purpose of selling memberships to a Go club covering all Japan, or a mailing list for newspapers/magazines).

How strong players discovered and played with local amateurs using incremental handicaps also corresponds to how dan level players at the time viewed amateurs in their writings. Below is a commentary for a two amateur players game from a book of a collection of real amateur games over the decade, published in 1911 by Nozawa Chikucho (野沢竹朝 1881 ~ 1931), at the time a 4-dan player.

He marked the strength of amateur player such as “準初段格” - a player almost at the level of shodan (probably a stone weaker and doesn’t have a diploma), or at some particular handicaps against certain dan-level strength, like 五段五子 - a player who can play 5 handicaps against a 5-dan player.

The most interesting thing is that these two players above can be found in the giant 5000 player list from 1908.

Where the quasi-shodan player is listed under the “4 kyu” section along an annotation said shodan, and the player of 5 handicaps against 5-dan is listed in the section of “8 kyu”. These two players were from a northern shoreline city (金澤市) in the North-East part of Japan, a relatively isolated city from major Go communities like Tokyo or Osaka.

The kyu ranks in the giant 5000 list are most likely not a good indication of local players’ strength, where they seemed to have their own rank structures and relative strength with each other. They probably also didn’t interact with players from other major hubs often if they were living in remote locations. This wasn’t a single incremental ranking system as we know today (from 9k to 1k and then dan ranks), but an attempt of aggregating various parallel and overlapping ranking systems from different parts of Japan.

A direct evidence of using several rank systems in parallel can be seen in a local Go club list from 1915 (大正四年)

The annotations on the list not only use the Hoensha rank such as 2-kyu shodan (二級初段), but also alleged strength like 二段格 (similar strength to 2-dan, but doesn’t have a diploma), and 12 kyu (十二級) from the old Hoensha rank system (equal to 3-kyu shodan 三級初段 of the new Hoensha system) where their names can be found on the Hoensha member list as far back as 1884 when they still used the old ranking system, corresponding to the first rapid expansion period of Hoensha from 1883 to 1884 to include players in wider areas, that more than doubled their members from 91 to 188 in one year.

I think due to the success of Hoensha in the late 19th century with their newspapers and magazines sold all over Japan as well as mail-Go, the old Hoensha kyu ranking system spread rapidly to various part of local communities in the 1880s to 1890s and became a popular ranking system for many local amateurs who started to play Go around that time. And they would report their local players using the old Hoensha system, where 4 kyu or 5 kyu Hoensha ranks represent the typical strong players, and then used them as measuring standard players for others (or if there were local known dan-level players, they would be the anchors, and most likely judged and organized local leagues)

However, when Hoensha changed back to a mixed dan and multiple kyu-shodan system, not everyone switched with them. Many still stuck with the kyu ranking or already made up their own local ranking system, like the South-West region used a 等 (pronounced “toe”) rank system, quite similar to the old Hoensha kyu rank, but only from 1等 to 9等. The result is that in the early 20th century, there were various ranking systems in different local areas, making the comparison and compiling of a digest list across Japan rather difficult. This is not that different from online Go servers today, where each has its own ranking system. If players don’t play in multiple servers and played with each other, it would be hard to sort everyone out.


With the knowledge of the existence of self-organizing local communities in the late 19th century, we can start to answer the first question I proposed:

  1. What was the population of Go players in the late 19th century to early 20th century and how did they change over the decades

A first estimation can be done using the ratio comparing known members in regional lists with the giant 5000 players list. By cross-reference the names in them, I estimate the ratio is about 10 to 30 times (hundreds of local members with only dozens in the giant list). Using the ratio of 20, we reach a rough estimate of at least 100,000 people had their names written somewhere in a local member list in the late 19th century (and likely participate in local competition games enough to demonstrate their strength). But more casual players like the poem Omachi Keigetsu probably wasn’t strong enough to be on a list (those around the level of current DDK to low SDK ranks), so the total number of players would probably be at least 2 times or 3 times as many, around 200,000 to 300,000 minimum.

Secondly, we can try to find sources at that time already did some estimation. Below is a paragraph from a very interesting but controversial proposal of a scholar/philosopher Inoue Enryo (井上円了 1858 ~ 1919, a very intriguing character) from a book - 甫水論集 published in 1902 (甫水 is the author’s style nickname he gave himself, 論集 means a collection of his essays)

The proposal was to call for the abolishment of Go games in Japan (囲棋廃止論), due to it was draining precious working hours of the citizens from the author’s point of view (he also called for abolishing tobacco, alcohol, fine clothes, backyards, luxury houses/items, and all recreational activities, where everyone should live a spartan monk-like life). But in order to establish his argument, he has to convince readers that it was taking enough time to begin with. And Inoue Enryo was very cleaver about how to achieve his proposal, and well-aware it won’t be effective just to ban an activity out-right. He proposed the use of tax to reduce the “consumption” (similar to how tobacco and alcohol tax work), and worked out the estimate annual flat property tax from Go games, based on his estimated number of Go boards all across Japan - around 500,000 to 1 million yen annual tax income with 2.5 to 5 yen per board from roughly 200,000 Go board sets (and his number was based on the number of the lowest administrative district units with 2 to 3 Go board sets per district unit). This is a very interesting number since it is about the same magnitude with the previous estimation of players, who had their names written in some membership lists.

However, the number of hundreds of thousands is actually still a pretty low ratio if we compare it to the total population of Japan about 45 million at 1900 (growing from about 35 millions to 55 millions from 1870s to 1920s) of just around 0.5%, or about one out of 200. So does this means Go was still just a fancy pass time hobby for the top 1% even in the 1900s?

In order to answer that, next we need to dive further into what kinds of people own a Go board and ask a very fundamental question - How expansive of a hobby Go was at the time? How much did it cost to own a Go board set? How much did it cost to learn to play Go? And where did people play their games?