While I was researching the early amateur Go community in the early 20th century in Taiwan, I found this very interesting newspaper Go game (played in April 1937) has a very "intriguing komi - black gives 6.5 komi (六目半コミ出し)
However, another game right before this one, just a week earlier, instead used a much smaller komi - black gives 2.5 komi (二目半コミ出し)
You can also check the final scores of these two games (both games went into scoring), and they indeed follow the komi listed in the newspaper. They were not typos or misprints. (these two games even have the same player in them - 羽廣清太郎, which actually gives us a clue and we will talk about what these matches were later)
Were they just experimenting and trying out different komi? Here is another game played the week after, and it was again using the larger komi - black gives 6.5 komi (六目半コミ出し)
So what is going on? Can you make an educated guess as to why they switched around komi in a series of games decades before large komi was in use? Did large komi arrive decades earlier than we previously thought?
Let’s jump into the rabbit hole of Go history and find out! (the initial post is already too lengthy, so I will break it up and continue the story next)
My guess is that Komi was adjusted from game to game, as a form of rank. Every time you lose you get 1 Komi more in your next game, while the winner gets 1 Komi less.
The Komi of the game is the sum of both players “handicap Komi”, which turned out to be 6.5 in the 2nd game.
On top of that, white gets 0.5 to avoid draw.
The person with the lowest Komi (or actually most negative Komi) in the country is the reigning champion.
You are talking about the Points Rating System (点数制の手合割) that tracks “rating points” to adjust komi for players of different strength (with a ratings/ranks conversion table). According to Yasuro Kikuchi (菊池康郎), he said after WW2, the amateur study groups had been using it where at the time people believe the difference is 5 komi for one rank with equivalent komi at 4.5. And he modified it in the 1980s when he established Ryokusei Go School, which he adapted from methods used in Ryokusei club (緑星会, established in the 1960s, which he was a member of), where the initial point system only use 5 to reverse 5 komi (when reaching 12 points difference, it jumps to the next handicap stone).
The original Oteai (大手合) for the Nihonkiin to rank their professional players also uses a point system. But instead of giving a point to “rating”, each game is evaluated with a “performance point”. Win or lose, or draw would give different performance points, and at the end of the year, the performance points are average to see how well each player performed. Those who performed above average and passed a certain threshold would get promoted to the next rank, and change their teai (手合 matching handicap) and the performance points would reset each year. Ninhonkiin used to publish these performance points and rank change annually (the oldest I can find publically is from 1932 records, but from magazines published by Ninhonkiin, they had been using it since its establishment in 1924). It could be much older and used before in other clubs or in-house in different associations back then.
It’s not hard to imagine the amateur Go community saw the “Nihon kiin ranking system” and just simplified it to suit their need. However, from Go books in the 1920s and 30s, even after WW2, the teaiwari was still mentioned very often and introduced as a way of differentiating ranks and play between different strength players.
The system I aim to unearth from old records and sources is different from the Point Rating System, and as far as I can tell only started being used sometime after the 1930s, and hasn’t been seen after WW2 records. Whether it is a local variant, or it evolved and inspired other systems, I cannot say at this time, and probably required more research. I’ll describe it more in detail next, but first, we need to take a detour for the background and context of these newspaper Go games before we continue. It certainly did not just appear out of a vacuum.
Very interesting, thanks.
The type of ranking used can greatly influence the challenge level during play, and the reward a player feels on a win. Talking about this seems very modern (ie. the gamification of everything), so it’s very cool to realise people have been thinking of systems way earlier.
Throughout history, Go has been linked with gambling, particularly gambling by comparing captured stones or winning margins. So it stands to reason players had been very sensitive to the margin of winning even if they knew they had won the game. It is not just enough to win but a bonus to win by a larger margin. After thousands of years, it stands to reason players had been keeping scores and have quite a good sense of the margin the first player was expected to get (先手利), and how to minimize it if playing the second. There is a reason why we have records dating at least to the 19th-century imploying komi (played in 1853, 5 points komi, and likely not the first).
However, ranking players using the dan system is a different matter. And for the longest time, it is more related to the guild-like structure of the Great House. Where establishing seniority and competing for authority was the main goal. They were not trying to make it fair but understood players naturally have different strength. Sometimes rank is not even linked to the teaiwari between players. A player can have different teai with different players, and only when one has beaten enough of the stronger players and changed all the teai through a series of games before their “official rank” can be recognized (and sometimes even if they were strong and beaten others, they still didn’t get the promotion, due to politics in the Great houses).
BTW, I’ve been weighing on how much context do I have to bring up. Since literally this was a research in the Taiwanese amateur communities first, and I have way more materials related to it than others. However, other contexts in mainland Japan at the time matter a lot for the subject of this particular komi system, since it originated there in a slightly different form, and it would require some background beliefs I talked a bit here - how they view the teaiwari at the time. It matters a lot why they choose certain komi values. I’ll try to concentrate on just one subject in one replay to make it easier to digest, and less taxing for me drafting. Less relevant subjects like the activities of the amateur communities in Taiwan or in mainland Japan in the early 20th century I can save for another post.
I have been working on the context and how far back we need to take to understand the rationale they got throughout history, and here is one of the digests I replied on reddit
The reason is sound, and I actually find records a bit old talking about ranking difference and teai (and if anyone wants to know the spoiler about what this almost forgotten equivalent komi system is : )
Summary
it is about measuring half a stone, and how they viewed the teaiwari gaps, it matters a lot in terms of why they changed it in 1927, and again merged it with the new komi system. The 2.5 komi is half a stone, 4.5 is 1 stone, and 6.5 is 1.5 stone, that is for 3 ranks difference player if the stronger player played black
Anyhow, my problem this week is that many of the older books are still in the early 20th century, instead of around the date when the teaiwari was established in the 1600s. Not even in later dates like the 1800s. And older publications became harder to find and search since beyond the point of 1850s, more and more old books were handwritten instead of printed, making searches in the database very hard since no OCR can recognize those written books properly and index them.
If this takes too long, I might just make a digest and continue the story in the 1920s and 1930s, and add them as supplementary if I find better records and books in the future to clarify what the ancient players thought and wrote about them before the 19th century.
