Ranking and Time Zone

General Go Discussion
21

As far as I know, historically the rank benchmark in Europe has been that the highest amateur level is 6d, while a few top amateurs of similar level as top insei or low ranking pros were considered 7d (as sort of an honorary rank). From there, other ranks are more or less determined by handicap.
I think this benchmark originally came from Japan, but Japanese ranks have since inflated somehow, so that in the early 90s some Japanese 6d were like 3d in Europe, while some other Japanese 6d were like 6d in Europe. When I was in Japan in 1990, I played as a 4d, while my Dutch rank would have been 1k-1d.

There are 2 Dutch players who are/were considered 7d in Europe: Ronald Schlemper and Rob van Zeijst, who have both been top insei in Japan and both of them were European champion 3 times.
Ronald Schlemper also took 3rd place in the WAGC in 1988, behind Japan and China.

In the late 80s and early 90s ranks in the Netherlands were about 3 ranks tougher than ranks in Japan (at least in the dan range).

At that time, there werenā€™t many Chinese amateurs playing in Europe, but the few there were did seem to have ranks that were similar as in the Netherlands, perhaps even a bit tougher (by about half a rank).

I donā€™t know much about lower ranked Korean amateurs at that time. In my memory Korean amateurs playing in the European Go Congress in the early 90s were all 7d and typically only lost to each other. The strongest Korean amateurs who are active in Europe today are ranked 8d EGF, overlapping with low-mid rank pros. Their level may be similar to the level of Baduk Doctor who is a Korean 7d.
From this it seems that at least Korean high dans are about 1 rank tougher than European high dans, and this hasnā€™t changed much over the last 3 decades.

I think the introduction of the European rating system around 2000 caused some deflation in Europe (at least in ratings, ranks to a lesser degree), but it was only on the order of a 0.5 rank deflation in the dan range over a period of 20 years.
In 2019-2021 I was involved in updating the European rating system, repairing this deflation by (among other things) inflating EGF ratings by about half a rank on average (in the low-mid dan range).

In the early 90s, I think AGA ranks were considered about 2 ranks softer than Dutch ranks, so AGA ranks would be more or less in between Dutch ranks and Japanese ranks. But I think that increased (online) contact between European and American players on KGS in the early 2000 reduced this rank gap. Nowadays the gap between AGA ranks and EGF ranks seems to be only about 1 rank around 1d level, though possibly still 2 ranks at mid-high dan level. A possible explanation for the latter is that KGS high dans were inflated compared to EGF high dan levels. In the 2000s it was not uncommon for 6d EGF to have a 9d+ KGS rank.

Iā€™m not aware of the rank gap between Japan and Europe changing significantly since the 90s, so Iā€™d assume that Japanese ranks are still about 3 ranks softer than European ranks (at least in the 1d-5d range).

In China a lot seems to have changed in recent years. Chinese dan ranks below 6d seem to be all over the place nowadays, especially for younger players.
My impression is that nowadays stronger Chinese 5d are similar to 5d EGF (perhaps even 6d EGF), while weaker Chinese 5d are more similar to 1d EGF.
Currently there is a high dan player from China active in the Netherlands. Given his rank history in the EGD, I assume he was 5d in China, but he is more like 6d in the Netherlands: https://www.europeangodatabase.eu/EGD/Player_Card.php?&key=19498325

As for Chinese kyu ranks, I know some cases of kids in Europe who would be about 15k here (as in needing about 9 handicap against a 6k EGF), who had their ranks estimated by their online Chinese teacher, and the estimate of the Chinese teacher was more in the range of 10k, possibly as high as 5k.

I also know of 2 Taiwanese amateurs playing in the Netherlands. From them Iā€™d think that Taiwanese dan ranks are about 2 ranks softer than Dutch dan ranks. One was 6d in Taiwan and he is 4d in the Netherlands, the other was 1d in Taiwan and she is 2k in the Netherlands.

5 Likes
22

Interesting pages, but i wonder why Schlemper musical performance isnā€™t mentioned at all (it was usual to hear that he had professional level in that too!)

3 Likes
23

Super interesting. Thanks for sharing some of the history behind the ratings system @gennan .

Iā€™m surprised to hear about the 6- or 7-dan cap. Somehow I expected the max to be like online servers (9 dan).

1 Like
24

Hum not always. IMHO strength was not always so easy to import online. Fact was some were 6 some 7 too. That was the point where i started to watch games. 8d was really strong crushing the 7d (and 7d the 6dā€¦) but still could be crushed by a 9d usually of the pro class (like milan milan, daoceā€¦ )

1 Like
25

I think the AGA also has amateurs ranked 9d+, as does KGS.
I suppose that servers attract more users by awarding inflated ranks. Who doesnā€™t want to be higher ranked, even if only informally?

26

I think you have a point that the EGF and AGF ranks are too restrictive.

Historically, the reason why they at all introduced amateur ranks in Japan was to help popularize Go after the old professional system lost state support through the fall of the shogunate. The whole reason for having amateur ranks at all is to popularize Go and it is sensible that amateur shodan should not be an unreachable goal for the majority of players, but merely indicate that one have a comprehensive understanding of strategy and tactics.

The reason why the Korean system is strict is probably that they got amateur dan ranks quite late (as can also be seen in the novel First Kyu) so in the Old Korean system 1 gup/kyu merely meant that you were a strong amateur player who had not won professional status in the professional examination tournament.

The reason why American and European ranks are deflated is probably that there was originally no professional ranks in Europe and the US, but now that there is also a professional system in Europe and America, it would be good if the amateur ranks could become more like the Japanese and Chinese in order to help motivate people and popularize Go.

2 Likes
27

Iā€™ll share a few stats so that one can judge whether Western ranks are too strict or not.

  • Among people who have a FFG rating and appear on the website, 14% are 1d or stronger, and the median rating is -600 (i.e. 6-7k).
  • In the Orsay tournament, the median rating of participants was -269 (3k).
  • In that tournament, among participants whose first rated tournament was 10 years ago or before, 39% were 1d or stronger, and the median rating was -151 (2k).
  • In that tournament, among participants whose first rated tournament was 5 years ago or before, 30% were 1d or stronger, and the median rating was -178 (2k).
  • The French Karate Federation has 13.5% black belts.

Notes:

  • I made those statistics by hand so may have made mistakes.
  • Not every club player has an FFG rating.
  • Reaching 1d is a wrong motivation. Low dan players on Fox or Tygem sometimes miss simple life and death problems or simple ladders, so I consider dan ranks there to be worthless.
5 Likes
28

Iā€™d day that todayā€™s Chinese ranks and Japanese ranks are not very much alike, nor are they very consistent internally.
As I said before, Japanese 6d can be anything between 6d EGF and 3d EGF and Chinese 5d can be anything between 6d EGF and 1d EGF.
Do you want ā€œuniversalā€ ranks to be on the softer side of Japanese and Chinese ranks? So ā€œuniversalā€ 3d would be like 3d Fox ~ 1k AGA ~2k EGF?

29

Klassen Go bond jaren 60
Klassen Go bond jaren 601723Ɨ1219 351 KB

Well, just for fun I will share a picture I took from a couple of pages in the magazine issued by the Dutch Go Federation (Tijdschrijft van de Nederlandse Go-Bond). Itā€™s from their founding year, Nr.2 (March 1961). All ratings are compared to the strenght of the world champion, Honinbo Takagawa Shukaku at the time, ranked 1. A novice would start at around 90 and was likely to improve to 70 soon. The European Champion at the time, Ir. Ciessow (from Berlin, Germany) was ranked at 16. The text mentions a formula which can be used to determine rank and / or handicap. There is no reference to the terms kyu or dan.
It also mentions that in an ideal situation, but unfortunately impossible, strong players from Europe should play the world champion regularly to keep ratings balanced.

I tried to get the text translated into English using some software (which can be done) but I lack the skills :grimacing: :joy:

Anyway, ratings seem to be subject of debate for a long time, and will be :wink:

2 Likes
30

Interesting stats, but I doubt reflective or useful without much more work.

A counter stat, for example, is that Iā€™m in the top 8% in GoQuest currently (~1800 where my ranking is 3095 out of 33159), but am under 1 dan in OGS.

Iā€™m not saying the GoQuest distribution is something we should follow, just pointing out that weā€™d need rankings, including historical values with active status, across many platforms and associations to tease objective insights.

Numbers aside, I feel that a bit less skew in the upper range would help the community.

Reaching 1d is a wrong motivation.

There is no such thing as a ā€œwrong motivationā€. Folks can definitely be motivated by ranking up, and there is nothing wrong with that.

31

1800 and under 1 dan thatā€™s consistent, no surprise here.
Still a lot of space to improve (like to 2750)

32

I donā€™t think it was the intention, but that sounds slightly condescending. :stuck_out_tongue_winking_eye: I of course already know that I still have much to learn. :muscle:

To clarify, my point isnā€™t about room for improvement, but that the FFG and GoQuest distributions are completely different distributions.

To use examples that donā€™t leverage my personal rank, I looked up the ranking for a 1900 individual on GoQuest (~ 1 dan). Their rank is 1848 out of 33159, so they are in the top 5.6%. This is very different from the FFG, where the top 1 dan are 14% of the population.

If we were to do the same exercise with other online servers / associations, weā€™d get different distributions as well.

Haha. Very true.

Still, this has been a super informative discussion.

Thereā€™s still lots to think about, but my current conclusion is that tournament strength and kyu/dan level should be independent of one another.

Again, this is similar to martial arts: in karate you donā€™t go down ranks.

A kyu/dan level could be something to strive for, but wouldnā€™t necessarily represent your tournament strengthā€¦ :thinking:

33

Sorry if i gave that feeling, not my intent.
It wasnā€™t about how you have still a lot of margin to progress but instead about how many players could still fit in this margin and modify the statistics

Those distribution may vary not by the choices of the system but by the population involved. If much more intermediate players come play on Goquest than in the FFG, then the 5.6% would not be a surprise.

Strong statement, all a debate.
Myself i like your level being determined by tournaments results and a dichotomy between these and a training time.
I did some martial art too, and it was great that after some weeks of activity in the club, we broke the routine, go into the unknown and check how our practice will or will not be efficient.
Now this is something fading online because we just rate all (and at the same level)

2 Likes
34

Go servers like GoQuest attract more casual players than go federations. One reason is that people who subscribe to a go federation pay a subscription fee and regularly play in a club or in tournaments (so have to move physically), and thatā€™s already a first effort. On the other hand, you can play every week in a club or participate in 10 tournaments per year without studying hard, and you can study hard and still keep bad habits. So Iā€™d say that the current FFG ranks are ā€œabout rightā€. Not necessarily the best in the world, if they were deflated by 1 or 2 ranks and thus align with AGF, IGS or KGS, I wouldnā€™t find that unreasonable. But I hope theyā€™ll never align with Fox or Tygem, that would feel very wrong to me.

Ranking up, sure, you can be motivated to rank up from 16k to 15k. But Iā€™d be against too much devaluation of dan ranks, even if it means I may never reach 1d FFG/EGF. A dan rank is like a driverā€™s license: you give it to someone you can trust to respect ā€œdriving fundamentalsā€ and have a low accident rate. In the same way Iā€™d expect a dan player to play very few moves that donā€™t respect go fundamentals.

3 Likes
35

I guess that is an inheritance from the old Japanese system when all the Go houses agreed on a Meijin who was officially the best player.

If I understand the old Korean system where there are originally no amateur dans it is more suited for Go clubs since it has not the best national player but the complete beginner as its starting point with 18 gup/kyu and then you work your way down to 1 gup in your Go club.

How did the old Chinese ranking system work before the revolution?

36

Do these numbers correspond to half ranks? So that the European champion was 7.5 dan weaker than the world champion?

1 Like
37

The text describes the class system that is basically an alternative ranking system to dans and kyus (but there is a clear correspondence between these systems).

Classes are half-ranks, in the sense that they are spaced by komi = half a handicap stone, so classes are more fine-grained than ranks.
The lower the class, the higher the level. Class 0 or 1 (= 11-10d EGF) may be the level of rare geniuses like Go Seigen or Shin Jinseo. The text mentions that there are indications that the level of the Honinbo title holder at that time might actually be class 3 or 4 (= 9-8d EGF). My 4 dan diploma from 2005 mentions class 13.
It seems like that text is from 1960 or 1961, because it mentions GĆ¼nter Ciessow being European Champion, who was class 16 at the time (= ā€œstrong 2dā€) according to the text. The text also mentions that raw novices are about class 90 (~35k), but they usually improve fairly quickly to class 70 (~25k).

The text also describes that the class system allows more accurate calculation of proper handicap, noting that 2 handicap stones is equivalent to 3 times komi and is thus proper for 3 classes difference between the players.

The page on senseis says the class system was abandoned in 2008, but I still use it in my childrens club and itā€™s also used in the Tilburg go club (and I assume there are more Dutch go clubs that still use it for some internal competitions).

5 Likes
38

The original intention of a ranking system is to provide a way to play anyone with a fair chance of winning, wether it is just by numbers between 1 and whatever or represented in a rather complex kyu and dan degree system. Itā€™s of course human nature to compete and improve but like in life, we grow, peak and decline. That seems a hard reality to accept. I think Go is the ultimate game to embrace that reality.

4 Likes
39

As a reverse komi fan, I absolutely love that!

1 Like
40

What else would a kyu/dan rank represent, other than strength in competitions (be it current or more a life-time achievement title)?
And how would you measure it? Would it be like XP points that many computer games have, giving awards for achievements like number of games played, number of years played, number of tsumego solved?