Server comparison (ranks)

I found this table online:

I don’t quite get it, is this really right? It seems to say an 8 kyu on OGS is almost a 1 kyu in Japan and China. Note that those 2 countries seem to be the outliers. I agree that KGS and Fox seem to be the easiest servers for a high rank.
Any comments?

1 Like

this survey is from 2018
while OGS ranks were changed in 2021


This Sensei’s Library article also discussed the survey and another earlier one from 2015: Rank - worldwide comparison at Sensei's Library

I guess it’s time for an updated survey to see how the 2021 rank system adjustments have turned out.


In OGS profile its possible to show what is your rank on other servers and organizations. Many users used this feature. Instead of survey, its possible to just collect that data from active OGS users.


ok, this table seems to be very similar to the one i posted. i am just wondering if China/Japan get high rankings really that easily, or is the sample size to compare OGS vs those countries not sufficient? i mean, i’d love to be a 1 kyu in China if that’s true

or are OGS players just that good?

Its possible to measure strength directly from games
it may be interesting to draw diagram like below, but by using games from different servers. Then it would be clear which 1k is really stronger

(mistake means move with at least 1 point loss compared to move of AI)

Definitely the latter is what I’d say. I learned and played Go in Japan for several years (2011 to 2014) and have played live with folks many times thereafter when staying for long periods. The gap isn’t as pronounced as what people English in forum folks believe.

Most players I’ve met in Japan refer to different ranks, e.g. their local club rank, their IGS rank. There’s also official ranks distributed by the Kiins, but I haven’t heard much about those (i.e. people don’t mention whether they have a certificate for rank X).

(Where? When?)

Not me I hope at least.,

Sorry. Not meaning to point fingers at anyone. I am just referring to the above reference of “8k OGS = 1K Japan” and how I have seen several folks saying ranks are “inflated” here (I’m currently in Japan), but without much backing.

I’m a bit defensive on this point given my first-hand experience and close ties to the country. :laughing:

In 2017, a few U12 French players were invited for a go competition in Japan. An FFG 13k registered as 7k, and an FFG 5k registered as 2d.
In 2018, three French teenagers were invited again in Japan. Two of them, who were FFG 5k and 6k, didn’t dare to register as dan players, so declared 2k and 3k rank. As far as I remember, one of them won the kyu tournament after beating the other at the final.


I think even in Europe French strong kyu players have a reputation for being underranked by some 1-2 ranks, compared to the average in Europe.


What were their KGS or OGS ranks if they had them? As someone that has never played Go in Europe, those ranks mean little to me.

I’ve also won a kyu tournament in Japan. At the time, I believe I was KGS 3K. 3K to 1K is only a two rank difference.

Regardless of the differences, I don’t think it warrants blanket statements saying “ranks are inflated in Japan”. That promotes the wrong image, and frankly it’s belittling.

Merriam Webster defines “inflated” as:

expanded to an abnormal or unjustifiable volume or level (e.g. inflated prices)

I don’t think the ranks in Japan are unjustified.

It’s a different country, with different standards / customs. There are also a lot more clubs than in North America and I presume Europe. Some are small, some are really big.

The Go population is also elderly, and unfortunately people do get weaker with age.

The last club I played at didn’t use ranks, but a number system for pairing people. We didn’t discuss ranks.

I think FFG 6k is about OGS 5k. So my guess is that OGS 5k = Japanese 1d. This estimate is not 100% precise of course, people with the same rank IRL may have online ranks that differ by up to 4 stones.

1 Like

I don’t think that’s what “inflated” and “deflated” mean in the context of rating systems. The terms just refer to the relative strength corresponding to nominally analogous ranks in separate ranking systems; no value judgement implied. Generally deflated ranks are considered almost as bad as inflated ranks.


That’s good to know. For the layman, it sounds judgmental.

I’d encourage folks that use such terminology to add precision and nuance, given that no international standard exists afaik for Go rankings.

e.g Japan’s go club rankings are generally inflated compared to OGS rankings, typically by a few levels.

Note that I didn’t say “Japan ranking” since that doesn’t exist. There are rankings within clubs, general rank ranges used in tournaments, and rank certificates handed out by the various Kiin (Japan Kiin and Kansai Kiin… not sure if there are others).

I don’t mean to be pedantic, but a bit of care with wording goes a long way. :smile:

1 Like

When people participate in tournaments, do they register with their club rank or something else?

1 Like

Indeed, this was my experience at the London open in the early-mid 2010s, in the low dans too. French ranks were strong, British ranks were weak (and Japanese much weaker). British ranks are now aligned with EGF ratings, to the chagrin of some long-time older players who have been demoted, we don’t have a separate rating system like France does.


Feel free to if you wish, but I can’t see inflated and deflated as sufficiently ambiguous in this context to warrant such prophylactic disambiguation (of course, clarifying when it’s misunderstood is a given, but that goes for any word). Words I would prophylactically clarify my use of would be mostly archaic or technical uses, like using kill in the archaic sense of murder, or using vertex in the technical sense of graph theory.


I think it’s a useful linguistic shortcut to group a variety of things of an associated set (in this case a country) together in order to talk about properties which generally hold for those things in that set, without going through the busywork of rigorously defining that you are claiming that the property in question applies to the members of the set as individuals (not corporately) at a rate exceeding the norm