I’m also curious how OGS determines automatic 9x9 handicap (and 13x13 handicap). I have asked a couple of times before, but I never got a response. Does anybody know where to find this information?
3 Likes
A different perspective: I’ve played less than 10 games, total, yet the site currently ranks me as 16k at time of writing. I think you’ll agree with me that this, genuinely, truly, cannot be correct. I know nothing about shapes and have only the most tenuous grasp on topics like seki. I regularly miscount liberties. There’s no way I should be 16k, and yet…
How is this right? I appreciate wanting to restructure the ranks to be more accurate to some standard, but promoting new players so highly is clearly wrong.
3 Likes
It is suspicious… one does wonder if the Japanese could be so wrong about the existence of 30k
System gives you ± 2.7 which means that its not sure about your rank. Play more games and if you are correct, your rank will become lower.
You won only 3 games - against 17k, 15k(bugged bot), 20k(timeout) - it surely not enough data for any system.
5 Likes
16k now is equiv. to 23k before so you’re probably not as far off as you think.
2 Likes
You’re saying someone who hasn’t played 10 games is 23k under the old system? I’m sorry, I find that preposterous.
6-10 games is roughly how long it took for the system to give you your first rank, it will obviously increase in accuracy over time, so what’s so preposterous about it?
1 Like
Keep in mind, 30k (colloquially, not necessarily OGS) means that you’ve just learned the rules, maybe not even that. It doesn’t take much to improve a few stones. And you’re actually winning against people so it means you know something about the game.
2 Likes
Try this: play a 19x19 against someone (family member, girlfriend, boyfriend whatever) who has never played before and give them 9 stones. Is it a close game? Are you brutally slaughtered? Do you brutally slaughter them?
If you get brutally slaughtered every time you try this with a completely new person, then maybe the system is a little skewed. Otherwise it’s just speculation.
4 Likes
LOL now we are having people saying the old system ranked preposterously high 
FWIW, 25-23k was definitely seen as “the ranks likely to play random moves” under the old system.
Also FWIW, as has been mentioned before, even in this thread, there are known problems being investigated with the behaviour of initial ranks.
5 Likes
IGS, Tygem, Fox, CyberOro all have a minimum rank of 18kyu and I think KGS absolute beginner was not far from 20kyu.
OGS was always the outlier.
3 Likes
…better 
7 Likes
I want to say that, from the oldest kyu rank documents I could find around the 1930s, the lowest kyu rank is 15k. And at the time, the rank determination is very rigid and determined by handicaps (手合い)
Some Go schools like the one I attended also still only have the lowest kyu rank class at 11k~15k, anyone who just joined would be assigned to the beginner’s class and don’t get any rank.
I wonder when did the kyu rank start to “inflate”?
8 Likes
This is fascinating, because it seems to say that OGS’s 25ks were way too low, and now that they have come up to mid-teens, it’s about restored to what it once was! 
1 Like
This comes from 鈴木為次郎『囲碁独習』日本棋院 1931 (“Self-studying Go” by Tamejiro Suzuki for the newly established/merged Nihon kiin at 1924)
The reason that amateur (素人) rank only goes down to 15k seems to be that the beginner(初心者) was given 25 handicaps(15k), with some big gaps - 20 handicaps(14k), 17 handicaps(13k), 13 handicaps(12k), till 11 handicaps(11k), and then jumped to 9 to 2 handicaps(10k~3k), and josen and tagaisen (2k~1k), where komi was yet to be introduced.
And dan players and Nihon kiin players had their own smaller handicaps 手合割, all dan players were considered “professionals” (專門家)
They seemed to already recognize that lower-ranked players would have trouble utilizing handicap stones.
13 Likes
16k is not a high rank.
Until you have a bit more experience with the game and understand seki, I think you’ll be happier if you just ignore the rank, play some more games and have fun. 
You don’t need a rank if you aren’t invested in the competitive aspect.
(Hint: you can even hide rank display in the settings if it distracts you.)
7 Likes
I have encountered many players that are clearly weaker than 15k according to handicap.
For example:
- Back in the day when I was about 2k I’ve played in some full handicap tournaments. In some games I had to give 9 handicap + something like 150 reverse komi (all adults). I even won those games by 100+ points (them ending up with only 1 or 2 small living groups). Such a result would suggest a rank gap of about 27, meaning those players were about 30k. These were probably beginners who had little more training than a beginners course.
- I consider myself 3d and I routinely give 13 stones on 13x13 or 6 stones on 9x9 to the intermediate players in my childrens club. These are not raw novices. They know some tactics, they know about life and death and seki and they can score their games without my help. I think that these handicaps correspond to a rank gap of about 30.
So I’m unconvinced when some go school or go association deny that players weaker than 15k exist, because I know from personal experience that beginners and novices are clearly (much) weaker than 15k when measured by handicap. And I cannot believe it was really different in the 1930s. Maybe they didn’t consider players weaker than 15k “real” go players back then? Or is handicap measurement an invalid method to measure ranks? If it is, what do ranks even mean?
My explanation is that in the 1930 they only just started to apply science to the traditional rules (such as komi, time control, adjourning etc). It took many decades to get things “right”. (And some things are still not “right”, such as the official formulation of Japanese rules and Chinese rules, professional promotion rules, etc.) I guess that tradition runs deep in Japan and China.
8 Likes
Nobody denies that weaker players exist just that they weren’t assigned ranks. Claire mentioned that they are just regarded as beginners.
The reasons that they used don’t seem to be very good but that doesn’t prove that a good ranking system requires 20k+ ranks.
3 Likes
In my experience it’s perfectly possible to determine ranks weaker than 20k by the handicap they need to beat a 20k. So I see no reason to put those TKP playes in one 20k+ purgatory (as @anoek called it in his inititial post in this topic).
5 Likes
I’ve replied here
The lower rank were given more than 1 stone per rank. It’s just how they devise the rank difference back then. As to modern Go schools, we don’t actually use the handicaps as measurement anymore, but assigning by teachers to different classes along with some tournaments results, and usually play normal komi games amount classmates in the same class.
4 Likes