Top OGS players (more serious this time)

Done.

N of games is number of games within last 3 months here. Oh, some ranks already changed from yesterday when I downloaded data. Oh, well.

Highest rated active players

Username Rating Dev N of games Registration date
1 peecatchoo 3074 (11d) ±123 27 2016-07-25
2 Sadaharu 2986 (10d) ±127 331 2012-04-24
3 잔세준 2936 (9d) ±128 13 2018-07-30
4 hhht 2891 (9d) ±109 19 2018-06-13
5 YHWQ 2878 (9d) ±103 17 2018-05-18
6 BobSong 2875 (9d) ±103 114 2018-05-14
7 rzhang20877 2872 (9d) ±122 11 2016-04-10
8 vlin 2872 (9d) ±139 13 2014-07-07
9 Tremendous 2867 (8d) ±101 38 2018-04-14
10 Leonix 2840 (8d) ±97 31 2018-04-06
11 ABlob 2839 (8d) ±135 173 2018-06-07
12 BOT Kyu_player, 2799 (8d) ±87 258 2017-03-20
13 5<2 2793 (8d) ±78 330 2017-08-05
14 k12h20m 2787 (8d) ±70 52 2018-04-17
15 KoreaInsei 2769 (7d) ±71 30 2017-02-01
16 BOT minusGo 2754 (7d) ±109 25 2017-10-30
17 80d8f16310b74052 2737 (7d) ±111 23 2018-04-19
18 BJYH 2735 (7d) ±83 163 2018-07-02
19 Winterfell 2733 (7d) ±88 24 2016-07-13
20 DizzyAlpaca 2732 (7d) ±79 312 2018-08-10
2 Likes

Where am I on the list? :stuck_out_tongue:
And feel free to put me on the list as bot if you like.

Of the active top 10, peecatchoo, Sadaharu, rzhang, vlin, and Leonix are the ones that aren’t playing almost entirely bots. They’re probably the actual top players on OGS, right now…

1 Like

Sadaharu is an interesting case. All his individual ranks are MUCH lower than the main ranking. Not sure what’s going on there, but apparently there is some weird discrepancy. I understand that this can happen for players on the rise, but for a stable top players I’d hope the ranks would convertge eventually.

This is true for almost all players (although maybe not to the same degree). The individual ranks behave somewhat funky. For instance, Ive often observed an individual rank go DOWN after I win against a weaker player.

I think highest rating one can get depends on number of players playing in that pool. If there’re only few players playing, their ranks can get only so far apart. So for certain high rating to be achievable, you need enough players to “support” it. Because overall rating has the largest pool of players, ratings can get higher than in categories. Next category would be 19x19 live, I assume.

I imagine it works a little bit like this. Compare these two fictitious and simplified rating periods:
[2,0,0,1,1,1] vs [0,0,1,1,1,1]. You will notice it’s 5vs4, a rating drop. Despite having one more win on your record, it’s an unsurprising win and therefore worth fewer points than the spectacular upset that gained you 2.

As for overall/overall ratings, they seem to be just about 100-200 points higher than the (most reliable) live 19x19 rating for the vast majority of people. For Sadaharu, that gap is 423 points though (would be 5d for live 19x). He doesn’t play enough live games.

I apologize for my ignorance. I’m trying to research things on the forum before asking them again, but I have a few questions after seeing this thread. Feel free to just link to threads with the answers instead of typing it all out (assuming you know where the thread is) (@S_Alexander or anyone else who could help)

  1. How is it somebody can create a fake 9d+ rating?
  2. How do you you get access to the OGS data? Is this something publicly available?
  3. How does one go about making a chart

Thank you greatly!

1 Like

The trick is to win (all) games :wink:

To be more precise, the rank of bots is to hight. If you can win easily against bots, you play against them. If you are not that strong, you have to know how to trick a bot to play the same bad moves in every game.



OGS has a publicly accessible api
https://ogs.docs.apiary.io/#
Search the forum for more information and help


This is a complicated topic. It depends on your experience, what kind of data you have and what you want to visualize.

I would recommend to search the web for a tutorial and software suited to you.

3 Likes

If we’re counting correspondence players, then what about Haylee (Hajin Lee)? She’s 4p and she always has a (teaching) correspondence game going with maufred. I’d call her an active player, she plays her moves regularly.

After Haylee, it’s hard to say who is strongest. My guess would be vlin (Victor Lin). Then you’ve got a lot of players who are probably at around 5d EGF, like vitality, Sadaharu, xhu98, TwoCats (Brandon’s half), Mark5000, and KDJ123. I couldn’t place them in order of strength but they’re all pretty far up there.

@bugcat If you read the above posts, you’ll see peecatchoo mentioned (a couple times) as the highest ranked account, above vlin.

And indeed pro players should be counted, even though they don’t have an OGS rank.

1 Like

There is no point in including (retired) professionals’ accounts that are marked as such, because they only have a title (_p) but no rating and do not enter the competitive player pool. The point of the exercise is to say “Here are the n highest rated accounts - they have attained and maintained their spots despite many a player’s best efforts in many rated matches”, as opposed to “This is a list of people who should be able to make top n in terms of rating, based on our best guesses”.

In other words, it makes no sense to ascribe to someone a place in the distribution of a population they are not a member of.

2 Likes

Ah, sorry, I misunderstood the purpose of the thread. I thought we were speculating on who the strongest OGS players were and using their Glicko rating as an aid to that, rather than looking at the rating as an end in itself.

1 Like

Putting pros on the same list with amateurs is an insult to pros though.

1 Like

@bugcat me too. I guess it also depends what you mean by “player”. Clearly Haylee is only here for exhibition games and lessons.

@S_Alexander in Asian countries, certainly. Elsewhere I guess it’s ok, see the European and American Go Congresses for example.

1 Like

Well, Haylee does play here. Sure she only plays one correspondence game at a time, and it’s handicapped, but she doesn’t always win. I don’t see the point of excluding professionals from ranking lists just because of their title.

Really, it seems to me that this thread should have been titled “Comparing the Glicko ratings of active amateur players using OGS for competitive even games.” But that would not be the same list as “the top OGS players.”

I’m surprised Ten hasn’t been on this list… isn’t he 1p or something?

Ten: 2859 (± 146) corresponding to 8.90d, cheekily taking spot #10 on the revised list
Haylee: 1500 (±350)

Ten hasn’t played here in a year, and I haven’t seen him in the chat for about as long. He used to participate in the chat on a semi-regular basis.

As I said, I don’t believe she is playing. These are most likely teaching games, considering the high handicaps. In such cases the teacher might pick a move for its educational purpose, rather than picking the best move.

That’s just the way it traditionally works. It is in particular meant as a mark of respect to older masters, whose current playing abilities might not reflect their true understanding of the game. This might actually apply to Haylee herself as she has retired.

1 Like