It is currently Thu Jun 30, 2022 5:52 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 69 posts ]  Go to page Previous  1, 2, 3, 4
Author Message
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #61 Posted: Wed Jun 03, 2020 7:45 am 
Lives in gote

Posts: 442
Location: Netherlands
Liked others: 243
Was liked: 127
Rank: EGF 3d
Universal go server handle: gennan
Tryss wrote:
"making a living" has less to do with competency, and more with the amount of money that is dedicaced to the discipline.

In some disciplines, the world champion isn't a full time pro, because there's no money, while in other, you have over 60.000 professional players


Very true. The top 5 in tennis and golf make a lot more money than the top 5 in ice skating and steeple chase.

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #62 Posted: Wed Jun 03, 2020 9:01 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3372
dfan wrote:
jumapari wrote:
But there seems to be a gap between professionals and amateurs at Go, which is much wider than in chess.

I think that the actual gap in skill is less than has often been thought, especially in modern times with online play and AI. There has traditionally been a mystical gap between "amateur thinking" and "pro thinking", but these days I think it's pretty well accepted that the line is fuzzy, with the strongest amateurs being clearly of pro strength.


In Japan in the mid-20th century the highest amateur rank (with a few exceptions) was 6 dan. The range of amateur 6 dan strength was wide, and definitely overlapped that of pros. Insei who did not become pros often became 5 dan amateurs.

Since then amateur ranks in Japan have inflated by a couple of stones, at least.

Edit: And as far as making a living at go is concerned, it was said that you needed to be at least 5 dan pro.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #63 Posted: Thu Jun 04, 2020 3:53 am 
Lives in gote

Posts: 559
Liked others: 129
Was liked: 106
Rank: OGS ddk
KGS: Ferran
IGS: Ferran
OGS: Ferran
Bill Spight wrote:
Since then amateur ranks in Japan have inflated by a couple of stones, at least.


Just to make sure I understand it, because I tend to think of stones as strength and rank as "degree": you mean that for the same skill the declared level has increased a couple of steps?

Quote:
Edit: And as far as making a living at go is concerned, it was said that you needed to be at least 5 dan pro.


I think I've read in a couple of places bits and bites that seem to imply that there IS a significant step between 4th and 5th dan. I think I did a chart some time ago with declared ranks of the Nihon Kiin, and there was a valley at 4th or 5th dan. It kinda felt that you were either good enough, and kept rising, or just not good enough, and you simply didn't reach that level. In fact, if you check the Kansai Kiin the rank with the lowest number of members is 5th dan. I mention that one because I can reach it fast[*]. And I think there were some tournaments that were for young AND below... 5th? dan. And this or that bit, that I can't seem to recall now, but the general feeling was indeed that there was a barrier [psychological?] at 5th dan.

Take care.

[*] I've mentioned before that I'm a bit of a Kansai fanboy. Only a bit, but...

_________________
玄 之 玄

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #64 Posted: Thu Jun 04, 2020 4:32 am 
Honinbo

Posts: 10905
Liked others: 3651
Was liked: 3372
Ferran wrote:
Bill Spight wrote:
Since then amateur ranks in Japan have inflated by a couple of stones, at least.


Just to make sure I understand it, because I tend to think of stones as strength and rank as "degree": you mean that for the same skill the declared level has increased a couple of steps?


Yes. You can see this reflected in some book titles. For instance, a book that originally was about becoming shodan is now about becoming 3 dan. ;) But rank inflation has been going on for some time. In one of his books, written in the early in the 20th century, Segoe says that an amateur shodan takes 4 stones from a pro. And if you go back far enough in time, IIUC, an amateur shodan was of professional strength.

Quote:
Quote:
Edit: And as far as making a living at go is concerned, it was said that you needed to be at least 5 dan pro.


I think I've read in a couple of places bits and bites that seem to imply that there IS a significant step between 4th and 5th dan. I think I did a chart some time ago with declared ranks of the Nihon Kiin, and there was a valley at 4th or 5th dan. It kinda felt that you were either good enough, and kept rising, or just not good enough, and you simply didn't reach that level. In fact, if you check the Kansai Kiin the rank with the lowest number of members is 5th dan. I mention that one because I can reach it fast[*]. And I think there were some tournaments that were for young AND below... 5th? dan. And this or that bit, that I can't seem to recall now, but the general feeling was indeed that there was a barrier [psychological?] at 5th dan.


Interesting. :) It could be that the barrier is not so much psychological as cultural/economic. By a certain age, if you can't make a living at go, because you can't get students or open a go club or get books or articles published, you drop out. That could cause the gap, with 5 or 6 dan and above making a living and those below mostly being young pros who still have a chance.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #65 Posted: Thu Jun 16, 2022 9:56 am 
Beginner

Posts: 1
Liked others: 1
Was liked: 2
Rank: OGS 3 kyu
Hi guys,

I put some research into this and here's my attempt at an answer.

As background, I've played chess since I was 7, just on and off, and I'm about a 1500 rating, and I've played go on and off (with some periods of heavy study) for the past 15-years, and I'm currently a 3k on OGS. So this topic has been of interest to me for a long time.

My intuition was that I've put more work into go than chess and therefore I'm stronger, and I think my research confirms that.

I looked at this from a couple different angles: first comparing OGS to chess.com, then comparing top pros in each, and then comparing OGS to USCF ratings.

My conclusion at the end.

Online servers: OGS compared to chess.com

Chess.com shows a small distribution next to your own rating:

Attachment:
rapid.JPG
rapid.JPG [ 17.83 KiB | Viewed 329 times ]


It says the avg rating is 800

There's a distribution of the rankings on OGS (albeit unoficial): https://forums.online-go.com/t/unofficial-ogs-rank-histogram-2021/35705

Attachment:
OGS2.png
OGS2.png [ 76.82 KiB | Viewed 329 times ]


Here the average rank is 9 kyu.

The problem with this comparison is that it's different populations. Chess is much more popular than go (about 800 million players compared to 40 million). And chess.com is way more popular than any go server. The population of chess.com is more "players who know how to play," where the population of OGS is closer to "players who actively play and study."

My own experience doesn't match this either. As a 10-year old, I knew the rules but not much else and I played around an 800 level at online chess, but it took me over a year of playing and some study to reach 9k in go.

Top player comparison

A better comparison would be to look at the progression of top players in each game and compare. The problem here is I couldn't find anything that showed the progression of top Asian go players, so I had to use EGF pros.

here's a great article on chess.com looking at the progression of several top GMs: [url]https://www.chess.com/article/view/chess-grandmaster-hours
[/url]

Top GMs study about 4 hours/day, 5 days a week, for about 1,000 hours per year.

Here's my chart:

Attachment:
chess gms.png
chess gms.png [ 49.6 KiB | Viewed 329 times ]


Here's the average of their progression:

Age Rating
7... 700
8... 1100
9... 1600
10... 1850
11... 2050
12... 2175
13... 2275
14... 2350
15... 2425
16... 2500
17... 2575
18... 2650

I then looked at several of the top EGF players and did my own analysis of their progression: https://drive.google.com/file/d/11oUpQrIRzYlEP90TEySb4XHonupCzLN9/view?usp=sharing

Average of their progression:
Age.... Go rank
6.... 20k
7.... 14k
8.... 16k
9.... 10k
10.... 5k
11.... 3k
12.... 1k
13.... 2d
14.... 2d
15.... 3d
16.... 4d
17.... 5d
18.... 5d
19.... 6d
20.... 6d
21.... 6d
22.... 6d
23.... 1p

Comparing the data yields this:
Age, Go Rank, Chess Rank
714k700
816k1100
910k1600
105k1850
113k2050
121k2175
132d2275
142d2350
153d2425
164d2500
175d2575
185d2650


Took the EGF players a little longer to reach pro than the GMs, but again, not an exact population match. The chess GMs above are top players, "super GMs," while the European Go players, no offense, are closer to average or below average Go professionals.

So I would say the chess ratings here are comparatively stronger than the go ratings for the same age.

Still, I think we're getting somewhere. Especiacially when you consider that "average" chess GMs reach GM level at about age 18: https://chess.stackexchange.com/questions/9891/what-is-the-average-age-to-become-a-grandmaster#:~:text=For%20players%20born%20after%201945,average%20is%2021%20years%20old

here's another good article discussing the GM progression in chess: https://coloradomasterchess.com/informant-ratings-and-expectations/

Magnus Carlsen Compared to Lee Sedol

When you look at two of the best ever at their game, in similar eras, I think it yields a good comparison. Problem is, I couldn't find data on Sedol below 1p.

But here's a summary:
Carlsen
- Age 14: Achieved GM (2500)
- 15: 2625
- 16: 2710
- 18: 2801
- 19: #1 rating in the world
- 23: World champion (probably could have been earlier)

Lee Sedol
- 12: 1p
- 15: 2p
- 17: 3p
- 20: 6p, LG Cup winner
- 20: 7p
- 20: 9p, won Fujitsu cup

Pretty similar arcs.

From this I think we can conclude 1p and GM (2500 rating) are pretty comparable, and it still takes many years for even the best to go from 1p/GM to world champion.

OGS compared to USCF ratings

Finally, I looked at a chart of rating distribution from the US Chess Federation: http://www.uschess.org/archive/ratings/ratedist.php

This data, to me, is the best comparison to the OGS data. I think the populations are closer: people who actively play and study. It also it lines up better with the conclusions from above and my own experience and intuition.

So, here's the final concise comparison I put together comparing OGS and USCF:

Percentile, Go rank, Chess rating
1%35k0
10%25k200
20%20k400
37%15k700
50%12k1000
60%10k1200
74%7k1450
82%5k1600
92%2k1800
96%1d2000
99%3d2200
100%9d2700


No perfect, but I think this is the best comparison I've seen.

Yes, I spent way too much time on this. No, I don't have anything better to do.


This post by Sharptack was liked by 2 people: Elom0, gennan
Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #66 Posted: Fri Jun 17, 2022 11:22 am 
Lives in gote

Posts: 442
Location: Netherlands
Liked others: 243
Was liked: 127
Rank: EGF 3d
Universal go server handle: gennan
My attempt at comparing go ranks (EGF) with chess ratings (USCF)

EGF rank, USCF rating, USCF class
34k500H
27k700G
21k900F
15k1100E
10k1300D
6k1500C
2k1700B
1d1900A
3d2100expert
4d2200master

Note that (deep) DDK OGS ranks may be significantly softer than (deep) DDK EGF ranks, so my table may not contradict your table much in the (deep) DDK range.
Also see https://forums.online-go.com/t/go-ranks-vs-chess-ratings/41594/42

As for FIDE titles, I think:

EGF rank, FIDE rating, FIDE title
5d2200CM
6d2300FM
7d/1p2400IM
3p2500GM
9p2700super GM

Note that FIDE ratings seem to be a bit tougher than USCF ratings (like ~100 rating points).
Also see https://senseis.xmp.net/?FIDETitlesAndEGFGoRatings (table near the bottom of the page)

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #67 Posted: Fri Jun 17, 2022 2:10 pm 
Lives in gote

Posts: 426
Liked others: 18
Was liked: 57
Rank: panda 5 dan
IGS: kvasir
What is the purpose of these comparisons? If it is not only to find a reason to feel good about yourself then you should define the purpose first and proceed from there. When you just make "my attempt" or compare something arbitrary without a clear rational it is literally not supported by any evidence. It is contrived to to compare rank/age of EGF pros to the average rating/age of five arbitrary chess players, and it is especially contrived to compare EUROPEAN Go ratings with USA chess ratings. Unsimilar things can be compared meaningfully only if there is a clear rationale (and a method), for example you could compare apples and oranges by weight if this serves any purpose but not really by which is more fruit.

Consider the following statement:

X kyu Go player is higher in my table than a Y elo chess player.

What is this supposed to mean? Is it true only if enough people feel good about it or could you define precisely what this comparison means?

Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #68 Posted: Tue Jun 21, 2022 11:40 am 
Lives in gote

Posts: 442
Location: Netherlands
Liked others: 243
Was liked: 127
Rank: EGF 3d
Universal go server handle: gennan
kvasir wrote:
Consider the following statement:

X kyu Go player is higher in my table than a Y elo chess player.

What is this supposed to mean? Is it true only if enough people feel good about it or could you define precisely what this comparison means?

@Sharptack used 2 approaches, one by comparing level of mental abilities (using minimum age as an indicator), and one by comparing rank distribution in the communities.

I tried to compare merit/prestige (as percieved by the respective communities of each game).

None of these methods is very accurate ofcourse. Also, there are significant rating differences between different platforms and organisations, in chess and in go. And do we even agree on what defines a "beginner" in go or in chess? Much of this inherenty inaccurate, but I don't think that makes it impossible to make any comparison at all.

As to why one would want to make such comparisons, I would ask, why not? Some people are just curious about such matters, and they may not mind if comparison estimates are inaccurate and somewhat speculative.

If you're not interested in such comparisons, you are free to ignore it.


This post by gennan was liked by: Sharptack
Top
 Profile  
 
Offline
 Post subject: Re: Comparison of go rank with chess rating
Post #69 Posted: Tue Jun 21, 2022 12:13 pm 
Lives in gote

Posts: 559
Liked others: 129
Was liked: 106
Rank: OGS ddk
KGS: Ferran
IGS: Ferran
OGS: Ferran
My apologies if I've written this before. My brain is mushy, right now.

I see an option. Find a KataGo/Leela Go network with strength equivalent to a certain human ELO. Train a chess robot with the same architecture. That's the equivalent rating.

It has its pitfalls, but it's relatively easy to quantify.

Take care.

_________________
玄 之 玄

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 69 posts ]  Go to page Previous  1, 2, 3, 4

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group