Life In 19x19
http://lifein19x19.com/

Iyama's world ranking
http://lifein19x19.com/viewtopic.php?f=13&t=13750
Page 5 of 6

Author:  Uberdude [ Tue Oct 30, 2018 10:38 pm ]
Post subject:  Re: Following Iyama Yuta (no world ranking discussions)

<admin> This and next post moved here from viewtopic.php?f=13&t=13751</admin>
TheCannyOnion wrote:
ez4u wrote:
Apparently we now want to say that Go Ratings (actually past go ratings) is the 'truth' and what we actually see happening in the world around us should conform to that or be subject to criticism. However, the Go Ratings algorithm itself rejects this idea since AFAIK it retrospectively adjusts past ratings based on current results.


And no, Goratings does not retrospectively adjust past ratings based on current results. What a ridiculous thing to say! If that were the case, there would be no past ratings.


Wrong. It does. Go read the Whole History Rating paper, or use archive.org to observe this effect.

Author:  TheCannyOnion [ Tue Oct 30, 2018 11:27 pm ]
Post subject:  Re: Following Iyama Yuta (no world ranking discussions)

Uberdude wrote:
TheCannyOnion wrote:
And no, Goratings does not retrospectively adjust past ratings based on current results. What a ridiculous thing to say! If that were the case, there would be no past ratings.

Wrong. It does. Go read the Whole History Rating paper, or use archive.org to observe this effect.


No, you are wrong. Nowhere in the paper does it say that it "retrospectively adjusts past ratings based on current results" or something to that effect.

Author:  Uberdude [ Wed Oct 31, 2018 12:29 am ]
Post subject:  Re: Iyama's world ranking

On the other Iyama thread and now moved to above this post, TheCannyOnion demonstrated his ignorance of the whole history rating method as used by goratings.org.

TheCannyOnion wrote:
Uberdude wrote:
TheCannyOnion wrote:
And no, Goratings does not retrospectively adjust past ratings based on current results. What a ridiculous thing to say! If that were the case, there would be no past ratings.

Wrong. It does. Go read the Whole History Rating paper, or use archive.org to observe this effect.


No, you are wrong. Nowhere in the paper does it say that it "retrospectively adjusts past ratings based on current results" or something to that effect.


Here is an attempt to remedy that.

As confirmed by Rémi Coulom, the creator of Whole History Rating and goratings.org, on this forum at https://www.lifein19x19.com/viewtopic.p ... 40#p193740: (nice explanation from ez4u why this is a good idea shortly after at https://www.lifein19x19.com/viewtopic.p ... 55#p193755).

Rémi wrote:
John Fairbairn wrote:
Remi: The 2000 rating list is headed 2000-01-01. So not only should it not reflect the results for 2001, surely it shouldn't include the results for 2000 either. On that basis, go4go has just two games, for 1997, scoring 1-1. It's well known I'm not a numbers guy, so what's the trap I've fallen into?


The WHR rating algorithm has retroactive effect: the rating of one day depends on games of the past and games of the future. That is what makes it more accurate than incremental rating algorithms.


WHR paper: https://www.remi-coulom.fr/WHR/WHR.pdf

First, some clues of what to read in the WHR paper. In the intro it describes the shortcomings of incremental rating systems such as Elo:
Quote:
Incremental Rating Systems. ... After each game, this data is updated for the participants in
the game. The rating of the winner is increased, and the rating of the loser is
decreased.

Incremental rating systems can handle players of time-varying strength, but
do not make optimal use of data. For instance, if two players, A and B, enter the
rating system at the same time and play many games against each other, and
none against established opponents, then their relative strength will be correctly
estimated, but not their strength with respect to the other players. If player A
then plays against established opponents, and its rating changes, then the rating
of player B should change too. But incremental rating systems would leave B’s
rating unchanged.


Then discusses more advanced approaches under "Accurate Bayesian Inference"
Quote:
Authors of incremental algorithms already proposed to correct inaccuracies by
running several passes of the algorithm forward and backward in time


and then WHR:
Quote:
”. Experiments presented in this paper clearly indicate that he underestimated
his idea: evaluating ratings of the past more accurately helps to evaluate
current ratings
: the prediction rate obtained with WHR outperforms decayed
history and incremental algorithms.


Evidence that past ratings change:

Iyama's rating on goratings.org on Oct 15th 2016, as captured by archive.org was 3520:
https://web.archive.org/web/20161015073 ... tings.org/

Iyama's rating on the current goratings.org site (https://www.goratings.org/en/players/601.html), looking in his game history for the relevant period shows his rating is now thought to have been 3533
Code:
2016-10-17   3533   White   Win   Yu Zhengqi   3364   ♂   jp   View game
2016-10-12   3533   White   Win   Takao Shinji   3336   ♂   jp   View game
2016-10-04   3533   Black   Win   Takao Shinji   3336   ♂   jp   View game


3533 != 3520

QED.

(Edit: thanks jlt, corrected)

Author:  jlt [ Wed Oct 31, 2018 1:02 am ]
Post subject:  Re: Iyama's world ranking

You mean 3533 != 3520.

Attachment:
Iyama.png
Iyama.png [ 38.05 KiB | Viewed 12414 times ]

Author:  dfan [ Wed Oct 31, 2018 6:44 am ]
Post subject:  Re: Iyama's world ranking

Ridiculous or not, it is absolutely the case that WHR, the algorithm used by goratings.org, changes its opinion about the rating of a given player on a given date once future results are made available to it. There are plenty of valid reasons to regard this as either a positive or a negative feature of WHR, but it is a fundamental property of the system.

(I have implemented WHR.)

Author:  hyperpape [ Fri Nov 02, 2018 6:52 am ]
Post subject:  Re: Iyama's world ranking

While it's cool to see Cho U making a bit of a comeback, it's somewhat disappointing that the recent title matches have been less a case of new players rising up to meet Iyama so much as him falling back to the level of the other Japanese pros.

Author:  gowan [ Mon Nov 12, 2018 10:10 am ]
Post subject:  Re: Iyama's world ranking

As for Iyama falling back, I think it is to be expected. Even the strong Chinese and Korean pros who win international tournaments don't stay at the top for ever. All of them have a run at the top and then fall back. It's a standard thing for a pro to win a lot of titles an hold them for a while but falling back, as you say, after a while. Cho U is having a comeback but he was at the top a while ago. Rin Kaiho was at the top for a while until Ishida dislodged him. Cho CHikun and Kobayashi Koichi held top titles for long periods of time but fell back. It seems to me to be a matter of being able to put out at a very high level of intensity for a long time. After all there is pressure from fans, journalists, other players, to stay at the top. And these title holders have life outside of go which eventually can make demands. Speaking of that, we don't read or hear much about the lives of our go idols away from the go board. I do recall hearing that Iyama was divorced. All the top international pros eventually fall back: Ke Jie, Yi Changho, Yi Setol, Ma Xiaochung, etc.

Author:  hyperpape [ Mon Nov 12, 2018 10:40 am ]
Post subject:  Re: Iyama's world ranking

I see a difference in those scenarios.

Before falling back, Lee Changho had Lee Sedol nipping at his heels. Then Park Junghwan caught up to Lee Sedol while he was still on top of the world. Shin Jinseo has now caught Park, while they're both among the best in the world. Even Cho, though he fell off quite rapidly after 2012, remained quite competitive for a few years: he beat Iyama in 2011 Kisei. In contrast, none of the players taking titles from Iyama seem likely to reach his former level of play, unless they surprise us.

As a fan, it's exciting to see a new upstart make the top player sweat while they're still at the top of their game. It's also bad for Japanese go on the international scene if they can only field one top-20 player in a decade.

Author:  TheCannyOnion [ Mon Nov 12, 2018 8:59 pm ]
Post subject:  Re: Iyama's world ranking

Holy cow, Iyama has dropped all the way to 55th on Mamumanu's latest ranking.

Author:  jlt [ Fri Nov 23, 2018 1:35 pm ]
Post subject:  Re: Iyama's world ranking

Iyama is back to the 24th place on goratings.org while Shin Jinseo is 1st.

Author:  TheCannyOnion [ Sat Feb 09, 2019 9:55 pm ]
Post subject:  Re: Iyama's world ranking

Iyama has fallen out of top 30 on Goratings. He's currently ranked 31st. On Mamumamu's ranking, Iyama is at 57th place.

Author:  macelee [ Sun Feb 10, 2019 12:44 am ]
Post subject:  Re: Iyama's world ranking

That reflects his relatively poor performance in 2018. His score against Korean/Chinese opponents is 5-12.

Author:  Uberdude [ Sun Feb 10, 2019 9:50 am ]
Post subject:  Re: Iyama's world ranking

Against his Japanese* opponents he also didn't do as well as the last few years: losing the Gosei (0-3) to Kyo Kagen, and the Meijin to Cho U.

Author:  TheCannyOnion [ Fri Feb 15, 2019 11:19 pm ]
Post subject:  Re: Iyama's world ranking

Iyama has risen to #29 on Goratings, probably due to his recent wins against Yamashita Keigo in Kisei title defense.

However, this means Iyama is ranked higher than Park Yeong-hun. Riiiiight... I would find it believable if not for the fact that Park has reached the last two Chunlan Cup finals in a row, beating Lian Xiao (twice), Gu Zihao, Dan Yifei, Zhou Ruiyang, and Ke Jie in the process. These were all strong opponents and, with the exception of Lian Xiao, all world-title holders.

What is up with Goratings? It can't be all due to incomplete data. There has to be something awfully amiss.

Author:  pookpooi [ Sat Feb 16, 2019 12:37 am ]
Post subject:  Re: Iyama's world ranking

TheCannyOnion wrote:
Iyama has risen to #29 on Goratings, probably due to his recent wins against Yamashita Keigo in Kisei title defense.

However, this means Iyama is ranked higher than Park Yeong-hun. Riiiiight... I would find it believable if not for the fact that Park has reached the last two Chunlan Cup finals in a row, beating Lian Xiao (twice), Gu Zihao, Dan Yifei, Zhou Ruiyang, and Ke Jie in the process. These were all strong opponents and, with the exception of Lian Xiao, all world-title holders.

What is up with Goratings? It can't be all due to the data source. There has to be something awfully amiss.

Iyama Yuta is the last year's LG runner-up while #6 Lian Xiao isn't even a runner-up in any international tournament. And the #1 Shin Jinseo didn't top any international championship (Crownless King).

Author:  TheCannyOnion [ Sat Feb 16, 2019 1:36 am ]
Post subject:  Re: Iyama's world ranking

pookpooi wrote:
TheCannyOnion wrote:
Iyama has risen to #29 on Goratings, probably due to his recent wins against Yamashita Keigo in Kisei title defense.

However, this means Iyama is ranked higher than Park Yeong-hun. Riiiiight... I would find it believable if not for the fact that Park has reached the last two Chunlan Cup finals in a row, beating Lian Xiao (twice), Gu Zihao, Dan Yifei, Zhou Ruiyang, and Ke Jie in the process. These were all strong opponents and, with the exception of Lian Xiao, all world-title holders.

What is up with Goratings? It can't be all due to the data source. There has to be something awfully amiss.

Iyama Yuta is the last year's LG runner-up while #6 Lian Xiao isn't even a runner-up in any international tournament. And the #1 Shin Jinseo didn't top any international championship (Crownless King).


And Park Yeong-hun was last year's runner-up in the 12th Chunlan Cup and the year before that (11th Chunlan Cup). Park beat two players in the world top-10 to get to last year's finals. Iyama beat NONE to reach LG Cup finals. Iyama had simply been pretty lucky with his pairings in the LG cup. Moreover, for the 11th Chunlan Cup, Park beat Zhou Ruiyang, Lian Xiao, and Ke Jie, ranked at the time 12th, 7th, and 1st, respectively. That's a pretty impressive performance from an 'old' war horse who, at the time, was ranked 17 places lower than Iyama, according to Goratings.

True, Lian Xiao has not reached a final of an international major, but neither has Honinbo Dosaku or Honinbo Jowa, or Go Seigen for that matter. But, it's safe to say that Lian Xiao has faced far stiffer competition than Iyama, and that he stands a far better chance of actually getting an international title.

Back to Iyama and Park... Park's great performance in the 12th Chunlan Cup is more recent: it happened a couple of months ago. Iyama's performance in LG cup was from the 2017 edition that concluded in early 2018.

All things considered, something is seriously amiss with Goratings: I think it inflates Japanese players' strength.

Author:  jlt [ Sat Feb 16, 2019 1:46 am ]
Post subject:  Re: Iyama's world ranking

Park Yeonghun has recently beaten strong opponents like Park Junghwan (3630) and Gu Zihao (3593), but also lost against low rated players like Kim Hyenchan (3246) and Song Gyusang (3329).

I wanted to check by myself if Iyama's rating reflects his performance against Chinese and Korean players. I considered a player with initial rating 3470, and I made him play the same 17 matches than Iyama's since January 2018 with the same result, and repeated that 10000 times. After each of the 17000 matches, the rating is incremented by c(match_result - 1/(10D/400+1)) where c=20 is an arbitrary constant and D is the difference between the opponent's rating and the current player's rating.

The player ended with a rating of... 3339 points, which puts him at the 98th place on goratings.

The constant c=20 is arbitrary, but changing it to other reasonable values (between 1 and 40) doesn't change the conclusion much.

I repeated the same experiment with the 14 matches in 2017 against Chinese and Korean opponents. The final rating is 3524, which corresponds to the 15th place.

If we take into account all the 31 matches since January 2017, we get a rating of 3410 (rank=54th).

My conclusion is that maybe Iyama is a bit overrated, but the variations are so wild that it's impossible to get a reliable rating by only taking into account non-Japanese opponents.

Here is a scilab code, for those who would like to check.

rating=3470;

opp_ratings=[3652,3652,3639,3519,3641,3587,3475,3496,3502,3502,3500,3609,3531,3631,3578,3488,3516,3517,3517,3431,3647,3565,3447,3588,3585,3448,3483,2969,3458,3465,3542];

match_results=[1,0,0,0,0,0,1,1,0,1,0,0,1,1,0,0,0,1,0,0,0,0,0,1,0,0,0,1,1,1,0];

n=length(opp_ratings);
c=20;

for number_of_loops=1:10000;

for i=1:n;
D=opp_ratings(i)-rating;
S=1/(10^(D/400)+1);
rating=rating+c*(match_results(i)-S);
end

end
disp(rating)

Author:  pookpooi [ Sat Feb 16, 2019 1:49 am ]
Post subject:  Re: Iyama's world ranking

TheCannyOnion wrote:
Park beat two players in the world top-10 to get to last year's finals, while Iyama beat NONE to reach LG Cup finals.

Iyama beat #26 Lee Yeongkyu in the first round, #24 Zhou Ruiyang (2013 Bailing Cup winner) in the second round , #8 Yang Dingxin (2019 LG Cup winner) in the third round, #2 Ke Jie in the semi-final, lose 1-2 against #11 Xie Erhao in final.

Author:  TheCannyOnion [ Sat Feb 16, 2019 1:54 am ]
Post subject:  Re: Iyama's world ranking

pookpooi wrote:
TheCannyOnion wrote:
Park beat two players in the world top-10 to get to last year's finals, while Iyama beat NONE to reach LG Cup finals.

Iyama beat #26 Lee Yeongkyu in the first round, #24 Zhou Ruiyang (2013 Bailing Cup winner) in the second round , #8 Yang Dingxin (2019 LG Cup winner) in the third round, #2 Ke Jie in the semi-final, lose 1-2 against #11 Xie Erhao in final.


Ah, I missed Ke Jie. However, Yang Dingxin was not ranked in the top-10 at the time of his game against Iyama, and as far as I know, Xie was not even in the top-15.

Author:  TheCannyOnion [ Sat Feb 16, 2019 2:02 am ]
Post subject:  Re: Iyama's world ranking

jlt wrote:
Park Yeonghun has recently beaten strong opponents like Park Junghwan (3630) and Gu Zihao (3593), but also lost against low rated players like Kim Hyenchan (3246) and Song Gyusang (3329).

I wanted to check by myself if Iyama's rating reflects his performance against Chinese and Korean players. I considered a player with initial rating 3470, and I made him play the same 17 matches than Iyama's since January 2018 with the same result, and repeated that 10000 times. After each of the 17000 matches, the rating is incremented by c(match_result - 1/(10D/100+1)) where c=20 is an arbitrary constant and D is the difference between the opponent's rating and the current player's rating.

The player ended with a rating of... 3339 points, which puts him at the 98th place on goratings.

The constant c=20 is arbitrary, but changing it to other reasonable values (between 1 and 40) doesn't change the conclusion much.

I repeated the same experiment with the 14 matches in 2017 against Chinese and Korean opponents. The final rating is 3524, which corresponds to the 15th place.

If we take into account all the 31 matches since January 2017, we get a rating of 3410 (rank=54th).

My conclusion is that maybe Iyama is a bit overrated, but the variations are so wild that it's impossible to get a reliable rating by only taking into account non-Japanese opponents.

Here is a scilab code, for those who would like to check.

rating=3470;

opp_ratings=[3652,3652,3639,3519,3641,3587,3475,3496,3502,3502,3500,3609,3531,3631,3578,3488,3516,3517,3517,3431,3647,3565,3447,3588,3585,3448,3483,2969,3458,3465,3542];

match_results=[1,0,0,0,0,0,1,1,0,1,0,0,1,1,0,0,0,1,0,0,0,0,0,1,0,0,0,1,1,1,0];

n=length(opp_ratings);
c=20;

for number_of_loops=1:10000;

for i=1:n;
D=opp_ratings(i)-rating;
S=1/(10^(D/400)+1);
rating=rating+c*(match_results(i)-S);
end

end
disp(rating)


Your speculation gels quite well with Mamumamu's ranking, which has Iyama at #64 currently. Incidentally, per a Reddit thread, Remi was investigating whether Goratings inflates Japanese players' strength. I don't know what became of his investigation.

https://www.reddit.com/r/baduk/comments/71je7h/world_rankings/

Page 5 of 6 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/