Underranked & overranked countries

[mohsart]

I'm mostly interested in European countries, though it is interesting on a global scale as well.
In another thread, I mentioned that the Swedish ranking system treats some countries special because they are believed to generally have overranked players.
The countries are: Japan, The Americas, UK, Portugal, and what used to be Jugoslavia.

Only one country (Korea) is believed to have underranked players.

I'm a bit interested in whether this (still) is true or not.

As for Japan and the Americas, I think it was and is true that they are overranked compared to Europe, but is e.g. UK players really "weaker" than, say, Germans?
(I took this example due to a Swedish player saying that he'd rather go play in a German tournament than to London Open, because it'd be easier for him to earn ranking points in Germany.)

Also, in a nearby thread (which is one reason for me to start this one) a couple of other countries were mentioned as "strong".

/Mats

[freegame]

some basic stats can be found on the EGD database
http://www.europeangodatabase.eu/EGD/Stats_Country.php

[mohsart]

While interesting, I don't believe that EGD provides the Absolute Truth.
This said, how do I read that? A low number in Difference implies overranked and a high underranked?

/Mats

[Kirby]

How do you expect to find "absolute truth"? The only thing you can really do is look at game stats. Even then, it's simply "the more data the better".

[Laman]

it slightly depends if you ask about ranks or ratings. ratings are more reliable but still not the same everywhere

i don't have much of a personal experience, but as far as i know, for example Czech ranks are stronger than European average (= Czechs are underranked) and in comparison German ranks are generally weaker than average (Germans are overranked). differences for other countries surely exist as well

it is due to two main reasons:

1) lack of games between players from different countries

2) different policies in gaining ranks. following the same example as before, in Czechia we reset ratings quite rarely and you have to get proper GoR to achieve a rank. (you get 3k when you hit GoR 1800, then 2k with 1900 or again 4k when you fall under 1700). from what i've heard, in Germany it is different - you have more freedom to reset your rank (and with the rank also your rating) when registering to tournament, without much supporting evidence for your improvement. (anyone, please correct me if i am wrong about this)
this practice injects a lot of points to the local system and relatively weakens the local rating

[mohsart]

Indeed, the more facts the better, and I'm not expecting to find the absolute truth.
But, I expect that everyone can agree that the GoR system has it's flaws?
Some think they are quite large, and others that they are minor.
The way I read the chart is that it compares declared rank with GoR, I could be wrong though but if I'm not, it's not really the kind of statistics I think helps much.
Better would be statistics from e.g. EGC's on how various countries performs.
If those kind of statistics are hard to produce without much manual work, at least a bunch of peoples view on it, e.g. "I'm always worried when I play in country xyz because I've never won more than one game there"

/Mats

[henric]

some basic stats can be found on the EGD database
http://www.europeangodatabase.eu/EGD/Stats_Country.php

That table is not relevant to the question. The table only deals with
the relation between kyu/dan ranks and GoR. In order to say anything about the
regional differences in ranks and ratings one has to look at actual winning
statistics in games between players from different countries.

In the article at page 10-15 in the Nordic Go journal number 2 2007
(http://homepage.mac.com/bjornwendsjo/go/2-07.pdf ) the regional differences are discussed,
as well as a few different methods to measure regional differences.
The data that are used there don't take into account any possible changes
with time, they lump together tournament data from 1996-2006.

However, the same kind of analysis was made using only data from the 2008
go congress and is presented in the Nordic Go Journal number 1 2009
http://goforbundet.se/ng/200901.pdf, in figure 2 on page 12 and in tables
4-6 on page 13.

There are problems with statistical significance, but I would conclude that
the old tendencies are that e.g. UK and a few other european countries, and japanese
participants at the EGC:s being relatively over ranked and over rated, while
some other countries, e.g. Finland and Sweden, France etc are relatively under ranked
and under rated. Those conclusions didn't seem to have changed much when the
analysis was performed only for (the EGC results in ) 2008. So I don't have the
impression that the regional discrepancies change much - there are too few games
being played between players from different countries.

In the case of Finland and Sweden I believe the reason for the underrating is that
they had a period with very fast increase in the number of players in
2001-2004, and consequently many young players improving simultaneously so that
the rating systems couldn't cope. The same I think is true for Spain, for example.

Normally the regional discrepancies are less than 100 GoR points and smaller at
higher ratings, but they can be larger in the extreme cases. I don't think they
are a big problem, but one should be aware of them and maybe check that they are
not growing.

The phenomenon doesn't carry much weight in the Swedish rating system, only there
is a provision to look more carefully in cases where someone has say many games
against "overrrated" countries in his/her portfolio.

cheers,
Henric

[mohsart]

It would actually be very interesting to see the statistics regarding both GoR and declared rank for every EGC, and possibly other "big", "international" tournaments e.g. Paris Open and London Open.

Worth to note in this discussion is that GoR is used in quite different ways in different countries. In e.g. Sweden and, I believe, Germany it is totally disregarded, while in other countries it is followed very strictly.

While this thread is not about GoR and it's qualities/flaws, it would still be interesting to see if using GoR increases the regional differences or decreases them.

I do understand that statistics from only a few tournaments doesn't prove much, but I'd still think It'd be a interesting read.

/Mats

[Phelan]

I'm mostly interested in European countries, though it is interesting on a global scale as well.
In another thread, I mentioned that the Swedish ranking system treats some countries special because they are believed to generally have overranked players.
The countries are: Japan, The Americas, UK, Portugal, and what used to be Jugoslavia.

Only one country (Korea) is believed to have underranked players.

I'm a bit interested in whether this (still) is true or not.

I noticed you mentioning that before, and I was curious as to what perception people had of Portuguese ranks. Some Spanish players that played in one of the last open tournaments were impressed with the strength of our 1-dans. We mostly use the EGF ranks, but there aren't many foreign players coming here, or vice-versa.

[tapir]

it slightly depends if you ask about ranks or ratings. ratings are more reliable but still not the same everywhere

when czech 9k / 8k play as 2k on KGS, it is likely that there is something wrong with the czech system. as said in the neighbour thread, discouraging / forbidding rank resets for kyu players makes the system less reliable not more.

[Magicwand]

i am kgs 4d. if i goto korean goclub then i will say i am 3kyu.
goclub rating in korea is very different from rating you are familiar with.
i guess it is the result of playing for money on everygame people sandbagging too much.
just my thought.

[HermanHiddema]

i am kgs 4d. if i goto korean goclub then i will say i am 3kyu.
goclub rating in korea is very different from rating you are familiar with.
i guess it is the result of playing for money on everygame people sandbagging too much.
just my thought.

THis, I assume, is because the club have not adopted the amateur dan ranks yet? As explain here: http://senseis.xmp.net/?RankGupKoreanExplained

[mohsart]

Chinese 5d can make KGS 9d, but this thread has nothing to do with online vs IRL ranks.

/Mats

[mohsart]

I noticed you mentioning that before, and I was curious as to what perception people had of Portuguese ranks. Some Spanish players that played in one of the last open tournaments were impressed with the strength of our 1-dans. We mostly use the EGF ranks, but there aren't many foreign players coming here, or vice-versa.

I guess I should comment this.
I really don't know, I personally have no experience playing against Portugese players and I don't recall having seen statistics proving Portugese players being overranked.

/Mats

[LovroKlc]

the Czech rankings are particularly strong. for example, take the Brno tournament(as far as I know the biggest tournament in Czechia) and look at foreign vs. Czech players on stronger kyus. I can also tell from experience. I am sure that a lot of Czech 3k-4k can make shodan in a lot other Euro countries.