European Kyu Inflation
Floris?: I am getting the idea European kyu players (esp. Dutch) have an inflated rank compared to dan players and strong kyus (1k/2k) in Europe. This I conclude from the fact that European dans are usually have a stronger rating on KGS than their EGF rating, whilst European kyu players usually have a lower rank (sometimes even by two or more stones) on KGS than their european rating. According to the diagrams here on SL KGS *should* be weaker than European ranks. Is this completely wrong or are these (a lot) Europeans just overestimating themselves?
Charles Based on a number of examples, I concluded that KGS ratings are less than one stone apart, with coincidence at around 1 k = 1 kyu.
mgoetze: As far as I can tell, EGF ratings are much more consistent than KGS ratings. Also note that many European players rank themselves higher than their rating on the EGF list...
Warp: Another (simple) explanation for the phenomenon described by Floris: It may be that KGS ranks are a bit less separated than EGF ranks (eg. if we think that EGF ranks are 1 stone apart, it may be that KGS ranks are for example 0.9 stones apart). EGF and KGS ranks are equal somewhere around 5k, which would explain the phenomenon. Just an idea.
Petri KGS version of maximum likelihood estimation is bit unstable. I would not trust it too much. Also behaviour of player polulation has impact on ranking system. People are not willing to play ranked games with handicaps, thus ranking algorithm does not get direct measurements results of strengths of people few stones apart, which in turn makes possible that rank step can be less or more than a stone. Even so that in some parts of curve it is less than a stone and in some parts it is more than a stone.
xue I don't know about the rest of Europe, but in Switzerland standards are really low: on KGS I'm ranked at 15 kyu, but I just participated in my first tournament, and they put me down as 10 kyu...go figure.
Niklaus: Being familiar with the Swiss Go scene as well as the rest of Europe, but I can tell you that Swiss ratings aren't any weaker than the rest of Europe. I don't think the problem lies with the real life ratings but rather with KGS. There's something strange going on if EGF 6 dans are 8 or 9 dan on KGS while EGF shodans are 4 kyu or so. At least that was the case with cgoban 2, let's hope that cgoban 3 fixes these issues.
Velobici: KGS ratings using Cgoban2 were finer grained than EGF ratings...so that there were a larger number of discrete steps between an EGF 1d on KGS and an EGF 6d on KGS using Cgoban2. The discrepancy, if any, between a single national federation, such as the Swiss, and the combination of all other EGF federations can be seen on the EGD site at
http://lnx.agi.go.it/EGD/ScatterRank.php?country=CH&when=now&vers=4 as you can see from 20k to 4d the Swiss ratings closely track the GoR, as opposed to say Thailand
http://lnx.agi.go.it/EGD/Stats_Country.php?ricerca=1&country_code=TH
maruseru EGF ranks aren't that reliable either. They work best if you play lots of international tournaments, and then only if you enter each tournament with a rank based on your current EGF points. A real-life 20k could, in theory, enter a tournament as 2d, lose every game in the tournament, and still have the points of a 1-2d. That's an extreme example, but it does happen that some people use their current EGF points, others use their KGS rank, still others use what they feel like. Further, when playing only local tournaments with no or little international participation the ranks of players from that area could be, for a lack of a better word, incestuous.
Maybe EGF points need a probability coefficient assigned to them to see how far they can be trusted. The more international tournaments you play while entering based on your current points the higher the coefficient would be. In Czechia for example, I've been told, players are only allowed to enter tournaments at their current EGF rank. In Austria dan players have to enter with their EGF rank; kyu players can use their EGF rank as the upper (weaker) bound and half their EGF rank as the lower (stronger) bound. For example, if you are an EGF 12k, you can enter as 6-12k. We've arrived at that rule because some players don't play a lot of tournaments, but of course they still improve.
Hu: This page was started in 2003. In December 2004 there was a ratings change on KGS that made kyu ranks much harder than before. On 18 September 2006, KGS ranks were updated and brought more into line with other ranking systems. So references to KGS ranks above are not very meaningful. However, as always, ranking systems are not directly comparable since people treat games in different ways depending on context and may be stronger over a wooden board than on the Internet, or vice versa.