Rating Histogram Comparisons
This article provides a table and two graphs comparing three Go rating systems (AGA, EGD and KGS). A chess rating system (USCF) is included for those interested in comparisons, or more familiar with it.
|Table of contents|
This table shows the rating you need to be at a given percentile of the players in each rating system.
|99.9%||9.64||2747||7–8d / 2–3p||pro||2643|
|top||10.12||2809||8d / 4–5p||9p||2789|
 KGS and AGA ratings are in AGA form: ``n`` kyu ≈ ``-n``½, ``n`` dan ≈ ``+n``½, with a jump from -½ to +½.
 The EGD cuts off at rating 100 = 20 kyu. This distorts the EGF graphs for the 1% and 2% percentiles.
 The 99.9% and top rating for KGS have been estimated as 2850 (8.5 dan amateur) and 3000 (10 dan amateur).
 USCF (United States Chess Federation) ratings, e.g. a master has a 2200 rating.
The following diagram plots the four rating systems as a cumulative distribution function: the vertical axis shows the percentage of players up to the level indicated along the horizontal axis. The top scale shows USCF ratings, while the bottom scale shows AGA ratings.
Usage Example: You have rating 2000 USCF, and are curious what that is equivalent to in go. Find the point where the USCF graph crosses 2000, which is at about 90%. Then move straight up to see where the other graphs cross that percentile, which is (roughly) 2200 EGF (2 dan), 3 dan KGS and 5 dan AGA.
The following diagram plots the same data as a cumulative histogram.
The correspondence of the Elo (left) and Dan/Kyu (right) scales is based on EGF assignments.
- AGA (American Go Association): http://usgo.org/ratings/RatingsQuery.asp - Ratings updated in or after 2005-11, as of 2007-11-6.
- EGF/AGD (European Go Federation / Database): http://www.europeangodatabase.eu/EGD/
- KGS (KGS Go Server): KGS Rank Histogram / Data
- USCF: http://www.geocities.com/siliconvalley/lab/7378/histogra.htm
The raw data for KGS and USCF ratings are in the form of histogram buckets. I used linear interpolation to estimate more accurate numbers. The raw data for the AGA and EGF was the entire list, so I simply used the rating of the players at the appropriate percentiles.