> IM and GM are not awarded by rating Correct. That's why I used phrases like "typically" and "International Master level". The mentioned ratings are part of

> IM and GM are not awarded by rating Correct. That's why I used phrases like "typically" and "International Master level". The mentioned ratings are part of the requirements, and are STRONGLY associated with those titles. "The requirements for becoming a Grandmaster are somewhat complex. A player must have an Elo rating of at least 2500 at one time (although they need not maintain this level to keep the title). A rating of 2400 or higher is required to become an International Master. In addition, at least two favorable results (called norms) in tournaments involving other Grandmasters, including some from countries other than the applicant's, are usually required before FIDE will confer the title on a player. There are other milestones a player can achieve to get the title, such as winning the Women's World Championship, the World Junior Championship, or the World Senior Championship. Current regulations may be found in the FIDE Handbook.[12" http://en.wikipedia.org/wiki/Grandmaster_%28chess%29 http://en.wikipedia.org/wiki/Grandmaster_%28chess%29 http://www.fide.com/fide/handbook.html?id=58&view=article http://www.fide.com/fide/handbook.html?id=58&view=articl... For a National Master (depends on country), here is the USCF: "The United States Chess Federation (USCF) awards the Title of National Master to anyone who achieves a USCF rating of 2200, and the title of Senior Master to anyone who achieves a USCF rating of 2400." http://en.wikipedia.org/wiki/Chess_master http://en.wikipedia.org/wiki/Chess_master A computer playing at well above 2500 can be said to be playing at Grandmaster level, despite not filing the paperwork for officially conferring the title. > When you do percentage based interest, you're making it easier to go up when you're at a higher rating, which is the wrong way round. Compound interest is obviously an oversimplification for learning, which everyone already knows is not a linear process. It's illustrative, however, of the vital point: small differences in a feedback cycle accumulate over time into a much larger absolute difference.