Saturday, January 16, 2016

MCF Chess Rating: A typical inquiry that I received after every new quarter release

After every rating release, there is always inquiry on how rating calculation is calculated. The first one arrived 3 hours after I released the Pre-release Article on Friday morning. Subsequently, there are couple of more inquiries until the latest one about 2 hrs ago.

Guess who did the inquiry? The one who who gained few hundred rating points or the one who lost few hundred rating points. Sorry ..... No prize for guessing correctly.

First of all, you need to understand F4/F5 Add. Maths on a topic called "Probability". If you have a rating of 1300 and happen to play someone at 1300, then you are expected to score half point i.e. 0.5. If you win, then you have +0.5 from that game,. If lose, then you have -0.5 from the game.

Say. you play all 10 games against players with an average rating of 1300 and won a perfect 10 score. Then, you have scored +5. The K-factor for player from rating 1000 and 1399 is 40. Then you are said to gain (+5 * 40), that is, a massive 200 rating points gained from the event [This is the same algorithm use in FIDE rating calculation]. Now, say you lost all 10 games, then you will lost 200 points too. Usually this kind of extreme scores result seldom happen, however, this quarter we have witnessed at least from one player from each extreme end.

 
Another example at the other extreme end:- A player who supposed to score no more than 1.5 points based on her rating (1145) but did well by scoring 5 points from 9 games including wins against players who are 200 to 300 rating points higher.