Why my opponent gain rating more higher (like me have 2300+ rating) even my opponent have much stronger rating?
And why i always gain rating only 4-8 only even with rating 1500?
This make my rating decrease significantly.
"I'm feel cheated here!"
Rating differential refers, basically, to the confidence behind your rating. A low RD means you have played many games against players rated similarly to you and the formula behind your rating is more sure of it. In this case, your rating with fluctuate much less with each result. On the other hand, with a high RD there is less certainty behind the rating, and greater changes with each result. If you and your opponent seemingly experience wildly different swings like this, it is probably because one of you has a much higher RD than the other.
What RD mean? And why my rating such like 2300+ not my current rating?
And for the original ELO rating rule, the sum of the fluctuation value will be the latest rating not by most games played.
lichess not uses ELO system. They use a variant of Glicko2. Read
en.wikipedia.org/wiki/Glicko_rating_system@AdityaMH “What RD mean?”
lichess.org/forum/lichess-feedback/not-fair-rating#2Nobody said anything at all about “most games played”. You clearly didn’t read what was written, why should anybody bother trying to help you understand?
the chess rating system models a persons rating as a bell curve with center on the current rating. RD is related to the width of that bell curve. CT probably what was happening is that you have a more stable 2300 rating, than your opponent did, so it makes smaller changes to your rating, while making big changes to his. Then his would go up or down more.
What's this 2300 rating thing? I didn't get that.
I thought you said you were playing another 2300 and his rating changed much more dramatically with the game results than yours. The rating system models each player by a gaussian distribution centered on your rating. Meaning it predicts your rating is the most likely actual rating but it has a bell shape tailing of.
If the gaussian distribution has high standard deviation it means the rating system isn't sure about it's predicted rating so the distribution is more spread out for that player and when they play other people even if a similar rating their rating will fluctuate more.
When the rating system is confident in your predicted rating (because you have been stable around a certain rating and you have played a variety of opponents, the standard deviation of the gaussian is smaller meaning that the bell curve predicting your rating is more narrowly centered on your actual rating.
so even if you both have similar rating, it could still be the rating system was confident in your rating (why it fluctuated less) and was not so confident of your opponents rating. (meaning each game would affect his rating more.)
Ah, I get it. So, more confident or stable rating you have will less rating you get (if win). So gaining rating score not by current rating but by stable rating statistics.
If someone have win/loss streak more, maybe can decrease RD or is it still influential if someone challenging a much stronger curent rating (if win)?
