- Joined
- 11 Jan 08
- Moves
- 28790
There may have been previous discussions of this topic, but unfortunately I do not have the time to review all threads.
I think the current way of calculating ratings - gives a distorted readout.
Let's assume I am a player with a "True" rating of 1400. I start 500 games at the same time. I play those 30 % games that go well quickly, and my rating increases to 1600 or so. Then I will have to finish the 70 % other games that don't go well, and my rating drops to 900. I end up
- having a too high tournament entry rating (it will never drop below 1500)
- when I am down at 900, I badly hurt the ratings of others (e.g. 1200 - players who I win against)
Should the rating be a more long term average (50 games or so)? It should definitely not be an average of all games, to allow for learning improvement.
Thoughts?
Originally posted by vgunzlertake it site ideas spanky
There may have been previous discussions of this topic, but unfortunately I do not have the time to review all threads.
I think the current way of calculating ratings - gives a distorted readout.
Let's assume I am a player with a "True" rating of 1400. I start 500 games at the same time. I play those 30 % games that go well quickly, and my rating increase ...[text shortened]... uld definitely not be an average of all games, to allow for learning improvement.
Thoughts?
Originally posted by vgunzlerWhat happens afterwards? You start 1000 games?
There may have been previous discussions of this topic, but unfortunately I do not have the time to review all threads.
I think the current way of calculating ratings - gives a distorted readout.
Let's assume I am a player with a "True" rating of 1400. I start 500 games at the same time. I play those 30 % games that go well quickly, and my rating increase ...[text shortened]... uld definitely not be an average of all games, to allow for learning improvement.
Thoughts?
Originally posted by vgunzlerWe should get an extra point for each clever posting and -1 for each stupid posting....
There may have been previous discussions of this topic, but unfortunately I do not have the time to review all threads.
I think the current way of calculating ratings - gives a distorted readout.
Let's assume I am a player with a "True" rating of 1400. I start 500 games at the same time. I play those 30 % games that go well quickly, and my rating increase ...[text shortened]... uld definitely not be an average of all games, to allow for learning improvement.
Thoughts?
'course, my rating would drop to below the mendoza line though
Originally posted by vgunzlera rating doesn't describe the current strength of a player. it describes how the player has performed on average in his PAST games. there's a difference, a big one.
There may have been previous discussions of this topic, but unfortunately I do not have the time to review all threads.
I think the current way of calculating ratings - gives a distorted readout.
Let's assume I am a player with a "True" rating of 1400. I start 500 games at the same time. I play those 30 % games that go well quickly, and my rating increase ...[text shortened]... uld definitely not be an average of all games, to allow for learning improvement.
Thoughts?
- Joined
- 11 Jan 08
- Moves
- 28790
@ wormwood and palynka: Thanks for your comments. I am not sure what you are saying is correct - please check "Help", "FAQs", "How is my rating calculated". I read this as there is NO averaging. The new rating is only based on your own LAST rating, the rating of your opponent, and win/loss/draw. I think this is similar or identical to the way the elo ratings are calculated in OTB chess, but there you don't play these huge number of rated games at the same time.
This is not a stupid idea of mine. There are quite a few real examples like the one I described - players with 500 or more simultaneous games, and players whose rating goes up to 1600 and then crashes to 900. I won't post their names in a public forum, but just look at profiles of players who play a lot of tournaments and you will find them. VG
Originally posted by vgunzlerVery elequent reply to thier comments IMO 🙂
@ wormwood and palynka: Thanks for your comments. I am not sure what you are saying is correct - please check "Help", "FAQs", "How is my rating calculated". I read this as there is NO averaging. The new rating is only based on your own LAST rating, the rating of your opponent, and win/loss/draw. I think this is similar or identical to the way the elo rating ...[text shortened]... ust look at profiles of players who play a lot of tournaments and you will find them. VG
Originally posted by vgunzlerBefore you play someone, view their graph. For the most part, the ratings we have are our real ratings.
@ yo its me - I like your attitude of "Ratings don't interest me" but how are you going to find players at your level, unless you have a reliable rating system? I want my games to be challenging, not easy wins or hopeless losses.
The graph will tell you if they are not.
RHP should have a ratings floor, the idea has come up... but nothing that sticks.
What should my rating floor be? See graph.
What is my average?
How is my rating not my rating?
P-