Rankings

Sort:
Oldest
doc48

G'day,

Wish to ask how rankings are made and what may be the average.

Thought I found a grapth but now can't find it again.

Thanks

doc

TadDude

Ratings not rankings. Graph - Online Chess | View Players.

 

 

Advertising: Exclusive Offset 701-1100: 100 games completed.

Dexter_Morgan

This is probably the graph that you found which shows the range of ratings at this site and the average :

http://www.chess.com/echess/players.html

As for the way the rating is calculated, well it is called the Glicko System.  I don't have the whole mathematic formula behind it at hand but I'm sure somebody else can provide it.  Basically though, the more games you play, the more the system assumes your rating is accurate, and the less your rating will spike and plummet from individual wins/losses.  If you play infrequently, then your rating will fluctuate a lot more as then the system feels your rating is not accurate.  I'm sure others can and will elaborate further on this.

You'll start off at a base rating of 1200 and go from there.  The higher the better of course.

 

Edit : Oops, I see somebody got to answering this question a minute before me.

rooperi

Hmm, a classic bell curve....

I'd be very interested to see the percentage of draws by rating, 3 and a bit percent seems very low, I'm sure its directly proportional to rating...

Dexter_Morgan
rooperi wrote:

Hmm, a classic bell curve....

I'd be very interested to see the percentage of draws by rating, 3 and a bit percent seems very low, I'm sure its directly proportional to rating...


Yup, the higher ups can be a bunch of draw buddies sometimes.

doc48

G'day,

Thanks for the info. Fellars. You've cleared that up for me.

doc

DrBobR

I have a general question about ranking:

I see: Today's Rank: #28080 of 71,250 (60.6%) it seems to me this calculation is backwards. What am I missing? It seems like a person ranked 10,000 of 71,250 would be 14%, 20,000 ... 28%, then 28080...39+%  ... or is the ranking a reverse pyramid Undecided

Nytik

It is reversed, DrBobR. Rank #1 of 71,250 would be 100%.

ichabod801
rooperi wrote:

Hmm, a classic bell curve....


Obscure piece of rating trivia: it's actually not a classic bell curve. It's a logistic distribution, not a normal distribution. Or at least, the logistic provides a better approximation of the actual distribution of chess players.

DOC-HOLLIDAY

http://www.chess.com/article/view/chess-ratings---how-they-work

Forums
Forum Legend
Following
New Comments
Locked Topic
Pinned Topic