Go back
rating

rating

Only Chess

w

Joined
10 Jan 10
Moves
44077
Clock
02 Apr 10
Vote Up
Vote Down

Is it just me or does my rating always go down twice as fast when I lose as it goes up when I win
I just lost to a player that was 21 points lower than I was and lost 17 points out of my rating. Someone needs to explain this to me.

s

Joined
27 Sep 06
Moves
3441
Clock
02 Apr 10
Vote Up
Vote Down

I've always thought the rating system here was messed up but I've never played enough games to find out.

u
Sharp Edge

Dulling my blade

Joined
11 Dec 09
Moves
14434
Clock
02 Apr 10
Vote Up
Vote Down

the new rating is based on you and your opponents rating at the start of the game. could have been your opponent had a much lower rating when it was started and won some other games

c
Grammar Nazi

Auschwitz

Joined
03 Apr 06
Moves
44348
Clock
02 Apr 10
Vote Up
Vote Down

not true

a

Joined
19 Feb 02
Moves
32077
Clock
02 Apr 10
Vote Up
Vote Down

Originally posted by willpeet
Is it just me or does my rating always go down twice as fast when I lose as it goes up when I win
I just lost to a player that was 21 points lower than I was and lost 17 points out of my rating. Someone needs to explain this to me.
It is just you. This is copied from the FAQ.


Players are rated using the following formula:
New Rating = Old Rating + K * (Score - Win Expectancy)
K is a constant (32 for 0-2099, 24 for 2100-2399, 16 for 2400 and above)
Score is 1 for a win, 0.5 for a draw and 0 for a loss.
The Win Expectancy is calculated using the following formula :
Win Expectancy = 1 / (10^((OpponentRating-YourRating)/400)+1)


So in your case, assuming your opponent is rated the same, it would come down to this:

Win Expectancy = 1 / (10^(0/400) + 1) = 1 / (10^0 + 1) = 0.5
New Rating = Old Rating + 32 * (0 - 0.5) = Old Rating - 16

And when your opppnent is rated lower, than you will lose more than 16. So giving the rating difference of 21 points, do the math yourself, and you will find out that 17 is the right amount.

Adje

o
Art is hard

Joined
21 Jan 07
Moves
12359
Clock
02 Apr 10
Vote Up
Vote Down

The rating that is used is the rating at the end of the game.

Paul Leggett
Chess Librarian

The Stacks

Joined
21 Aug 09
Moves
114051
Clock
02 Apr 10
1 edit
Vote Up
Vote Down

Very roughly, ratings between equally-rated players will go up/down 16 points in a decisive result (winner +16, loser -16).

When there is a rating gap, the variance is plus or minus 4 points per 100 point difference in the players, so if a 1700 player beats a 1600 player, he goes up 12 (16-4) and his opponent goes down 12, but if a 1700 player loses to a 1600 player, he goes down 20 (16+4) and his opponent gains 20.

The max variation is a 400 point gap- if a 1600 player beats a 1200 player, he gets 1 point and the loser loses 1 point, but if the 1600 player loses, he loses 32 points and the 1200 player gains 32.

I have oversimplified, but this is a pretty accurate ballpark approach to figuring it out.

This is why lower-rated players should be absolutely fearless when playing a higher-rated player, if ratings are important to them. They have little to lose and lots to gain.

I also think that if players understood this better, open invites would be approached with a very different attitude.

Paul

NMD

Joined
29 Aug 09
Moves
1574
Clock
02 Apr 10
1 edit
Vote Up
Vote Down

The way I think about the rating is easy and you don`t need a calculator for a formula.

Just think if I lose to someone rated the same ---lose 16 points.
And If I lose to someone rated 25 points higher---lose 15 points.

Just continue thinking 16 points and adjust 1 point every 25 points.

Disclaimer...This may not be as accurate as the formula but it is close and easy to calculate without a calculator.

Oops I see Paul typed the same thing quicker than me by a few minutes.

w

Joined
10 Jan 10
Moves
44077
Clock
02 Apr 10
Vote Up
Vote Down

Thank you. I will keep that formula in my head and may not go into shock every time I win/lose a game. It will be great knowing there is something besides that incredible calculus formula I always am refereed to.

Cookies help us deliver our Services. By using our Services or clicking I agree, you agree to our use of cookies. Learn More.