Ratings
Hello Yotta, could you please explain what rating system is in use on CC. Thanks.
Hi Corvidae,
ChessConquest uses the ELO rating system using a K-factor of 32.
New players receive a starting rating of 1000 when they sign up for the first time.
Below is an example of how the rating is calculated. For those new to computer science, ^ indicates "to the power of", and * indicates "multiplication".
R1 = 1200 (White player)
R2 = 1100 (Black player)
K = 32
Black wins:
Score1 = 0
Score2 = 1
Step 1:
transR1 = 10^(R1/400) = 1000
transR2 = 10^(R2/400) = 562.341
Step 2:
expected1 = transR1 / (transR1 + transR2) = 0.640
expected2 = transR2 / (transR1 + transR2) = 0.360
Step 3:
New R1 = R1 + K * (Score1 - expected1) = 1180
New R2 = R2 + K * (Score2 - expected2) = 1120
ChessConquest uses the ELO rating system using a K-factor of 32.
New players receive a starting rating of 1000 when they sign up for the first time.
Below is an example of how the rating is calculated. For those new to computer science, ^ indicates "to the power of", and * indicates "multiplication".
R1 = 1200 (White player)
R2 = 1100 (Black player)
K = 32
Black wins:
Score1 = 0
Score2 = 1
Step 1:
transR1 = 10^(R1/400) = 1000
transR2 = 10^(R2/400) = 562.341
Step 2:
expected1 = transR1 / (transR1 + transR2) = 0.640
expected2 = transR2 / (transR1 + transR2) = 0.360
Step 3:
New R1 = R1 + K * (Score1 - expected1) = 1180
New R2 = R2 + K * (Score2 - expected2) = 1120
Thanks, I got most of that, To save using up all my grey cells, could you answer another rating query please. Smith (1200) and Jones (1100) are playing two games. Smith as White is about to mate on next move. Jones is not far off from winning the second game (as White) and Smith is thinking of resigning. Should Smith win the first game before resigning the second, or vice versa in order to get a better rating, or doesn't it matter?
I would resign before mating. I think the rating system should be simplified as much as possible.
After doing some quick back-of-the-envelope math... the results seems to be:
Scenario 1: Smith wins first, and then resigns second game.
The new ratings will be:
Smith:1191
Jones:1109
Scenario 1: Smith resigns second game, then wins the first game.
Smith:1193
Jones:1107
Note that I did do some rounding to save time, but in my opinion it probably won't make a big difference in the long term.
Scenario 1: Smith wins first, and then resigns second game.
The new ratings will be:
Smith:1191
Jones:1109
Scenario 1: Smith resigns second game, then wins the first game.
Smith:1193
Jones:1107
Note that I did do some rounding to save time, but in my opinion it probably won't make a big difference in the long term.
I've often wondered that because it's quite annoying some times to drop a couple of points, but it's good to see how you've worked it. Many thanks.
With a relative small number of players and a small number of games the ratings can be unreliable. I've come to realise that you can't take too much notice of ratings.
I play on many servers. On some of them I'm terribly underrated because of timeouts. Now I have reduced the number of servers I play on and am reducing the number of games on every server I play. Once I'll back to the tops at least to the most of them.
I'm with you on that one. I had the same experience. When I started I played ten games and lost them all. Didn't do much for my rating. Now I play for the love of the game. No chess engines for me.
Our top ranked player Chinchin120 has a FIDE active rating over 2000.
Yes, I’m still hoping for the day where I can win a game against him. ;-p
He now has zero active games. Hope he's not gone for good. We're stuck on 39 players for a while.