ENG Rapid Ratings in Swiss Manager

General discussions about ratings.
Roger de Coverly
Posts: 21301
Joined: Tue Apr 15, 2008 2:51 pm

Re: ENG Standard play 4-digit ratings

Post by Roger de Coverly » Tue Aug 16, 2022 12:50 pm

Peter Harrington wrote:
Tue Aug 16, 2022 12:23 pm
And this is what I was expecting to see, and 1563 is much better than 1477.
It isn't really that much at less than 100 points. You can do it directly using what's sometimes called the 400 point rule. Add up the ratings of your opponents, add or subtract 400 times your excess or otherwise of wins over losses, divide by the game count. If the method sounds familiar it's because the old Otto Clarke system is a linear approximation to Professor Elo's version which used probability tables.

The Elo performance measure is that rating which gives an expected score of what you achieved against the opposition you faced. Another point is that the figures shown by the chess-results sofware aren't always reliable. Better to see what the ecfrating system comes up with. Incidently that could be accused of slacking in not yet having the results of the completed tournaments, but perhaps they are waiting for the Congress as a whole to complete.

(edit) There may be a simpler explanatuion. The chess-results calculations may be using the International ratings, and not all your opponents had them and in any event they are mostly different from the max(Elo,English) used in the pairings. (/edit)

Ian Thompson
Posts: 3551
Joined: Wed Jul 02, 2008 4:31 pm
Location: Awbridge, Hampshire

Re: ENG Standard play 4-digit ratings

Post by Ian Thompson » Tue Aug 16, 2022 1:00 pm

Peter Harrington wrote:
Tue Aug 16, 2022 12:23 pm
The new 4-digit rating system appears to undervalue player ratings. For example at the current British Championships in Torquay, I played in the week 1 U1750 event. My grade going into the event was 1531 (or 111 in 'old money'). I was disappointed and confused by the system-generated performance rating of 1477.
I assume you got that number from chess-results. The accuracy of its TPRs have been questioned before.

chess-results has got nothing to do with the new ECF rating system so you need to wait until the ECF rate your games and see what it shows. Presumably you expect a small increase in your rating if your TPR was slightly more than your current rating.

Peter Harrington
Posts: 15
Joined: Sat Apr 28, 2018 7:41 pm

Re: ENG Rapid Ratings in Swiss Manager

Post by Peter Harrington » Tue Aug 16, 2022 1:10 pm

Yes, I did get the number from chess-results. That's right I was expecting a small increase on my current rating. Thanks for your comments.

Ian Thompson
Posts: 3551
Joined: Wed Jul 02, 2008 4:31 pm
Location: Awbridge, Hampshire

Re: ENG Standard play 4-digit ratings

Post by Ian Thompson » Tue Aug 16, 2022 1:25 pm

Roger de Coverly wrote:
Tue Aug 16, 2022 12:50 pm
Better to see what the ecfrating system comes up with. Incidently that could be accused of slacking in not yet having the results of the completed tournaments, but perhaps they are waiting for the Congress as a whole to complete.
The results of all the events completed on 10 August have been submitted to the rating database - see https://www.ecfrating.org.uk/v2/new/list_events.php.

Peter Harrington
Posts: 15
Joined: Sat Apr 28, 2018 7:41 pm

Re: ENG Standard play 4-digit ratings

Post by Peter Harrington » Tue Aug 16, 2022 1:36 pm

Roger de Coverly wrote:
Tue Aug 16, 2022 12:50 pm
Peter Harrington wrote:
Tue Aug 16, 2022 12:23 pm
And this is what I was expecting to see, and 1563 is much better than 1477.
It isn't really that much at less than 100 points. You can do it directly using what's sometimes called the 400 point rule. Add up the ratings of your opponents, add or subtract 400 times your excess or otherwise of wins over losses, divide by the game count. If the method sounds familiar it's because the old Otto Clarke system is a linear approximation to Professor Elo's version which used probability tables.

The Elo performance measure is that rating which gives an expected score of what you achieved against the opposition you faced. Another point is that the figures shown by the chess-results sofware aren't always reliable. Better to see what the ecfrating system comes up with. Incidently that could be accused of slacking in not yet having the results of the completed tournaments, but perhaps they are waiting for the Congress as a whole to complete.

(edit) There may be a simpler explanatuion. The chess-results calculations may be using the International ratings, and not all your opponents had them and in any event they are mostly different from the max(Elo,English) used in the pairings. (/edit)
Thanks, I will remember the 400 point rule - very helpful.

Peter Harrington
Posts: 15
Joined: Sat Apr 28, 2018 7:41 pm

Re: ENG Rapid Ratings in Swiss Manager

Post by Peter Harrington » Tue Aug 16, 2022 4:47 pm

Thanks for your comment.

Post Reply