http://universalrating.com
Just announced, it will be used for the next Chess Grand Tour. They make it clear they want to expand beyond just chess. Which seems like a good idea to me. FiveThirtyEight has made some waves applying Elo to basketball and other sports in recent years. Nielsen and FICO make good money from their branded ratings, why shouldn't Glickman and Sonas?
-Carl
"Universal Rating System"
Moderators: hgm, Rebel, chrisw
-
- Posts: 186
- Joined: Fri Oct 10, 2014 10:05 pm
- Location: Berkeley, CA
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: "Universal Rating System"
They seem to use time-dependence (the moment the game takes place) and faster games dependence in some decaying manner in weighting, then proceed iteratively. Interesting, but I wouldn't mix blitz games with classic TC, disregarding the correlation which exists. Being used to engines, we know what different scaling means, and human-computer ratings at Blitz and Classic would look very different at these TC, by hundreds of ELO points.clumma wrote:http://universalrating.com
Just announced, it will be used for the next Chess Grand Tour. They make it clear they want to expand beyond just chess. Which seems like a good idea to me. FiveThirtyEight has made some waves applying Elo to basketball and other sports in recent years. Nielsen and FICO make good money from their branded ratings, why shouldn't Glickman and Sonas?
-Carl
-
- Posts: 4833
- Joined: Sun Aug 10, 2008 3:15 pm
- Location: Philippines
Re: "Universal Rating System"
Major rating factor seems to be the performance rating and they apply weights to it. They can generate rating at a given time frame of say from 1965 to 1970. Attempting too much time difference and it will suffer the connectivity issues as emphasized in Ordo.Laskos wrote:They seem to use time-dependence (the moment the game takes place) and faster games dependence in some decaying manner in weighting, then proceed iteratively. Interesting, but I wouldn't mix blitz games with classic TC, disregarding the correlation which exists. Being used to engines, we know what different scaling means, and human-computer ratings at Blitz and Classic would look very different at these TC, by hundreds of ELO points.clumma wrote:http://universalrating.com
Just announced, it will be used for the next Chess Grand Tour. They make it clear they want to expand beyond just chess. Which seems like a good idea to me. FiveThirtyEight has made some waves applying Elo to basketball and other sports in recent years. Nielsen and FICO make good money from their branded ratings, why shouldn't Glickman and Sonas?
-Carl
I thought it is something like take Sf8 (as reference say 3000 rating points) analyze particular moves of every player in question and generate rating based on score difference of the best score / best move of Sf8 and the score (based from Sf8) of the move played by the player, and also include to factor the final result of the game.
-
- Posts: 186
- Joined: Fri Oct 10, 2014 10:05 pm
- Location: Berkeley, CA
Re: "Universal Rating System"
That method produces so-called "intrinsic ratings", but I do not think URS uses it. It's bound to be something like Glicko.Ferdy wrote:I thought it is something like take Sf8 (as reference say 3000 rating points) analyze particular moves of every player in question and generate rating based on score difference of the best score / best move of Sf8 and the score (based from Sf8) of the move played by the player, and also include to factor the final result of the game.
-Carl