Which rating to use?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Re: Which rating to use?

    Tom, I think you are on to a very good idea. I'd like to see a separate discussion along those lines - how to use the rating/results system to recognize players' accomplishments.
    Paul Leblanc
    Treasurer Chess Foundation of Canada

    Comment


    • #32
      Re: Which rating to use?

      I'll second the value of the new USCF system. I live in the US and was quite pleased when I discovered their newly-implemented norm system a few days ago. I have one "1st category norm", and my (lowly) chess goal now is to earn four more!

      As to which rating to use, CFC or FIDE or both: my view is "none of the above" if no effort is made to take into account the +/- error associated with ratings. Ratings are *estimates* of strength, where even for established players true strength could be 30 or more points above or below current rating. The SSDF rating list (for chess programs) gets this right, as they include estimates of rating error. On the current SSDF list Deep Rybka 3 has a rating of 3227 +/- about 26 points after playing 1005 games. For people, who unlike machines are subject to emotions/sleep deprivation/other distractions, Jeff Sonas of chessmetrics.com has estimated that FIDE ratings have a standard deviation of 50 points.

      If the goal is to select the strongest team, it is not possible to do so by mechanically selecting players based on rating without some regard to the error inherent in those ratings. The current system creates the illusion of a completely objective selection procedure, but it isn't. Because adding two ratings together makes the +/- error even larger, this strikes me as the worst of the three choices.

      Being the current Canadian champion is a pretty unambiguous selection criterion. As is having a rating much higher than everyone else. Once ratings +/- error begin to overlap, though, the process of identifying the "strongest" players inevitably becomes rather grey.

      Comment


      • #33
        Re: Which rating to use?

        bump up.

        (The latest news from the FIDE rating committee - no technical difficulties to lower the base rating to 1000 and do monthly ratings :)

        Comment

        Working...
        X