BGonline.org Forums

ELO

Posted By: Bob Koca
Date: Friday, 18 December 2009, at 7:02 p.m.

In Response To: ELO (Jake Jacobs)

"You would want to use only matches between players with enough games for the rating to have had enough time to reach its true value."

I'm not so sure. If you have a large enough sample size it ought to take care of itself. One of the questions I posed was whether a given difference, say 200 points, turned up with different win percentages at different levels.

---

For players with little experience there is no reason to expect that the Elo formula predicts win chance accurately. Such games should not be included in the data (for what I think you want).

Messages In This Thread

• ELO
  Jake Jacobs -- Wednesday, 16 December 2009, at 6:39 p.m.
  • ELO
    Bob Koca -- Wednesday, 16 December 2009, at 7:24 p.m.
    • ELO
      Jake Jacobs -- Friday, 18 December 2009, at 5:24 p.m.
      • ELO
        Bob Koca -- Friday, 18 December 2009, at 7:02 p.m.
  • ELO
    Rich Munitz -- Wednesday, 16 December 2009, at 7:35 p.m.
    • ELO
      Bob Koca -- Wednesday, 16 December 2009, at 7:43 p.m.
      • ELO
        Rich Munitz -- Wednesday, 16 December 2009, at 9:25 p.m.
        • ELO
          Bob Koca -- Wednesday, 16 December 2009, at 11:14 p.m.
          • ELO
            Matt Cohn-Geier -- Thursday, 17 December 2009, at 7:17 p.m.
    • ELO
      Bob Koca -- Wednesday, 16 December 2009, at 8:16 p.m.
  • Elo, please
    garyo -- Thursday, 17 December 2009, at 7:00 a.m.
  • ELO
    Christian Plenz -- Sunday, 20 December 2009, at 6:42 a.m.
    • ELO
      Rich Munitz -- Sunday, 20 December 2009, at 7:14 a.m.