BGonline.org Forums

Cube action in a volatile position

Posted By: Timothy Chow
Date: Thursday, 20 August 2009, at 3:36 a.m.

In Response To: Cube action in a volatile position (John O'Hagan)

I trust GNU 2-ply rollouts and so does everyone else I know.

This is just an aside, but part of the reason I trust the bots less than many others do is that I've seen people trust bots in other games more than they should. If you study the history of computer chess, Othello, and checkers, you will find that there is a rather large gap between "being able to trounce the best human players" and "being able to play perfectly." Backgammon bots are strong, but not so strong that we can unconditionally trust their verdicts, especially on the particularly tricky positions that are so interesting to readers of this forum. It is important to understand what is happening during the rollout to make sure that the bot is not suffering from one of its known weaknesses.

If you're still not convinced, I recommend reading this Wikipedia article on the long list of "known facts" about chess endgames that were disproved once we acquired the ability to analyze those endgames perfectly.

Messages In This Thread

• Cube action in a volatile position
  Timothy Chow -- Tuesday, 18 August 2009, at 5:00 a.m.
  • Cube action in a volatile position
    Matt Cohn-Geier -- Tuesday, 18 August 2009, at 5:11 a.m.
  • Cube action in a volatile position
    Stick -- Tuesday, 18 August 2009, at 11:45 a.m.
    • Cube action in a volatile position
      Timothy Chow -- Tuesday, 18 August 2009, at 10:49 p.m.
      • Cube action in a volatile position
        Stick -- Tuesday, 18 August 2009, at 10:53 p.m.
    • RHR??
      Chuck Bower -- Wednesday, 19 August 2009, at 4:44 p.m.
  • Cube action in a volatile position
    John O'Hagan -- Wednesday, 19 August 2009, at 4:12 a.m.
    • Cube action in a volatile position
      John O'Hagan -- Wednesday, 19 August 2009, at 5:02 a.m.
    • Cube action in a volatile position
      Timothy Chow -- Thursday, 20 August 2009, at 3:36 a.m.
  • XG Rollout
    Neil Robins -- Wednesday, 19 August 2009, at 5:52 a.m.
    • This doesn't look right...
      Chuck Bower -- Wednesday, 19 August 2009, at 5:08 p.m.
      • What MET used for the XG rollout? *NM*
        neilkaz -- Wednesday, 19 August 2009, at 5:14 p.m.
        • What MET used for the XG rollout?
          Neil Robins -- Wednesday, 19 August 2009, at 8:59 p.m.
        • GNU rollout confirms
          neilkaz -- Wednesday, 19 August 2009, at 11:55 p.m.
        • What MET used for the XG rollout?
          eXtreme Gammon -- Friday, 21 August 2009, at 1:33 a.m.
  • Looking at the rgb post
    neilkaz -- Wednesday, 19 August 2009, at 9:18 p.m.
    • Looking at the rgb post
      Chuck Bower -- Wednesday, 19 August 2009, at 9:31 p.m.
  • 3-ply RO's in the hopper
    Chuck Bower -- Thursday, 20 August 2009, at 12:48 p.m.
  • {3,2} RO
    Chuck Bower -- Thursday, 20 August 2009, at 11:25 p.m.
  • {3,3} RO
    Chuck Bower -- Thursday, 20 August 2009, at 11:28 p.m.
  • 3-ply RO commentary
    Chuck Bower -- Thursday, 20 August 2009, at 11:34 p.m.
    • 3-ply RO commentary
      Matt Cohn-Geier -- Thursday, 20 August 2009, at 11:48 p.m.
      • 3-ply RO commentary
        Chuck Bower -- Friday, 21 August 2009, at 2:06 p.m.
    • 3-ply RO commentary
      Timothy Chow -- Friday, 21 August 2009, at 1:43 a.m.
      • 3-ply RO commentary
        Chuck Bower -- Friday, 21 August 2009, at 2:01 p.m.
        • 3-ply RO commentary
          Timothy Chow -- Saturday, 22 August 2009, at 3:48 a.m.
          • 3-ply RO commentary
            Chuck Bower -- Saturday, 22 August 2009, at 5:59 p.m.
    • 3-ply RO commentary
      Daniel Murphy -- Saturday, 22 August 2009, at 6:18 a.m.