BGonline.org Forums

A new variance reduciton technique

Posted By: Bob Koca
Date: Monday, 30 March 2009, at 2:57 p.m.

In Response To: A new variance reduciton technique (MaX)

"But if Dahl's Var.Red. is perfect (i.e. it factors out luck perfectly), this would be unnecessary, right ? "

And rollouts would be also. It would mean that gnubg always got it right things exactly right on 2 - ply so just ask for evaluation.

I think that either method by itself would show a very large improvement over doing nothing fancy. Used in conjunction the extra one gives a small improvement over using just one by itself.

"1. Instead of going for the same rank, you could (at negliectible cost) select the roll for play B that gives the closest luck to the one obtained for play A. Should be more efficent than just taking the roll with the same rank."

Doesn't work as a roll is not randomly chosen. For example suppose that position A has the property that it is extremely unvolatile whereas position B has the property of having a few jokers with everything else slightly below average. Those jokers would never get used.

"2. I'm not sure gnubg uses 2ply for luck evaluation. If it uses 0 ply (as I remember, but I could be wrong, mabe it is 0ply just for match analysis luck evals), introducing your trick would cost something in terms of speed (actually, quite a lot I would say, knowing that each additional ply cost more or less a factor 21). If somebody on gnubg mailing list can confirm that ...."

If so then just use 0-ply to rank.

Messages In This Thread

• A new variance reduciton technique
  Bob Koca -- Sunday, 29 March 2009, at 7:47 p.m.
  • A new variance reduciton technique
    David Rockwell -- Monday, 30 March 2009, at 4:08 a.m.
    • A new variance reduciton technique
      Bob Koca -- Monday, 30 March 2009, at 4:43 a.m.
    • A new variance reduciton technique
      Bob Koca -- Monday, 30 March 2009, at 4:51 a.m.
  • A new variance reduciton technique
    Matt Cohn-Geier -- Monday, 30 March 2009, at 4:56 a.m.
    • A new variance reduciton technique
      Bob Koca -- Monday, 30 March 2009, at 5:34 a.m.
      • A new variance reduciton technique
        Stick -- Monday, 30 March 2009, at 5:38 a.m.
        • A new variance reduciton technique
          Matt Cohn-Geier -- Monday, 30 March 2009, at 5:42 a.m.
  • A new variance reduciton technique
    MaX -- Monday, 30 March 2009, at 2:13 p.m.
    • A new variance reduciton technique
      Bob Koca -- Monday, 30 March 2009, at 2:57 p.m.
      • A new variance reduciton technique
        Maik Stiebler -- Tuesday, 31 March 2009, at 10:21 a.m.
      • A new variance reduciton technique
        MaX -- Tuesday, 31 March 2009, at 11:15 a.m.
  • A new variance reduction technique
    Walter Trice -- Tuesday, 31 March 2009, at 5:37 a.m.
  • I tried it
    Tom Keith -- Wednesday, 1 April 2009, at 1:18 a.m.
    • I tried it
      Bob Koca -- Thursday, 2 April 2009, at 3:15 p.m.
      • I tried it
        Tom Keith -- Thursday, 2 April 2009, at 6:37 p.m.