BGonline.org Forums

Number of games to establish two players have equal skill

Posted By: Frank Berger
Date: Saturday, 5 January 2013, at 5:41 p.m.

In Response To: Number of games to establish two players have equal skill (leobueno)

One problem might be that you don't know the standard deviation. For cubeless game it's trivial because you can simply calculate it if you make assumptions about the gammon/bg percentages.

For play with cube you can't do this, and IIRC Zare and another one (Bob Koca? Timothy?) have doubted that it's even defined. (they constructed a scenario what was somehow related to the ST. Petersburg paradox). I believe without standard deviation you can't calculate the confidence interval, but I'm to much of a mathematical layman to make an assertion.

As a pragmatical programmer/economist I am I would estimate the standard deviation from a larger number of games (>40, the more the better) and do a simple estimation with that values.

Messages In This Thread

• Number of games to establish two players have equal skill
  leobueno -- Friday, 4 January 2013, at 11:21 p.m.
  • Number of games to establish two players have equal skill
    Timothy Chow -- Saturday, 5 January 2013, at 12:13 a.m.
  • Number of games to establish two players have equal skill
    Frank Berger -- Saturday, 5 January 2013, at 5:41 p.m.