BGonline.org Forums

XG2 MET Q for Neil

Posted By: neilkaz
Date: Sunday, 24 April 2011, at 5:03 p.m.

In Response To: XG2 MET Q for Neil (eXtreme Gammon)

I rolled out each score 100,000 times. I used 3 ply play/XGR cube noting that XGR cube is even a bit stronger than 4 ply cube and certainly faster. However, for the post-crawford even scores, I stepped it up to 3 ply ply/XGR+ cube since at -1-2PC XGR+ cube was doing a better job of playing on for a G than was XGR cube.

I stopped this work after 9 pts since it was so close to R/K that I felt simply using R/K was fine. I also checked a few of my extrapolations from R/K (rolled to 15 pts) to higher scores and was quite happy with the results.

Maybe at some point, I'll extend to a longer match, or maybe I'll roll a few critical early scores out to 400k to halve the CI, but if I do I won't be going past a 4 or 5 pt match. If I do any more RO's for this, the updated MET will be included with XG2 release or made available for users. However, no human can evaluate anything except maybe a very short bearoff to the accuracy where it matters whether the ME differs by +/- .03% or so anyhow.

.. neilkaz ..

Messages In This Thread

• XG2 MET Q for Neil
  Matt Cohn-Geier -- Sunday, 24 April 2011, at 9:28 a.m.
  • XG2 MET Q for Neil
    eXtreme Gammon -- Sunday, 24 April 2011, at 3:01 p.m.
    • XG2 MET Q for Neil
      neilkaz -- Sunday, 24 April 2011, at 5:03 p.m.