| |
BGonline.org Forums
Some bot rollouts of -2-1C ..obtaining the ME at GG
Posted By: neilkaz
Date: Sunday, 4 October 2009, at 2:30 a.m.
Ian Dunstan's rollouts with GNU 2 ply give the trailer 32.23% ME.
David Rockwell's ~40K rollouts with GNU 2 ply Supremo give the trailer 32.3112%
My ~26K rollout with a slightly expanded GNU 3 ply search space give the trail 31.15%
My 30 x 5184 game rollouts with XG 3 ply compiled to give the trailer 32.3235% ME.
One thing is certain and that is that the ME at this score certainly isn't 30%.
We all like to think, and it still may be true in real life that most humans don't get as many G's as bots, but it is also true that the trailer is a game favorite here since he can stay back for ever to wait for hits.
If I were going to calculate a new MET (and since I'd be doing it by hand and would want to keep it reasonably simple) I'd do the following.
70% cube efficiency
The leader at -1-2P has 51.2% (then I'd make a stab at other free drop ME's)
Each side wins 36% singles, 13.5% G's and 0.5% BG's at all scores.
The G rate is what I use OTB for -2-1C where I use 32% vs an equiv. opp. The G rate may be tad high later on, but this should be close enough to balanced by the BG rate being a tad low.
I won't be doing said calculations since I am working with David Rockwell on rolling out a 15 pt MET using GNU 2 ply Supremo. Those with a MET generator program feel free to do so, but I stress NOT to use 100% cube efficiency and that my work a few years ago indicates that 70% is about right on average.
Of course I expect the rolled out MET to be somewhat more accurate !
.. neilkaz ..
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.