| |
BGonline.org Forums
XGR++?
Posted By: neilkaz In Response To: XGR++? (mtuhtan)
Date: Wednesday, 23 March 2011, at 6:41 p.m.
XGR++ is the high level roller we've been working on for XG2. It is another truncated RO, just like XGR+ and XGR. It is several times slower than XGR+ but, IMHO more accurate. It is probably a couple Elo stronger (difficult to measure since mistakes are few) than XGR+.
On my hexacore I can play matches vs XGR++ at ABT time controls and it won't time out. On your OC'ed i7, it might time out or not. Of course, being a bot, you may wonder sometimes, what the heck it is thinking about.
XGR++ with the new Gigantic SI, will be a great setting for first pass of custom analysis for those with i7 chips for overnight match analysis.
I find that XGR++ is often a little more accurate at appraising the position. Here's a XG2 XGR+ eval for comparison. Of course to make a more accurate comparision, I'd like to quadrupal my XGR+ RO length to halve the CI.
is Player 2
score: 0
pip: 137Unlimited Game
Jacoby Beaverpip: 89
score: 0
is Player 1
XGID=aBBaBbCBB----AA-b--c-b-bb-:1:-1:1:33:0:0:3:0:10 to play 33
1. XG Roller+ 14/11 6/3*(3) eq: +1.148
Player:
Opponent:87.02% (G:46.52% B:0.42%)
12.98% (G:1.76% B:0.04%)Conf.: ± 0.000 (+1.148...+1.148) 2. XG Roller+ 13/10 6/3*(3) eq: +1.143 (-0.005)
Player:
Opponent:86.66% (G:47.10% B:0.45%)
13.34% (G:1.83% B:0.04%)Conf.: ± 0.000 (+1.143...+1.143) 3. XG Roller+ 7/4 6/3*(3) eq: +1.106 (-0.042)
Player:
Opponent:85.83% (G:44.94% B:0.41%)
14.17% (G:1.38% B:0.05%)Conf.: ± 0.000 (+1.106...+1.106) 4. XG Roller+ 13/4 6/3* eq: +1.073 (-0.075)
Player:
Opponent:84.13% (G:47.54% B:0.57%)
15.87% (G:3.41% B:0.07%)Conf.: ± 0.000 (+1.073...+1.073) eXtreme Gammon Version: 2.00.156.alpha
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.