| |
BGonline.org Forums
New GNU Strength Some More Data
Posted By: Chuck Bower In Response To: New GNU Strength Some More Data (Michael Depreli)
Date: Saturday, 6 April 2013, at 3:02 p.m.
Using the default 2-ply filtering of 8 moves within 0.16 for NewGnu the correct move doesn't make it past 0-ply but when forced to analyze it at 4-ply it finds the correct play. XGR would have missed this too at the same filter settings but XGR++ lowest setting (Normal) has an initial filter of 16 moves within 0.32.
You seem to be comparing apples to oranges in this paragraph. Isn't it more relevant to compare GNU-bg evaluations with XG evaluations? True, XG allows a human to play against an opponent using its mini-rollout settings but that doesn't translate to strength of engine, which I assume is what you are more concerned with.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.