| |
BGonline.org Forums
Better tools, less disinformation
Posted By: Rich Munitz In Response To: Better tools, less disinformation (mtuhtan)
Date: Monday, 1 November 2010, at 4:44 p.m.
I'm not saying any of today's rollouts are "correct", I'm simply saying that they are in all likelihood small errors. Stick once said that he would not be surprised if 64R turned out to be correct on the opening roll. So what if it is? How much of a mistake will it be if anyone plays 64P or 64S? Is there some great truth we are all missing because today's bots have major gaps in knowledge? Sure there are positions that arise infrequently that bots might not have a good grasp on. But the key word is "infrequently". My point is simply that I do not believe that there is another revolution of truth out there.
Imagine that we had a backgammon oracle that could tell is the exact equity of any position. What would be the average PR of XGR+ in matches against itself? Maybe 1.5? And how much better than that would rollouts be? So how much error are we really making here when we're willing to invest the computing time? How much room for improvement in the tools (which includes the hardware they run on) is there really left?
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.