
BGonline.org Forums
A programming exercise to address Albert Steg's question
Posted By: Timothy Chow
Date: Monday, 16 May 2016, at 10:35 p.m.
Albert Steg recently asked, "Is all equity equal?"
It occurred to me that one fairly easy way to get an intuitive feeling for these questions is to restrict attention to noncontact races and then experiment with different kinds of nonequilibrium players. For example, the Aggressive Doubler might double whenever the equilibrium ND equity is within 0.100 of the D/T equity. (Similarly one can imagine a Timid Doubler, an Aggressive Taker, a Timid Taker, or various combinations of these tendencies.) It's perhaps harder to come up with a plausible systematic checkerplay bias, but if we don't care about plausibility then there are various possible rules one could devise to generate errors in a systematic way. For example, perhaps the Ace Hater could refuse to bear off a checker with an ace unless it is forced.
Let X and Y denote nonequilibrium players of two different types (perhaps one makes checker errors and the other makes cube errors), and let E denote an equilibrium player. Then one can empirically test claims such as:
If E enjoys the same quantitative edge over X as E enjoys over Y, then X and Y will be equally matched.
Anybody interested in such a programming exercise?

BGonline.org Forums is maintained by Stick with WebBBS 5.12.