Closed NicMcPhee closed 9 years ago
The early results on, for example, the Ordertree problem suggests that inverse bias is much the same as no bias at all. That seems slightly odd to me (I was expecting it to be worse), but my guess is that we may want to step away from this in order to keep things from getting even more complex than they already are.
Inverse bias is part of the "standard" set of options at the moment, so I'm going to mark this as "done" in the sense that we're collecting data on it.
In my early runs I did lots of values for "positive" crossover bias, but never included inverse/negative crossover bias, and I should just to see what that does.