Several people (Tom Helmuth, Bill Langdon, etc.) suggested that we look at the difference in the fitness of parents and see if that helps explain why bias does or doesn't work. The idea is that if there are larger differences between the fitness of parents (likely true for binary tournaments) then XO Bias would have a bigger impact than when the differences were smaller (likely true for larger tournaments). I collected some data on this but haven't looked at it, but it could be really important for explaining things.
After a whole pile of false starts and mistakes, I think we finally got some reasonable plots that support the idea that larger tournaments + elitism give us smaller differences in parent fitness.
Several people (Tom Helmuth, Bill Langdon, etc.) suggested that we look at the difference in the fitness of parents and see if that helps explain why bias does or doesn't work. The idea is that if there are larger differences between the fitness of parents (likely true for binary tournaments) then XO Bias would have a bigger impact than when the differences were smaller (likely true for larger tournaments). I collected some data on this but haven't looked at it, but it could be really important for explaining things.