[phenixbb] Can simulated annealing remove model bias after MR?