Depends on what is supposed to be judged.
If we are to judge how good the science behind the model is, and the performance of the climate-modeling, then it is very much reasonable to say that the differing factor between A,B and C developed in accordance with A, and that as such, A is the relevant prediction to judge against the resulting temperature trend.
If we are to judge how well Hansen can predict political actions, sure , then B was the "most probable" from him. But that isn't really relevant to most discussions.