Thanks. Cached thoughts seem applicable, but also too broad for what I'm describing. After all, if I failed to update on A and B exiting on -2, and continued thinking C may get out either on -3 or -4, that could also be described as a cached thought which I retained even when new evidence contradicted it. But I didn't do that, and was in no danger on doing that. I think that it's the necessity to roll back to the previous state, rather than just, in general, update on new evidence and get rid of the cached thought, that seems important here.