NOTES
If you run the same experiment a million times, the final answers are distributed equally across all the possibilities. In each case, there will be a settling-in period, and the process looks like it's going to a predetermined answer.
I should note that I read about this version of the thought experiment in a science magazine about 10 years ago - and I haven't been able to find the original reference; the rule reversal version is original to me as far as I know.