21208_AndNowForSomethingCompletelyDifferent!

This showed that Bart had done 3 out of 4 chores correctly and Lisa had done 2 out of 3 chores correctly. Then for a second month, Bart and Lisa did chores and the record keeping looked like this:

The father made the following chart to support the hypothesis that Bart was doing the most chores:

Lisa looked at the tables then said "No no no, that's not right, I've done more." Bart pointed at the statistics and said, "Numbers don't lie." Lisa said, "But look, I've done 5 chores and you've only done 4. You just look good because you always run away before somebody asks you to do something. The total should give me 55% of the total chores done right." Bart laughed and ran away. Now Lisa needs your help to show the unfairness of these chore evaluations. She wants a program which computes the averages over the entire dataset and displays where the before and after averages support the opposite hypothesis than would be indicated by considering the total dataset. To be fair, she only wants to consider record keeping times when both she and Bart have chores assigned both before and after the record keeping.

where BBefore is Bart's correctly performed chore percentage before the record keeping, LBefore is Lisa's correctly performed chore percentage before the recordkeeping, BAfter is Bart's correctly performed chore percentage after the first record keeping until the end of the dataset, and LAfter is Lisa's correctly performed chore percentage after the first record keeping until the end of the dataset. Note that chores are numbered from 1 to N in order of appearance in the dataset. All percentages must be rounded to the nearest .0001. Print a blank line between the outputs for two consecutive test cases.

Clarification: Because Lisa wants a program which “displays where the before and after averages support the opposite hypothesis than would be indicated by considering the total dataset”, then any averages which, after rounding, are printed as equal must be disregarded.