Possible issue with `mbar-block` and `fractions`

[slochower]

It did not seem like mbar-block was correctly analyzing fractions of the data. Revisit on simple system.

[wwilla7]

In my most recent test, failures occurred intest_mbar_block and test_ti_block, as AssertionError. Do you think they are related to this possible issue?

[slochower]

No, this issue is about analyzing only pieces of the total simulation data (for example, say you want to analyze the first half and second half of a trajectory separately). The issue you faced, as I think we discovered, was due to the random number seed being reset when running multiple free energy methods at once.