Inconsistent future_t for different models in expected_probability_alive

That makes it more clear.
Thanks for the response, @ColtAllen !