[resampling] unbiasedness of ge estimator

[ludwigbothmann]

It should be clearer what is meant if we say the estimator of ge is unbiased. This is wrt to a specific data set size because otherwise, it is pessimistically biased. (And "unbiased" rather says that we got rid of the optimistic bias). I would suggest copying content from the tuning paper since it is quite nicely described there.