If I calculate a separate result for each of a number of individual conditional simulations having used a normal score transform on the sample input data and a back transform on each individual simulation in order to make my calculation - and then form a distribution from the indivdual calculations - could anyone tell me why the mean of that distribution will differ markedly from the mean of the same calculation carried out on the output from the back transform of all the simulations taken together - and which result is preferred?

Since I had no response for this - is there nobody out there versed in geostatistics or is my question too dumb?