Results show that increasing the standard deviation by 20% implies assigning only one bike more in station E with an increased cost of 2.81% (511.45 instead of 497), while decreasing the standard deviation by 20% implies assigning only one bike less in station H with a decreased cost of 2.75% (483.30 instead of 497). From the test we can conclude that if we know the support and the probability distribution of the rental demand, it is not so crucial to have an exact estimate of the standard deviation to find a good fleet size.