One aspect which must never be neglected are the boundaries and edge cases. It is valuable to know where the methods cease to be usable; for example, when noise levels supersede the signal. Having identified the limitations of the method, their robustness can be improved. For instance, in the case of NRR assessment, one can improve the range of displacements where results can be differentiated by increasing the size of the shuffle neighbourhood. Prior experiments showed that the performance is degraded beyond a certain shuffle neighbourhood size, but there are other parameters that can be varied, so their effect on the shuffle neighbourhood is not mutually exclusive.