This paper deals with some statistical methods for the comparison of multivariate time series of arbitrary dimensions, with particular attention to the SMETS method. As regards the application in the financial field, the case of missing data in the historical series is first dealt with, then the use of the multi-scale permutation entropy is presented. Finally, it ends with a quick methodological comparison on how to treat time series of different lengths, in particular reference is made to the spectral domain method.