Group multiple rows in 3 categories for each mont

My data looks like the image below. I have 60 different values for every month  and I'd like to sort them from highest to lowest (by BarCMgMed, for example) and divide them in three groups of 20 rows each, in order to calculate their average. My objective then is to have just three values where each one corresponds to the average of those groups of 20 rows. For example:

A    2019    Ene     Cardones 220    35.06

B    2019    Ene     Cardones 220    40.34

C    2019    Ene     Cardones 220    48.16