cells evaluation during a mace

Hi All,
I have two environments and when I run the mace of a rule group, in the first one the execution takes few minutes, in the other environment the excecution takes 30 minutes.
I notice that, in the log, for each rule, the number of evaluated cells in the first environment is less than in the second environment.
Without taking in consideration the number of the cells, why in the first environment rpas evaluates 0.% of cells and in the second environment rpas evaluates 50% ?
I notice that the navalue parameter in the log is different, "navalue = [true]" in the first and "navalue = [false] in the second". What is this parameter in the cells evaluation and how is it calculated ?