The "older generation" tools only allow you to calculate with ranges that are constructed by hand-ranking, instead of actually picking the best hands based on the situation. (This makes the calculations much easier to carry out.)
If you assign a Push range to be 20% in Wiz, those 20% will always be the same hands based on some hand ranking. (I believe it uses Sklansky-Chubukov as ranking? Not exactly sure though.) Those are usually *not* the best 20% of hands for the situation.
The consequence for users is that they become used to seeing ranges of a fixed structure, independent on the situation. And using a fixed structure like this can be a considerable mistake, as you see in the example you posted.
If you select a 20% range in HRC it uses the best performing 20% as a default, based on the situation. (You still have the option to add/remove individual hands.) This helps users to recognize that the structure of optimal ranges depends on the situation.

