Both tasks are made up of two filters being applied to your keywords in sequence. One of those, for Value Demand, is exclusive [highest 60% vs lowest 30%], while the other, for Real Supply, is not [range between 20% and 80% vs highest 50%].
The order in which the filters are applied is important, that is, applying them in a different order will result in a different result set. With the filter on VD applied first, you'd get different results for both tasks, as this first step would produce sets with keywords exclusive to each task.
Instead, it seems that the filter on RS is applied first. This leaves many keywords in both result sets after the first step [those with RS between 50% and 80%]. Since VD score is independent from RS for the same keyword, the next step won't make results mutually exclusive.
TLDR: The order in which the filters are applied makes the difference. With VD filter applied first, results wouldn't contain any matching keywords.
You can try it yourself in MKL by clicking on Filter to make the conditions visible, and changing them in sequence.