Tied notes speed up arpeggios

In some versions of MuseScore it has been the case that an arpeggiated set of quarter notes executes faster than a set of whole notes (though this does not seem to be true in 3.6). This is not what I'm talking about here.

When it comes to tied notes, it would be expected that the arpeggio speed would be predictable based on either the total duration, or the first of the tied notes: i.e. tying two notes of equal duration would either double the arpeggio time or keep it the same. Instead, the opposite happens: tying two notes of equal duration makes the arpeggio execute faster than it would on a single note. This is not the case if only some of the notes in the arpeggio are tied.

Obviously this can be worked around by adjusting time stretch, but it is a bizarre and unintuitive correction to make. Is there some reason this is not a bug? Have I missed any patterns?