I rather sympathize with list1 on this occasion! If you start from a certain set of assumptions then the option-valuation problem becomes like a certain type of maths, called change of measure. If you had a different set of assumptions and you changed measured willy-nilly then it might not make any sense. So the word 'necessary' in the title of the thread is a bit misleading.
So do you have to change measure? No. You don't need to know anything about this. (And I suspect that list1 does know quite a lot of the theory.)
Does it have any advantages? Yes. Some people find this way of thinking helpful.
Does it have any disadvantages? Yes. If you think every model turns into a change of measure, or if being tied to this concept limits new ideas.