From my own perspective, and many of my friends, the MX1 test was really not hard, maybe the slightest amount harder, if not the same difficulty level of the previous two years (I didn't do the 2019 paper). However, MX2 was definitely hard, more specifically the last few questions. Without the jump in difficulty in MX2, I would say it wouldn't really be that controversial. Looking at my friends' reactions, I saw that people who did MX2 didn't really feel the difficulty in MX1, and those who did MX1 didn't feel the difficulty in advance. Even though the content is different, the mathematical skill/ability does reflect in which course they take. I don't believe NESA has the obligation to make an easy test, and a curveball when people least expect it is what's going to differentiate people who try to grind for the test, and not gain a mathematical understanding across the whole syllabus for a well-rounded skillset in maths. There are people that are like, "there wasn't much integration, and that's unfair", but integration is objectively the easiest to do in MX1, where you can just do thousands of questions or just follow a method, and eventually, you will get to an answer (i.e. 'u' sub, t-formula, trig identities, or just use the data sheet (f'(x) and f(x) reversal shiz) (Try one of these and you will find the correct path eventually)). People just don't give stats that much attention (I was guilty myself) and so it felt harder to a lot of people, but that's only a subjective thing, and not really something you can say is a determinant of the scaling that will occur for the MX1 test.
--> So when the scaling is not as high as people expected, I wouldn't be surprised. I mean, it really wasn't that much harder in an objective sense.