Unfortunately, the design change from the original MCAS to the accident MCAS variant resulted in the hidden trim rate change, where the MCAS rate at high speed was much greater than the trim rate using the normal trim system by the pickle switches. That alone altered the potential impact of a failure, and if it had been evaluated in the sim it is reasonable to expect that the analysis would have increased the categorisation above major.
Not only that, but the designed run time for the MCAS input (9 seconds), far exceeded the assumed pilot reaction time to runaway trim (3 seconds). Yet there was no attempt to educate the pilots that some new system might be trying to activate the trim system for this length of time while at the same time bypassing the control column cutout switches. I realize with the new software that MCAS will only activate if both AOA's indicate an approach to stall, but this still seems very questionable from a design standpoint.