Uh, It's a crude simulation -- *on-axis and not accounting off-axis and many more complex movements of the head and body. I thought that part was kind of obvious.
Sure, but that is the problem with down-mixing, or better, with mixing in general. In my book it is in-human to expect the listener to sit his head fixed at a certain position all the time. But if she moves, that particular equalisation You gave as an example becomes wrong in that it even worsens things. The same applies to any mix, may it be for still common stereo alone.
The only way out of this inherent mishap with the design of the stereo technology is to allow for tolerances with seating position, orientation of the two ears, ..., room acoustics and more.
The stereo needs the listener to give some input. It is work left to be done on the consumer side! The listener has to actively ignore cues that would due to the intelligence of human's hearing capability, regularly allow for alternative interpretations of the physical sound field.
Actively identify the cues which direct the senses to the intended stereo panorama, actively (!) ignore any other spatial cues that make the speaker identifiable as individual sound sources.
Best means to do so is to listen to the music rather than to concentrate, while listening "critically", on the "excellence" of the very stereo "system".
People with an "audiophile" attempt act paradoxically. They need the speakers to be stolen in order to make them disappear (see somebody else above, thank You