That's a perfectly valid idea! But the thing is: there is no ultimate "right" or "wrong". :-) MIR 3D is "just" an empty room which lends itself to any type of instrument seating and many different recording approaches.
A source position that might cause a phase problem with one microphone setup (e.g. a so-called "Blumlein pair" of two crossed Figure-8 capsules) might sound perfectly fine for the same source as soon as the Main Microphone is used in form of an M/S array with an omni-capsule for the M(id)-component. Similarly, it will make a difference whether the source's stereo width is reduced to mono, or increased to 10 or 12 meters in case of a huge ensemble. ... all of these classical coincident stereo techniques can be modeled easily within MIR 3D, all kinds of source positions, rotations and sizes will have and influence, too.
Our options get even more diverse as soon as we leave the domain of 1st Order Ambisonics and proceed to the world of Higher Order Ambisonics. Here it makes more sense to map the source positions to a spatial "sphere" than to think in terms of simple virtual microphone capsules. Consequently, possible phasing problems will most likely occur with a completely different set of parameter choices.
Finally, all of these points are moot when mixing in surround or 3D. While phasing between individual speakers is still possible on a technical level, it will be much less of an issue due to the obvious fact that all the other speakers will counteract any audible problems.
Bottom line: While I absolutely understand your quest for "once and for all" solutions, it rather makes sense to think through some basic audio engineering aspects to exploit MIR 3D to its full extent. There are full-blown, tried-and-tested Venue Presets with "Roles" for 450+ sources available which can serve as staring points for your own creations. And I'm happy to tell you that many, many Venue Presets for "empty" stages are about to be released in an update in the near future. :-)
Kind regards,