If we're going to the level of QM, then no. Because QM is inherently a random process. Like radioactive decay is a great way to make true random number generators. The real world has truly random events, and while it bothers many people, physicists don't generally argue about this point (anymore).
Maybe if you had a perfectly homogeneous air density (which you'd never see, even in a really good vacuum and near absolute zero temperature. You can thank QM for that), that constantly stayed homogeneous. Problem in real world conditions is that nothing is perfectly homogeneous (entropy exists), so things will heat up unevenly, certain parts will gain more velocity than others, and just a small offset can create a large change in outcome. Short answer is that these systems are chaotic in nature, so they are not stable.
But that doesn't have anything to do with simulations (except what you are trying to emulate). As far as simulations, numerical accuracy will play a role, but really what you look for is if it is realistic, because the real world has random events. That's more what I was trying to get at. You want something that represents reality, not an overly simplified example that you can't use in a meaningful way. Even with these inaccuracies you can get representative models (they will reflect what happens in a physical experiment). And I say representative, because you aren't going to account for all those factors in a simulation, but you are accurate enough to make extremely effective conclusions. I'll even note that some people will add random noise into their simulations (I don't know if this author did, but numerics can play that role).
Maybe if you had a perfectly homogeneous air density (which you'd never see, even in a really good vacuum and near absolute zero temperature. You can thank QM for that), that constantly stayed homogeneous. Problem in real world conditions is that nothing is perfectly homogeneous (entropy exists), so things will heat up unevenly, certain parts will gain more velocity than others, and just a small offset can create a large change in outcome. Short answer is that these systems are chaotic in nature, so they are not stable.
But that doesn't have anything to do with simulations (except what you are trying to emulate). As far as simulations, numerical accuracy will play a role, but really what you look for is if it is realistic, because the real world has random events. That's more what I was trying to get at. You want something that represents reality, not an overly simplified example that you can't use in a meaningful way. Even with these inaccuracies you can get representative models (they will reflect what happens in a physical experiment). And I say representative, because you aren't going to account for all those factors in a simulation, but you are accurate enough to make extremely effective conclusions. I'll even note that some people will add random noise into their simulations (I don't know if this author did, but numerics can play that role).