That doesn't seem fair. The article doesn't say, but in this case it sounds like the theories were based on previous experiments, not on computer simulations. And the deeper, underlying theories are correct. The problem is that we can't efficiently simulate large enough systems to discover phenomena like this computationally. But in the future, with better simulation algorithms, or possibly even large quantum computers, it should be possible to get results like this purely from simulations.

(Of course, I qualified my argument by saying "in the future." So maybe you are right that computational research isn't yet a third pillar of research. But I do think it will become one, as our simulation tools incrementally increase in power.)