Every journal has strict rules about duplicate submissions, and I do not mean publishing the paper in 2 venues.
My approach is mostly from an economics point of view: Given the fact that most reviews take between 2-3 months, plus any number of months for a resubmission, limiting yourself to only one journal seems not only a waste of time for the individual, but an overall drag for the scientific community. Even if you make a groundbreaking discovery, it won't get published until about a year later.
Multiple submissions dramatically increase the chances of getting your work published, and if you get accepted in both, you can always pull out your paper. From a marketplace point of view, this makes sense, since in this way, journals would be fighting for authors and not the other way around.
Models like Arxiv have proven that this is not such a crazy idea