Particularly since the confounding issue is really enormous in science. And sits at the core of the example the article gives in introduction... That would be an achievement in itself to build an experiment without confounder(s).