I think it really depends on the initial state of the code, and on who is doing the refactoring.
If this code base was in good shape to start with, this is probably just illustrating that there is a point of diminishing returns on refactoring.
Something about this study feels contrived. The size of the project makes a big difference, and it seems that the study focused primarily on making the code less linear and more abstract (adding polymorphism, etc.) in cases where the overwhelming benefit is seen in huge projects involving many developers. I don't feel that was even close to being the case here.
Dumb study, IMO. Often the value of refactoring is to promote re-use of the code (generally by adding flexibility), but that comes at the cost of some additional complexity. I don't see that any of their metrics measure that. But I guess the lesson of the paper is, don't refactor for the sake of refactoring, which should be obvious.