This paper, and its conclusion seems remarkably premature.
The study starts with the assertion "The statistics of extrasolar planetary systems indicate that the default mode of planet formation generates planets with orbital periods shorter than 100 days and masses substantially exceeding that of the Earth. "
That's a pretty substantial, definitive statement right there. Yet, two of the very basic rules of statistical analysis are collection bias and contextual sample size.
In reverse order, then, first we're looking at a TEENY sample of systems, given the potential population. This is quite literally, taking a drop of water from the ocean and drawing massive, systemic conclusions therefrom. Now, one could even perhaps make such conclusions decisively from a small sample if one could be sure that one's sample was representative - leading to a Godel's Theorem sort of problem in which you can't know enough about your sample to be sure that it's representative without knowing more about the context, which implies a larger sample anyway. Our testing methods allow reasonably certain detection (of a minimum size, more on that later) to what, the nearest 40,000 stellar candidates? That's 4/100,000ths of 1%. Do we know that our system is 'typical'? Do we know our stellar neighborhood, the Orion arm, or even the Milky Way is 'typical'? Without knowing that, we can hardly be categorical that our pinprick of data is at all representative of everything else, even relatively nearby.
Second, and I believe more important, is selection bias. We have a number of different methods to detect planets today, but I think it's relatively accepted that we're still in the very early stages, where our methods are - at best - looking through a glass darkly. If your methods can only detect relatively major stellar motions (requiring massive perturbing bodies) or gross dips in luminosity (consequential of substantial occlusion taking place) then logically all the samples you'll get are large. It's like casting a 1" crid net into the ocean, and concluding that none of the fish out there are less than 1" long