Recent studies have shown that research by small teams is more likely to lead to disruptive results than research by large teams. Disruptive research challenges established paradigms. This paper offers a possible theory to explain this paradox. We argue that individuals in possession of research ideas with great disruptive potential have incentives to form small teams and compensate potential group weaknesses with a greater research effort rather than considering additional co-authors. Additional co-authors have the advantage of bringing more overall effort and expertise to the team, reducing technical difficulties, and increasing the chances of success and the potential value of the ideas. We show that individuals in possession of potentially disruptive research ideas prefer to keep teams as small as possible, because the resulting credits per co-author decrease as the value of the project is split among more co-authors.