Decision theory has been long applied to project management for risk and uncertainty reduction. Among the foundations, the manager is considered following axioms describing his rationality the most prominent ones being transitivity and independence. The order in preferences is not supposed be reversed yet unknowns events of nature, seen as exogenous, may perturb our understanding of the given situation and may require designing new decisions going against decision theories, hence increasing uncertainty. In this paper we show that in an innovation project management, traditional decision making is not able to grasp expansion and generativity phenomena as a manager senses the unknown and endogenises it. To highlight this phenomenon we use Bayesian Nets with Wald's foundations to sense the reordering preferences in an industrial case and the benefits of designing one's playground and being intransitive. The purpose to contribute to the idea theories studying generative processes (design theory) by opposition to optimisation (decision theory) can help extend the underlying logics of innovation management and untangle the tipping point, the necessity to explore/exploit.