Literatronic is an adaptive hypermedia system for hypertext fiction. Its adaptive features are based on an algorithm that simulates a Hamiltonian cycle on a weighted graph. The algorithm maximizes narrative continuity and minimizes the probability of loosing a readers attention. The metric for this optimization is defined as the minimization of hypertextual friction and hypertextual attraction. We consider the challenges involved with modeling such hypertext, and we offer specific examples of this type of adaptivity.