The goal of this study was to determine quantitative relationships between electrophysiologic parameters and the propensity of cardiac tissue to undergo atrial fibrillation. We used a computational model to simulate episodes of fibrillation, which we then characterized in terms of both their duration and the population dynamics of the electrical waves which drove them. Monte Carlo sampling revealed that episode durations followed an exponential decay distribution and wave population sizes followed a normal distribution. Half-lives of reentrant episodes increased exponentially with either increasing tissue area to boundary length ratio (A/BL) or decreasing action potential duration (APD), resistance (R) or capacitance (C). We found that the qualitative form of fibrillatory activity (e.g., multi-wavelet reentry (MWR) vs. rotors) was dependent on the ratio of resistance and capacitance to APD; MWR was reliably produced below a ratio of 0.18. We found that a composite of these electrophysiologic parameters, which we term the fibrillogenicity index (Fb = A/(BL*APD*R*C)), reliably predicted the duration of MWR episodes (r2 = 0.93). Given that some of the quantities comprising Fb are amenable to manipulation (via either pharmacologic treatment or catheter ablation), these findings provide a theoretical basis for the development of titrated therapies of atrial fibrillation.