Many features of an intact functioning heart provide a margin of safety with respect to the generation of cardiac arrhythmias. We have been investigating Ca2+ dependent arrhythmias and the requirements for such pathological activity including features that focus on the margin of safety. One of the primary issues is the number of cells involved in the initiation process. If too few cells are needed to initiate the arrhythmia, arrhythmias will be rampant. If too many cells are needed, then it is difficult to explain how known arrhythmias occur. Thus the balance is one of the critical parameter to be estimated to validate the model and also to develop a useful set of computational tools to examine lethal cardiac arrhythmias. It has been suggested in published work that, a large number (∼800,000) of ventricular myocytes be depolarized to overcome the electrotonic load of the heart. In our recent work we have re-examined this question in light of a revised stochastic model of electrical and Ca2+ activity. We have found that as few as 12 myocytes may be needed to initiate an ectopic cardiac arrhythmia. This reduction of over 4 orders of magnitude reflects important improvements in the modeling. These changes include the stochasticity of the Ca2+ modeling, improved assumptions regarding the ionic currents in heart cells and updated geometries. We will present and discuss these critical changes and the justification for them. Our new findings regarding the initiation of spontaneous arrhythmias are in line with functional data from the heart. This new finding suggests a mechanism by which the generation of arrhythmias is possible while still maintaining a high safety margin.