Models for predicting an individuals risk of developing cancer are complicated by a lack of well-conserved mutations present in most types of disease. To address this challenge we developed a simple lottery simulation of cancer development based solely on the abundance and relative survival advantage of mutations present within individual stem or progenitor cells. In simulations requiring few mutations or that involve mutations exhibiting a strong survival advantage cancers develop in our simulation in a progressive manner amenable to oncogene-based risk prediction. In contrast, in simulations involving numerous mutations that lack a strong survival advantage, a stochastic process of neutral drift and punctuated equilibrium determines eventual tumour formation. In these situations, the development of cancer is largely attributable to the rate of mutational variability amongst progenitor cells rather than the identity and relative survival advantage of the mutations present at any single time point. These results suggest that measuring rates of stem cell mutational variability represents an important component of predicting an individuals risk of developing cancer.