This excerpt taken from the ALL 10-K filed Feb 25, 2010.
The nature and level of catastrophes in any period cannot be predicted and could be material to our operating results and financial condition
Along with others in the industry, we use models developed by third party vendors in assessing our property exposure to catastrophe losses. These models assume various conditions and probability scenarios. Such models do not necessarily accurately predict future losses or accurately measure losses currently incurred. Catastrophe models, which have been evolving since the early 1990s, use historical information about hurricanes and earthquakes and also utilize detailed information about our in-force business. While we use this information in connection with our pricing and risk management activities, there are limitations with respect to its usefulness in predicting losses in any reporting period. These limitations are evident in significant variations in estimates between models and modelers, material increases and decreases in model results due to changes and refinements of the underlying data elements, assumptions which lead to questionable predictive capability, and actual event conditions that have not been well understood previously and not incorporated into the models. In addition, the models are not necessarily reflective of actual demand surge, loss adjustment expenses and the occurrence of mold losses, which are subject to wide variation by event or location.