ABSTRACT
The reliability of any model is dependent upon the quality of input data. Risk models are no exception. If they are to provide a reliable guide for decision making, they must be supported by a robust approach to risk estimating. Unfortunately, there are a number of weak approaches that are used in practice. Where they are used, it becomes a moot point as to whether a risk model provides results that are better or worse than deterministic models. Hubbard (2009) describes some forms of risk analysis as being worse than useless on account of their being convincing to decision makers whilst generating misleading results.
