Users of hazardous gas models are interested in whether any of them perform significantly better than others. They are also interested in the expected magnitude of the uncertainties in these models. These questions can be answered through application of standardized model evaluation procedures based on statistics. However, in order to confidently apply statistical procedures, certain criteria must be satisfied, such as data independence. The model evaluation exercise should take place in an equitable fashion, i.e. no models should have unfair advantages over others.