The transformers are very complex devices, consisting of different subcomponents e.g. winding, tap changer, bushing etc. During operation, all of the subcomponents are stressed and subjected to aging. When the aging reaches certain level, the subcomponent may fail. By application of statistical tools it is possible to investigate the failure occurrence in transformers. That means, that by fitting the mathematical model to the data, the parameters of the transformers' population can be estimated. By doing so, it is possible to see what is the stage of life of transformers, and what can be expected in terms of coming failures. However for a proper analysis, it is important to know how to deal with failures that do not seem to follow the statistical distribution chosen. Such outliers may affect the analysis considerably. In the paper, the influence of the outliers on the results of expected failures is presented. The latter is very important for the utility, as such information can be used as an additional information when deciding about replacement of transformers. Also, when deciding about spare transformers, knowledge about failure expectation is very valuable.@en