It was suggested that five is the minimum number of data points for the least likely cate gory in constructing histograms. In many cases, however, we must estimate probabilities that are extremely small or for which relatively few data are available. In the example about machine failures at the beginning of the chapter (pages 356-357), suppose that we had one day out of 260 in which three failures occurred. What do you think we should do in such a situation?