In this note the relation between two simple approaches to estimate the extreme ship response used when no, or a limited, amount of data are available is discussed. The first one employs the long term distribution of the local maxima of ship response while the second one uses the expected number of upcrossings of a level by the response. It is mathematically demonstrated that the two approaches are equivalent. However, the upcrossing method is more straightforward and convenient for practical applications, particularly for non-Gaussian responses. The full-scale measurements of a 2800 TEU container ship during the first six months of 2008 are used in the comparisons.