The simulation of 50 years wind-induced fatigue loadings of a slender steel structure is presented. The benchmark annual fatigue damage is obtained by applying Rainflow counting to concatenated 3 days long time series. The benchmark solution is adopted to discuss the classical fatigue analysis approach. The role of the length of the stress time series for annual fatigue damage evaluation is discussed. The variability of the annual damage over the 50 year simulation is highlighted and discussed.