Design methods based on a statistical approach to the prediction of dynamic effects of wind on frameworks of tall building are steadily gaining acceptance. Similar concepts can be applied to wind loads on cladding, although there are important differences such as exposure, size, and dynamic properties of the elements considered. Wind pressures and suctions at several points on a tall building recorded during strong winds are used to study the problem of predicting peak gust loads on windows and cladding elements. The statistical approach to the prediction of peak gusts, based on random process theory, is investigated. Measured peak pressures are compared to values calculated by a simple wind gust factor method and by a more detailed gust effect factor method. It is concluded that the gust effect factor method is preferable because it promotes a better understanding of peak pressures, but further field data and experimentation are needed before entirely satisfactory design estimates can be made.