In the case of flood determination the commonly used methods are often questionable as to their accuracy and sometimes even based on disputed concepts. Basic questions concerning flood formation remain to date unanswered. In the following paper the question as to the validity of the methods used in Switzerland to estimate flood discharges is examined. A comparison is made with the results of rainfall simulation experiments. lt is shown that the travel time method almost always results in an overestimation of flood discharges. The SCS method has a more realistic basis, though the parameters used to differentiate soils, namely vegetation and soil type, are insufficient and unable to provide a reliable prediction of catchment behaviour. The Kölla method approximates the main flood forming processes well but can not replace a more detailed catchment evaluation. Rainfall experiments show that flood discharges are often determined by still little understood subsurface flow processes and that even under extreme precipitation there may be no contribution from large parts of catchments. The observed processes are too complex for a simple procedure to be proposed for flood evaluation.