When analysing data from fast ion measurements it is normally assumed that the Larmor radii of the energetic ions are small compared to the width of the field of view of the measuring instrument. It is demonstrated here that this approximation is not always valid, by analyzing neutron emission spectrometry data from a JET experiment with deuterium neutral beams together with radiofrequency heating at the 3rd harmonic of the deuterium cyclotron frequency. In the experiment, the neutron time-of-flight spectrometer TOFOR was used to measure the neutrons from the d(d,n)3He-reaction. Comparison of the experimental data with Monte Carlo calculations shows that the Larmor gyration of the fast ions needs to be included in the modelling to get a good description of the data. Similar effects are likely to be important for other fast ion diagnostics, such as g-ray spectroscopy and neutral particle analysis, as well.