Criticisms of bicycle helmet design are reviewed, and changes in the design since the 1990’s explored. Finite Element Analysis is used to model the impact of a generic helmet on flat and kerbstone anvils. The performance of current helmets was investigated using oblique impacts, in which the liner and rotational acceleration of a headform, fitted with a compliant scalp and a wig, were measured. Most of the design criticisms are shown to be invalid. The peak values of rotational acceleration were of the order of 5 krad s-2. However the coverage of the side of the head is not optimal.