EC2 presents two alternative methods for designing short-span beams without shear reinforcement that have multiple point loads close to supports. The simplest option is to reduce the component of shear force owing to loads applied within 2d of the support by the multiple a_v/2d (where a_v is the clear shear span and d is the effective depth). Conversely, the previous UK code BS8110 increases the shear resistance of sections within 2d of the support by 2d/a_v. The two methods are equivalent for single point loads but not for multiple point loads. EC2 also allows short-span beams to be designed with the strut-and-tie method (STM), raising the question of which method to use. This paper compares the predictions of BS8110 and both EC2 methods for a series of beams without shear reinforcement tested at Imperial College London. The beams were loaded with either one or two point loads placed within 2d of the support. This type of loading occurs in structures such as the cross-head beams of bridges. Strut-and-tie modelling is shown to give significantly more realistic predictions of shear strength than the beam design equations of BS8110 and EC2.