BACKGROUND A new locking-plate for femoral-neck fractures that provides angular stability to three screws in an inverted triangle configuration was evaluated. The plate is not fixed to the lateral cortex and therefore represents a new treatment principle. METHODS Twelve pairs of cadaver femurs (mean T-score -1,95 (range -4,5-0)) with subcapital femoral-neck fractures angulating 60° were randomly allocated to fracture-fixation using either three individual screws or three interlocked screws. Subject-specific axial force and torque were applied by a hip simulator and three-dimensional migrations were recorded. The femurs underwent 10,000 cycles of simulated partial weight-bearing, followed by 10,000 cycles of simulated full weight-bearing and stair climbing. RESULTS On average interlocking reduced femoral-head centre migrations 1.6mm (95% CI 0.1-3.1, P = 0.04). The intra-pair correlation of migration was 0.953 (Pearson's r). Interlocking did not change rotational stability (P = 0.87). Adding a locking plate did not affect the risk of failure, however all failed femurs were fixed using the smallest-sized aiming guide. CONCLUSIONS Adding a lateral interlocking plate to three screws might improve the fracture stability. However, none of the implants were able to resist the unwanted deformation of the proximal femur. Regardless of the fixation, female sex, reduced bone quality and small sized femurs appear to increase risk of failure.