We examine the level of substructure and mass segregation in the massive, young cluster Westerlund 1. We find that it is relatively smooth, with little or no mass segregation, but with the massive stars in regions of significantly higher than average surface density. While an expanding or bouncing-back scenario for the evolution of Westerlund 1 cannot be ruled out, we argue that the most natural model to explain these observations is one in which Westerlund 1 formed with no primordial mass segregation and at a similar or larger size than we now observe.