Some time ago one of my cousins described a technique for flight simulator displays where the direction the pilot was looking was monitored. The center of his field of view was generated in high resolution while things in the pilot’s peripheral were generated in much lower resolution. An observer watching the pilot in the simulator could clearly see the high and low resolution areas of the screen. However, to the pilot, it appeared that the entire screen was in high resolution. This technique was adopted to maximize the use of limited computing power in rendering an image for the pilot. Although a smart watch may not have the same computing power limitations, it would seem that the screen area limitations could be addressed by a similar technique.
The current generation of motion sensors is very small and very precise. They could be used to create virtual screen area to compensate for very small screens. This is standard in “near to eye” applications but could also be useful on a wrist mounted device. But there is no real substitute for just using a larger display. The first wrist watches were pocket watches with a wrist band. They kept pretty-much the same size although they were eventually engineered to be much thinner. Given their single function, there was never much of a point to making them bigger. That is not the case with a smart watch. A large cylindrical display with the virtual screen area enhancement would have interesting 3D effects as well.