[andrew.roos]

@Andrew, you brought up light color that might add further to the fuzziness (I didn't know about that). I have tested on night street scenes, lit with mercury lamps, thus heavily biased toward the red end of the spectrum.

So, as far as I understand the optimal pinhole radius is the point at which the diffraction and the circle of confusion balance each other. If the distance from pinhole to film increases, light rays diverge further and the diffraction effect becomes more apparent, and you need to open up the pinhole to compensate that. But if you keep the same frame size, the COC becomes larger on the film.
So, apparently the "optimal" pinhole size for a 10° FOV is less "optimal" than the one for a 50° FOV - correct?

Red light has a longer wavelength (around 700 nm) and so is more affected by diffration than blue light (around 460 nm). Theoretically you could get about 50% more resolution if you used only blue light, and optimised the pinhole for the diffraction of blue light. In practise, I doubt it would be worthwhile.

The size of the pin-hole determines the angle at which the light will diverge due to diffraction. The smaller the pinhole, the greater the angle. However when your FOV is 10 degrees, then a 0.1 degree divergence represents 1% of your FOV. The same pinhole on a 50 degree FOV camera would still cause the light to diverge by 0.1 degree, but now it's only 0.2% of your FOV. This is why diffraction is more of a problem with a narrow FOV. So, as you put it, the optimum pinhole is less optimal on a narrow FOV than on a wide FOV camera.