If you’re using a game that renders each frame at an instant in time, and the aim is to get a better approximation of true motion blur to your eye, the theoretical maximum for getting smoother motion blur is gonna be when the thing is moving at one pixel a second, which is higher than the rate at which we can distinguish between individual images. Well, okay, maybe a bit more, since you could hypothetically have sub-pixel resolution.

But point is, more rendered frames does buy you something even past the point that they’re not individually distinguishable, unless the game’s rendering engine can render perfectly-accurate motion blur itself.