I also dislike the idea of using only a portion of the sensor and lens. That's why I believe the Scarlet is primarily intended as a B-cam for Epic shooters, to shoot 24 fps at 4k only. They can use the same glass, and only have a small crop factor. But I am certainly the one with the least experience and knowledge here, so take everything I say in this context with a rain of salt.
This is correct, of course, on a frame-by frame basis, but there are trade-offs in motion that can make it worth making the trade. Reading a smaller portion of the sensor can be done faster, so frame rates can go up. When frame rates go up, the image steps into an "unreal" place where resolution at the frame level makes a little less difference because the effect of slow motion takes such a big role in the image. Shooting 96 fps in 5K is stunning, but shooting 96fps at 2K is not a bad-looking image, and because it is not "real" time, it cuts together with 4K or 5K better than you'd imagine.
Yes, there is a loss of resolution, but how much it matters depends on what your deliverable is. It actually may matter very little, because the actual trade-off may not be less resolution, but less super-sampling. If you are delivering for a 4K projector, shooting in 2K is a huge loss in resolution and a significant compromise. But if you are delivering for TV, there is less compromise. To be sure, there are benefits to be gained from shooting 5K or 4K or 3K, but they are not primarily in resolution—even 2K provides nearly as much resolution as most any 1080p camera. The benefits to shooting in higher resolutions are the noise reduction and apparent sharpness that comes from super-sampling (when the scaling is done by a good-quality scaler). The higher the original frame resolution, the more of those benefits you will see. But if you give the 2K image plenty of light, so that it has little noise to begin with, it will stand up very well to the un-cropped frames.
There are two reasons why large-format glass is so much heavier and more expensive. Think about it. Going from 2K (S16) size to 4K (S35) size, you have to deliver light to an area that is 4 times larger. So the elements have to be 4 times as big, and just as optically perfect over that entire area. The second reason is the angle-of-view change. a 50mm lens at 2K will give a similar angle of view as a 100mm lens at 4K. A longer lens is generally bigger, heavier, and more expensive.
So to summarize: If you are, in fact, blowing up the image, you'll see a quality loss. But if your deliverable requires less resolution than any of the crops, there is much less of a problem, and in practice, the reasons you use the cropped sensor mask many of those issues. Cropping the sensor does, however, put a premium on good exposure, because you don't have super-sampling helping you out with noise reduction.
