I must be really thick, but I don't understand the advantage of applying stabilization to the nested version of the media rather than to the original media in the original project. What has been gained?

Even more puzzling is the idea that you would stabilize the nested project which contains both shaky and NON-shaky footage. Since stabilizing always degrades the image somewhat, why would you stabilize footage that doesn't need it?

Like I said, I must be thick because I'm not understanding the advantage.

[edit] OK, this is a way to get around the inability of Mercalli to batch stabilize, yes? If so, it unfortunately still isn't a solution because you are stabilizing footage that doesn't need it. Stabilizing non-shaky footage is a bad idea for three reasons:

1. It forces a re-render, which always introduces degradation.
2. It will slightly crop the footage, resulting in a loss of resolution.
3. It can actually make the footage more shaky.

The last one may be counter-intuitive, but as many people in these forums know, I spent a lot of time working on stabilization (with Deshaker) and learned a thing or two. Any deshaking algorithm assumes there is shaking and therefore can easily find something to correct, even in footage that is taken on a tripod. If someone enters the frame close to the camera, for instance, the deshaking will start tracking that object, and your locked down shot will start to move.

So, if this is an attempt to batch stabilize, it is clever, but I don't think it is viable.