>And pedants might insist that "lasting longer" is the definition of stability, not the result of it.

Exactly.

>I would postulate that stability is preferable to chaos and requires an effort.

This is where you miss my point. The stability can happen accidentally, from within the chaos. You must admit, if atoms are bouncing around in chaos for a several minutes, the chances of something lifelike forming are imporobable. But do that for a million years or so and now its very likely that a star can form. Once the star is established and stable, it will take alot longer to scatter its atoms than if it wasn't a star.