> Velocity in a gas is a distribution; raise temperature and increasing fraction exceeds escape velocity. It's not going to be a significant fraction to matter

> Velocity in a gas is a distribution; raise temperature and increasing fraction exceeds escape velocity. It's not going to be a significant fraction to matter, except of truly geological timeframes. > "Realistic" for values including "let's build a mirror the size of a planet, in space, and keep it together for centuries". Probably thousands of years. But not hundreds of thousands.