>Again, I'm hard-pressed to imagine how this would be better >than simply getting the exposure right in the first place. >You'll have motion artifacts either way.
Actually if that sample was a multiple exposure merge, it's pretty good in that there is to me little appearent motion artifacts especially in the masts.
But there is an advantage in stacking multiple shots with regard to SNR I believe.
The predominate noise source in images is shot noise. The benefit of multiple stacking is that the stronger signal will be accumulative and show little difference, but the lower signal will be averaged and shot noise would be improved (especially in fine gradient areas such as the sky).