There are notions out there about Star Wars where hyperdriven ships are super-quick, have virtually unlimited range, years of provisions, and are effectively untrackable in hyperspace.
If so, then assuming a lack of good intelligence about opposing forces and their intent, the model of warfare one ends up with is this game:
http://games.kde.org/game.php?game=konquest
In the game, you build ships at your planets and send them off to other worlds. You are unable to see opposing ships or even your own in open space . . . you can only count the number of ships at an enemy planet (unless blind mode is on) and see if it suddenly declines to know an opposing fleet has left. You cannot know its course . . . all ships move at the same speed and have unlimited range, so you can only narrow down a fleet's possible targets by simply watching the clock. "It's not there yet, so it must be headed to the next one", et cetera.
As a result, every maneuver is a gamble. If you send your fleet of 1000 ships toward enemy space, you can't know for sure that the computer AI hasn't targeted your homeworld with 1000 ships from 20 different worlds all due to arrive at the same time.
And while you can play the game in an island-hopping capacity, moving from planet to adjacent planet out of habit, you can also just go hog-wild willy-nilly if you wish, moving a large force to conquer a distant cluster of worlds with no significant need to worry about supply chains or any other such strategic considerations. Yes, a contiguous empire makes deploying reinforcements somewhat faster, but that's about all.
War thus becomes completely chaotic. Rather than massing forces at a front line, one must ensure that every world is adequately garrisoned against attack, or else gather forces at central points close enough to other planets that a mass reprisal can occur if a planet is conquered.
Were we to compare this with what we've seen in Star Wars, I don't think it adds up. To be sure, this should not come as a surprise, given that the claims upon which such a model would be based are known to be false, but it is nevertheless helpful to have the reminder.