I don't have a problem with naval battles being decisive. The problem arises from doomstacking, so the entire USN dies in one bad battle.
There is also the incredible difficulty in modeling WW2 naval combat in a game where the entire range of naval mechanics consists of "find the bad guy ships" and "find the bad guy convoys"! Commerce raiders? I mean, how the heck do you model a German heavy cruiser out in the woods shooting convoys once in a while and make it worthwhile, and then how do you balance that out with the difficulty of having the Allied player finding the ship and killing it?
Naval combat is really, really tricky to model, especially since a single bad fight due to luck (like Midway, if I may be so bold to open that can of worms) can cost a player literally years of dedicated NF production. Make a bad division template or lose a battle in North Africa, all right, lick your wounds and keep fighting. Lose all 4 of those Shokakus you spent the last 2 years building? Hah, good luck!
I suppose the point I'm making is as follows: in a game where the player has little control over how the actual battles go (we only make division designs and operational decisions, production, etc, we don't control the battles themselves), naval combat is subsequently extremely dicey, given the large amount of time it takes to produce capital ships. The stakes are higher, so to speak. We need a lot more control over our navies with the stakes as high as they are.