Three things:

1: Thanks for the calculations SquintGnome!

2: Would you be interested in collaborating on a Google Calc sheet(s) that go over these different scenarios for a more comprehensive reference resource?
I prefer playing with people who understand the odds. ;)
The main trick will be the formulas which I will be able to help with once I know the format we are working with... plus a ton of copy/paste/number entry

3: 3/3 versus 4/2 is never totally clear cut, and especially after seeing the odds laid out (namely how close the two are) your really need to ask another question to choose between the two:
Do I intend to attack from that position if the first territory is compromised?
The reason for this is that actual dice rolls do NOT follow the statistics.
4/2 Will work out slightly better offensively if you wish to counter attack, but there is one major advantage that is easy to overlook with 3/3, namely that it maximizes the potential for a "lucky break" in the form of repeated 0-2 wins in your favor.
It allows the first territory to go 1-1 and still have two dice to defend with, which is huge when fighting an attrition battle, it may not work out (in which case if it follows the statistics you have effectively broken even, and if it does not your opponent got lucky, cest le vie.) but it provides a slightly greater opportunity for luck to play out in your favor than 4/2 does.

There is some truth to saying that we make our own luck. 

Also, as a predator in that case: I might look at your 3 and 3 and think that you are well defended, but your 4 and 2 as not so much, I just gotta get through the 4 and then the 2 will be easy.  

I am way more likely to attack a 4 and 2.  Because it looks easier if I can break it.