Quote Originally Posted by FlyingScanian View Post
I think I actually saw math somewhere that indicated that it did at least some difference against massive dakka (the example was 30 wounds vs. 5 terminators), due to the fact that while each model will on average fail one armour save, some will instead fail two, while others will fail zero...

Since I suck horribly at statistics (or, at least at calculating statistics and variances), and don't remember where I saw it, however, take it with a bit of salt...
Assuming two 3+ save models take four wounds...

If they are identical (two or more saves have to be failed for both of them to die): 11/27=40.74% probability for both of them to die (if I didn't miscalculate).

If they are different (both of them have to make two saves; in order for them both to die, both have to fail at least one of these saves): 5/9 probability for each of them to die individually, so 25/81=30.86% probability for both of them to die.

So, yeah, it improves your chances quite a bit. Though it makes the process of rolling saves for this unit incredibly tedious.