Ok, was going to put this with my last post, but figured it was different enough (and the last one was long enough), that I should make a separate post.
When I was a senior in college, I saw this really interesting talk by a game designer who was trained as an engineer. I forget his name, but he was working on the City of Heroes CCG. The talk was about applying what you learn in college in novel ways. His example: a CCG is basically a chaotic system. As he put it, a set contains a couple hundred cards, each of which breaks the rules in its own unique way, and it's not possible to test out every card combination. You only get occasional times when you as the creator can control it (by releasing a new set, or by banning cards). So, how do you keep particular card combinations from being optimal? You put in negative feedbacks. In other words, he said about 10% of his cards are specifically designed to undo the remaining 90%. His example was, if a combination of blue power-ups gives you a super-super hero, you just put in cards that specifically destroy blue power-ups. Then, if everyone starts using that combination, people can win by playing an anti-blue deck.
In my game, in some ways I've tried to do the same. There are a lot of abilities, but for the most part there are ways of countering every kind of special ability (except maybe melee). For example, Aura of Chaos, Take Control, and Unsummon un-do animated creatures and summoned creatures; Aura of Mist and Sunbolt makes archery much more difficult; Antimagic Aura makes magic dangerous. Thus, if you base your group around any single type of attack, it can be undone by a single character.
So, the question is, what would stop a party from using these every time? Well, the answer is actually very simple, and I kind of discovered it in my last playtest. I made my group anti-summoned creatures, anti-constructs, and anti-archery. None of these abilities worked perfectly, but worked well enough that they could have shut down any of these types fairly well. The problem: nullifying all of these forms of enemy attacks required me to use most of my party's abilities. In the end, it was almost like neither of us had any abilities at all (or at least it would have been had I not spent so much time in hiding, see next post).
So the answer becomes this: When you plan your party, if you are particularly afraid of one form of attack, you can protect against it. However, doing so has an opportunity cost great enough to discourage you from trying to protect against everything.