Thoughts on magic and adjusting the research cost

In my tests for viability it is standard victory with 6 Indeps, I find this is strong (or weak depending on how the AI is made) for Impossible AI's to hyperexpand and prove the hardest challenge for me. I usually play 1 under the maximum recommended of AI's for a map as it gives a little room but not very much so you will have fairly quick conflict or at least territory. I also avoid using cheeze diplomacy, as it negates some of the effects.