I wasn't arguing against that, I was arguing against a single slider with a whole bunch of different unrelated mechanics on it.
Then we weren't argueing lol. I would want a range of atleast a few sliders, though for simplicity they would likely make it 2 at most, who knows eh?
When an AI is clearly losing, they might have a priority shift, and assume a new decision making algorithm that dictates their behavior. For instance, if they are slowly losing cities to you late game and it appears that they cannot gain allies to turn the tide, they have a good chance to switch to a desperate-times-call-for-desperate measures algorithm in which they might, say, move all of their units into one massive army, leave all but perhaps their capital defenseless, and book it toward your channeler/capital city in a risky move to seize victory from the jaws of defeat (think Battle of the Bulge). If they are lucky enough to achieve this specific objective, it might improve the odds of other opponents (who were sitting on the fense or have poor diplomatic relations with you) being inspired to join with the cause against you.
Not a horrible idea, but truth be told there are non-military ways to win. If your close to a magic victory (or whatever) and the AI hits your borders with a horde of soldiers... im personally not going to be "happy". The AI should see your going to win, and start a limited war with you, but realistically suicide attacks to stop the player scream gamey.
How should it be handled? The AI should start problems with you and maybe even be given minor bonuses as the end begins to dawn, but as I said earlier, the end should be the end. Don't prolong a well deserved victory artificially. The games real shine will be in the fight the AI puts up as the game plays out, not in its final moments. I have played civ since civ2, and ive started probably a few hundred games. Ive finished maybe 5? 6? the point is, the endgame really isn't where its at.