Question: - When a capital is destroyed what determines which city the new one is assigned too?
Is it the oldest, largest closest or just random?
I don't think it's anything so simple. In the game I submitted, I was paying reasonably close attention to where the capital went several times when I took one and it definitely didn't always go closest or largest. My guess is that Civ4 has an algorithm for determining city quality and it goes to the highest quality remaining city. I know in the past I've occasionally lost my capital and my memory is that it moved to a nearby highly developed city.
I'm going to give a brief game report of my current game. This is a Classic start, Aggressive AI to try to slow the tech, just the eight specified opponents. The game plan going in was to expand briefly while teching to Guild and then go to war with Knights/Catas and some Macemen and War Elephants mixed in. Then continue on to Cavalry and wing it after that. The games I've studied to formulate this plan generally end around Cavalry.
My capital was on the far northwest of the map on a peninsula with DeGaulle just to the south of the peninsula neck. With a normal mix of bad luck (no iron in my cultural borders, had to settle on iron the turn after I discovered guild, no religion for a long time) and good (DeGaulle seemed to be rather busy expanding and I managed to attack while his stack was far away), I was able to take all but one of DeGaulle's 13 cities, leaving just an island city to keep him alive to gift him cities later. Unfortunately, while I was doing this, Mansa and Freddy were depriving Isabella of her cities, eliminating her from the game. So I was left essentially with the six peaceful AI when I would rather have had at least one and maybe both of the hate targets around for their slow teching and easy later wars.
Tech-wise, I was doing well, keeping up with the top AI until I ran into the dreaded "We fear you becoming too advanced" response. Elizabeth was already slightly ahead of the pack, and she's continued to pull away. I've managed to stay within reach of Mansa and generally on par with Ghandi, Hats, and Freddy. But I've definitely been fearful that one or more of them might establish such a tech lead that I either wouldn't be able to reach a capital (since the map is so wide) when a spaceship launched or my military would be too inferior to conquer the capital.
I decided to target Lincoln next since he's close, the least advanced, and had a relatively small army. Just after my DoW, Mansa started planning a war, and I was almost positive I was the target. I figured my best path was to try to take Lincoln down quickly, and hope Mansa held off. Given my tech situation, and the timeline of the game, I didn't feel I could take a breather from war and gaining cities. Well, just as I prepared to take New York (the third city from Lincoln), Mansa declared war and took one of the French cities (Tours) right on his border. This is the present situation, screenshot below. The next ten turns or so will be pivotal. If I can turn back Mansa and grab a few cities from him, the only remaining obstacle will be stopping Elizabeth's spaceship when it launches. If Mansa slowly starts to eat away at my cities, or if we get into a deadlock, this will mark the beginning of the end.
I've found this combination of Deity/Quick Conquest to Time victory especially brutal. For Conquest normally you want a small, packed map with maximum opponents who you can immediately begin taking cities from. For a Time victory, you want a larger map with minimal opponents so that the domination land cap is higher, and there's lots of coast where you'll get a good number of sushi resources and be able to pack in a lot of coastal/island cities. And on Quick there just isn't that much time for conducting wars.
Against DeGaulle, I was always just managing to take the next city, never able to get an overwhelming army, feeling the years slide closer to a spaceship launch on a spread map, getting to use Knights than Cuirassiers then Cavalry for just ten or maybe twenty turns before the next upgrade came along. When the conquest was finally over, I had to take a breather to try to get my cities in better position to tech and eventually to pump out tanks.
I think all this is made harder by this particular combination of AI's, which in hindsight is better suited to a peaceful space game. Elizabeth and Mansa, often joined one or two or all of Ghandi, Freddy, and Hats, just like to sit there and tech and trade and tech and trade more, the tech train gaining more and more and more speed. Before this game, I quit a number of (non-Aggressive AI) games because my war just wasn't going fast enough and I was falling too far behind tech-wise. I think if everything goes just right, this kind of Classic start spread map Pangea Time game with these opponents can be won (especially if Mansa or Elizabeth is an early target). But with round about three weeks left, I think if I falter in this game, I won't have time for another Classic start, I'll try to turn to another modern start, this time with eight opponents and no vassals. (And no, I'm not passively lobbying for an extension, two months on this one game slot's enough for me.)
All that said, I'd like to thank Noble Zarkon for running this gauntlet. I've learned so much about starts across the eras, the way the AI trade with each other, AI selection, different ways to tech to various military combos, and more. Both the game I submitted and this game have had long, tense stretches where the outcome was in doubt, and have been fun, stressful, challenging, and interesting all at once. I definitely plan to use what I've learned to make a run at the Deity/Quick/Huge + Conquest/Domination/Time slots.