About the computer cheating in battles: some people around here have made tests for lots of combats and found out that a large number of battles tend to have the predicted results. For example 3 out of each 10 70% battles are lost. Nothing was close to cheating. It's our selective memory again.
When playing on lower levels I used to save/reload a lot. A lost battle? Reload, try something different. Missionary failed? Reload. Lost a wonder race? Reload from an earlier save. Soon it started to get really long and really boring.
So I decided to try a different approach. Lose a lot of battles? It means that I didn't plan my attack well or my troops aren't appropriate. So I'm trying to think what other combo would work better, or wether I could choose another attack path. And use all this for the NEXT battles. The current game gets a little more challenging, but hey, where's the fun if you win everything easily? I expect to lose around a quarter/one third of the battles anyway, so I'll bring in more troops to compensate.
Lost missionary? Well, why exactly? Too many religions in the city? Shouldn't have used it there in the first place. The religion's influence isn't strong enough? Look for closer cities next time.
Lost wonder race? You know, not every wonder is an absolute must. So I tried to plan my wonders a little better. Going for a wonder means I prioritize that tech. I'll also look at what the other AIs have to offer. If they've already researched the tech a long time ago I consider the wonder lost. I've even played a game with the special rule of not building any world wonder. Quite painful since my Ironworks city was surrounded by flood plains and had 6 Engineers in it and could get Eiffel Tower in 5 turns!!
I'm not saying that reloading is cheating. Hey, whatever floats your boat! But I think without reloading I've learned more about how to manage my army, when to stop attacking like a mad-man (lost two good odds battles? maybe the remaining troops aren't up to the task of taking the city anymore), when to build a wonder and when not to. So reloading feels more like cheating yourself out of learning these things, than cheating the computer. Trust me, he doesn't care.
Plus, eliminating reloads gives me more civ, so more fun.
Reloading to try another long-time strategy falls into a different category for me. As Bluetooth said, you can learn a lot from this. It's like in chess when you try different lines. In an over-the-board game you might go for the calm one, especially if you're in time trouble, but later you could analyze it to see if the risky sacrifice could actually pay off. Right now I'm in the process of perfecting the Great Wall/Pyramids gambit in Warlords. Instead of restarting every time to try a different approach I'd rather play the same start over again to see what would get things done the way I wanted. By no means do I consider this a "serious" game, and will probably quit most of the tries when I achieve my objectives anyway.