I somewhat expected a thread like that, and I want to take the opportunity to throw in my 'two cents' here.
this cotm05 is pretty easy to win, and cotm04 was, too, but the competition is always there. It's more like playing against yourself, it's about how fast and effective you manage the game and your civ, in comparison to fight, dominate, and beat the ai.
cotm03 was not very inviting because of several reasons: it was demigod, it had a pretty unpromising start position and, after initial map discovery, it became clear quickly, that the map has its special difficulties. cotm05 is the other way round: regent, very prosperous start location, pangea maps are the easiest world forms to play.
I was thinking about cracker's games. I don't want to imply they were somehow 'better' than the recent games - the saves ainwood have been giving us were set up in a fine way with beautiful maps (e.g. gotm31-france), or, just think about the choice of opponents in the running game. They are fine with me. But cracker's regent games sometimes had some tweaking that made them pretty difficult, more difficult than regent in a random game though the ai building ratios were the same. on the other hand, his deity games sometimes were easier to play than you'd expected. i would say, some negation of the difficulties took place, that led me - personally - to disregard the given difficulty level in the first place, have a look at the game announcement and slip back to conquest if i expected the open class game to be too hard to have fun. i found it quite satisfying to win on difficulties higher than monarch for the first time though I 'knew' the game was tweaked and I 'could' win it if I would avoid doing too many dumb moves. Hey, the ai was still building for 80% or 60% of my improvement costs!
Maybe, it depends on one's attitude towards 'cheating' and the purity of the game in general. My personal view is that the basic settings provided by Firaxis are just one opinion of how the game should work. If the one who's designing a competition game thinks an ai opponent needs a start settler more or less for game balance it's fine with me when I get more sensible scenarios that lend from history or bring a special task I have to find and solve. If one thinks the best way to play gotm/cotm would be to give out random games probably disagrees with me.
I tend to support Raider, and, among others in this debate, solenoozerec, for the suggestion to make extended use of the different players' classes.