I'm certain as a deity player you are aware of the speed differences. However, the assertion that less-cost-than-settlers early city conquest on deity/normal is CONSISTENTLY possible, or even possible a majority of the time is ridiculous based on everything I've seen. I've not seen that out of any other high-level player on this forum on a consistent basis above marathon, especially given the still-to-be considered RNG dice throws in such a situation. Depends on whether you're facing creative opposition or not (a diety 2nd city will take only a little more than 50 turns to hit 40% if the AI is creative), and which city you're taking. Regardless, more spammy AI personalities can easily throw a bunch of archers post-haste. Yes, if you use perfect micro you might take your 2 cities and 10 axemen against their 3rd city or so at 20% culture defenses. But then you realize that they still have more cities than you, possibly have metal (AI on deity can easily have metal units in the turn 20-30 range if you're not lucky), and can now make almost a unit/turn in remaining cities. If you're just picking off a city, they're not going to let you buy out cheaply. If you're trying to kill somebody off who has 4 cities by 2000 BC (you can NOT rush someone with decent #'s by 2000 BC on normal consistently. Nobody can do that.) and probably double that by 1000 BC, you better EXPECT to see a minimum of that kind of defense in the majority of the AI cities after you declare and take that first one. Just like you, I can cite tons of experience here, but can pile on the fact that very few deity/normal players early rush unless the situation favors it severely or they're desperate. Okay, so now the rush is predicated on the opposition not having one of 2-3 traits, not being a 35+ unitprob AI, not having a resource-less UU that screws you, not having lots of hill cities, and being situated in such a way that rushing doesn't feed another AI more cities. All of that and if the RNG treats you wrongly, you are STILL going to lose if you fail there. Yes, you get wonders/workers/etc if you successfully capture them (though if it's truly an early rush you cut into your odds of capturing a lot of goodies, which hurts your cost-benefit argument side), but you're talking 2-3 axes per diety AI archer or you risk losing. You're trying to tell me that's consistently possible? The numbers aren't adding up here. Though for straight $$$ rush, towns beat everything (although windmills with env aren't bad). Don't even get me started on random events, cottages or whatever. They would be the most broken, ill-conceived trash ever to make it into civ IV, except they're in line behind global warming, the apostolic palace, and the user interface/controls. Ever had the bermuda triangle even wipe out tens of thousands of hammers via sinking a naval stack? Just realizing such a thing is possible is enough to deem them broken and turn them off (nevermind the auto-loss vedic aryans that can still glitch-hit humans if unlucky enough), so let's not use trash features to sully this otherwise good discussion. Spies attacking towns is admittedly annoying. It's so bad that the pillage argument isn't worth considering from the AI, except for fish boats where it sells its soul to pillage (but that's not relevant to land tile improvements). The AI will only pillage with extremely small sub-5 unit (usually just 1) stacks, or with its 2-move units traveling with its primary SoD. Neither of these things take any effort to prevent, assuming you have enough military to survive (if not, pillaged tiles are irrelevant). In MP the consideration is greater; pillaging is a legit threat. Again, however, I emphasize that players who can't prevent pillaging of flatland tiles are almost 100% to lose regardless of what tile improvements were actually getting pillaged. There's no reason to expect anything is getting put back there. It's fair to argue that you simply prefer not to use cottages and have workable methods around them; that still doesn't justify the general advice against using them at all. The best advice is to only use them when appropriate, and to give players an idea that they are not NEARLY as good as "spam and forget" tiles as they believe...well unless you're playing FIN.