Not to sound completely crazy, but doesn't doing things like selling strategics to other civs completely defeat the purpose of playing on deity? You would never make such a trade?
Actually, selling strategics is perfectly sensible on both sides of the trade. AIs that have surplus copies won't buy them; AIs that have no desire to build the associated units won't buy them.
The luxuries are another matter entirely.
I agree that it is quite difficult to say that greece's UA is somehow equivalent to a 4 GPT bonus. Not to take anything away from Matrin's math, but as you pointed out and others there's a lot more to it than just that. You can lose a CS ally, it can be killed, you can get allies through quests, etc.
Yeah, the point of the exercise was simply to show that
at best that's the most you can derive from the bonus. When you compare the Greece UA to other civs' abilities, it comes up short both in terms of raw power and due to low flexibility resulting from the requirement of a city-state-centric strategy.
Greece looks pretty because you can take Patronage and basically fire-and-forget on a city-state. That isn't an option with anyone else, and so it feels really powerful. Unfortunately, if you get under the hood and look at what you actually get out of the UA against what you could be getting with another civ in an identical situation, Greece's UA starts to look very weak.
have not done the math but 4 gold per turn is big if you have ten plus city state allies I usually play a large map and try to buy out as many maritime and culture states as possible 20 city states can be 80 gold per turn by that math.
As already noted, if you have enough

to throw around to get that many allies and your enemies don't have enough

to compete with you for the favor of those allies, then you already won. You just haven't spent enough time on that particular game to generate a win condition yet. Discussing the power of a strategy or ability under those conditions is like discussing chess strategy when you're already up a rook, which boils down to: "Force exchanges whenever possible and don't do anything really stupid." It's trivial.
What you should care about is what's going on during the period in time when you haven't generated a winning position yet.