This is the story of my game of Civilization IV Adventure Thirty-nine from Realms Beyond Civilization.
Believe it or not, I've actually never tried a game on Deity difficulty. I missed the few previous RBCiv events on it, and have never had reason to try a solo Deity game.
Deity is winnable, but not every map and not every time. Judging by opinion on Civfanatics, the way to handle Deity is to ignore military entirely at least until drafting. If someone declares war before that, well thanks for playing, here's our home game, go try a new map.
I originally hashed out what seemed to be a decent plan for this scenario: build the Great Wall and go for espionage economy with the spies. Great Wall is exactly what you need to ignore military, and espionage economy shines best on high difficulty when the cost discounts are critical and the AIs have more to steal and you won't have any lead to build wonders. But then I found the torpedo in that plan: I'd confused the Maya for the Inca and didn't realize we start nowhere near Masonry. I won't get the Great Wall first after three Deity-cost researches.
Anyway, we also need to metagame against Sullla's sponsorship style. His previous ideas of "extreme" (Winter Wasteland, The Gauntlet) have been winnable without much difficulty at all for top players, and I even won Holiday Surprise (though thanks to spoiler knowledge.) Ruff speculated in the game info thread about a "valley of 20 gems" - which would make Deity eminently winnable, of course. I doubt Sullla's gone that far, but there must be a catch. It's not an isolated start with an anti-invasion moat, since we know it's a pangaea map. My bet is more on a set of neighboring AIs that should be easy to befriend and manipulate, through elements like early religion and Hed Rule civic.