Have you ever tried to go back where you use a GE and get beaten by one turn (as I did in the previous post). It could have been a sheer coincidence in my case that I got beaten again, or it may be the AI has a nefarious hatred of people using a GE to thwart their plans
I know some might call shenanigans on this, but I've never been beaten to a Wonder after I have burned a GE on it. (...or, well, that's true for Civ V, at least (vanilla, G&K, and BNW); I can't claim to remember back to Civs IV and III.) So I have no data on post-GE burning changes.
(To shenanigan-callers, let me repeat what I said earlier: when I play these kinds of games, it is only in the early part of the game that I have any competition for Wonders...and most/all of the early Wonders can be built in 1 turn by a GE. So I've had very few instances of using a GE and then having to wait additional turns while under threat of Wonder competition.)
However, if I'm reading you right, this isn't quite relevant to what you experienced. You burned a GE during both attempts to build the Wonder, and the post-loading adjustments you made were to other things. So, whatever the AI did differently the second time could not have been in response to you using a GE, because that factor was present in both situations.
Small things can often make all the difference in this game.
Definitely true, but unless I'm mistaken about how the RNG works, those small things will only be changed if the player introduces a new roll into the loaded game. The things you mentioned--selection of tiles for border expansion, barb razing/movement, etc.--are determined by the RNG, and since the RNG is determined well ahead of time, simply reloading a game will not change the outcome of these things. Border expansion will grab the exact same tile, barbs will do the exact same things, etc., each and every time. Unless a new RNG roll is introduced via player action.
So, if a player reloads a game and then changes nothing except working a few different tiles, or burning a GE, then the AI behavior should remain completely identical. These actions do not involve an RNG roll, and so the pre-determined order of RNG results will be the same across reloads. It is only when the player does something
else that is new, something involving a roll, that AI turns could possibly proceed differently across reloads.
Of course, all this goes out the window if "New Random Seed" is selected as a game option. Obviously. If a game is played with this option, then definitely these minor things could produce something different. Although again, I must add that even with a new random seed, I still have never ever experienced what is being described here. My last such "world pwnage" game, in fact, was with a new random seed: Prince level, Huge/Marathon, with New Random Seed activated. I found competition for Wonders to be particularly strong early on--this was my first post-BNW game of this type, and the linking between SPs and Wonders presented a challenge--and I was forced to engage in quite a bit of backtracking in order to round up all the Wonders. Even with the New Random Seed, tons of close calls, and lots of reloading, I never experienced as much as a single turn's difference in AI build times. Which is consistent with all my other experiences with such games, over a decade-plus of time.