Shouldn't we continue the loop and check if there is another player with also a wonder more advanced in the countdown, instead of choosing the first one with a built wonder?

If keeping it that way, I would change it to "turnFactor *= 1 + relicsCount;" so that already 1 relic gives more chance to be chosen

but i wonder if you shouldn't do as for wonder, i.e. use a constant penalty rather than one depending on the number of units, so

turnFactor += relicsCount * 150;

because if you have two allies, one with 151 units and one with 50 units and 2 relics, i would rather choose the latter first.

And finally, as 150 is used several times, could be a parameter (defined either at the top of this function or in the config file, could also be changed to something like maxPop/2 if that's how you've tuned it).

Well, this check for wonderProgress is based off of the one in getEnemyPlayer in attackManager, which also breaks at the first wonder found with an undefined foundationProgress (I've found that it is usually a good idea to follow such patterns that exist in the code already). And in any case, I don't believe there is a way to check for a wonder with a more advanced countdown once it has been built, and even if such a way was added, this may be taking the AI too far (and it becomes "cheating" in a sense). So I'm tending to believe this isn't needed.

Okay.

Also I forgot to use the updatingGlobalCollection "allRelics" instead, so I will use that instead.

It's quite a long time I've played with wonder, but iirc the countdown is displayed for all users (maybe not as soon as the wonder is built, but once it reaches some limit), so i don't think it would be cheating :)

But i'm also fine with keeping the code as is. The real solution will then to ensure that, when a countdown is displayed for all users, this info is sent to the ai and, and when such a countdown is running, the ai chooses the enemy with the more advanced countdown. But that looks like a next patch.