Average turn that the AIs win at on each difficulty?

[LoneRebel]

Assuming Standard game speed, on average, on which turn (or range of turns) do the AI players generally win on each difficulty level? I want to know what to expect so that I can pace myself and take steps to win before then.

Apologies if this has been posted elsewhere; I haven't seen any thread asking this question yet.

 

[Unresolved]

They will win on Prince around turn 450. Subtract 50 turns for each additional difficulty level.

 

[budweiser]

If left in a vacuum they will win early. You, as the player may interact with them via various means during the course of the game to push the date back.

 

[LoneRebel]

Thanks for the answers. I guess we can assume that on Standard speed, on average the AIs win at: Turn 450-500 for Prince, Turn 400-450 for King, Turn 350-400 for Emperor, Turn 300-350 for Immortal, and Turn 250-300 for Deity.

Of course, not counting outlier situations such as an AI being reduced to one city or another AI becoming a super-runaway.

 

[dopingman]

The estimates are too high. Emp 300-350, Immo 250-350 (there is a large difference between earliest and last finishing time as the human player usally starts to interact with the AI quite early, preventing certain runaway moments and becoming more powerful meanwhile) and 220-280 for Diety. Unless the player puts considerable effort in cutting AIs down, you will see runaways nearly every time on Emp+.

AI estimates with human interference are not very meaningful imo as you are going to beat the living daylights out of an AI which is running away. Better rule is to take non-human interference values and calculate how much you delay the AI each game individually.

An example from my last game (Emp diff): Persia became a runaway around T190 on the other half of a Pangea continent. He started to tech like a madman and hit Industrial era T200 iirc. Cleaning up my half I was knocking on his door around T250, at this point he had rocketry and started to build the Apollo programm. I put a serious dent in his dreams of space vic by wiping him out (GWB FTW btw), but if left alone, I'm pretty sure he would've finished around T310-320. Maybe I'm just going to skip through the turns to see how it would've worked out

 

[pilot00]

Id say that till king I have yet to see a CiV win by turn 400 (and that their last chance since by then I start to get bored and launch the final part )

However there can indeed be exceptions. I still remember one of my fist games from vanilla on chieftain. Bismark had conquered his continent and was encroaching mine as well with a huge army. He had us outsienced too.

 

[Ninakoru]

The estimates are too high. Emp 300-350, Immo 250-350 (there is a large difference between earliest and last finishing time as the human player usally starts to interact with the AI quite early, preventing certain runaway moments and becoming more powerful meanwhile) and 220-280 for Diety. Unless the player puts considerable effort in cutting AIs down, you will see runaways nearly every time on Emp+.

Interacting with the AI is part of the game. However even in continents with the runaway on the other side, I've never seen such times for computer. I played a LOT of emperor games, they never were near to win before T400, even big runaways going SV.I am more inclined to this numbers: Emperor T400-420 Immortal T320-340, Deity T260-280

 

[Unresolved]

The estimates are too high. Emp 300-350, Immo 250-350 (there is a large difference between earliest and last finishing time as the human player usally starts to interact with the AI quite early, preventing certain runaway moments and becoming more powerful meanwhile) and 220-280 for Diety. Unless the player puts considerable effort in cutting AIs down, you will see runaways nearly every time on Emp+.

AI estimates with human interference are not very meaningful imo as you are going to beat the living daylights out of an AI which is running away. Better rule is to take non-human interference values and calculate how much you delay the AI each game individually.

An example from my last game (Emp diff): Persia became a runaway around T190 on the other half of a Pangea continent. He started to tech like a madman and hit Industrial era T200 iirc. Cleaning up my half I was knocking on his door around T250, at this point he had rocketry and started to build the Apollo programm. I put a serious dent in his dreams of space vic by wiping him out (GWB FTW btw), but if left alone, I'm pretty sure he would've finished around T310-320. Maybe I'm just going to skip through the turns to see how it would've worked out

Much of that has to do with how you deal with the AI. If you let them runaway like a madman, then sure, they can win on Deity sub 200. But even a moderate runaway won't generally win before 300 on Immortal, 350 on Emperor. And I've played a ton of Immortal games and they are never on track to win at turn 250 for me.

 

[Justice1337]

Map Size is another very large variable.

The AI's production, tech, growth, and gold bonuses are constant in each city. On the other hand, the benefits from the player's skill in management come from more focused cities, such as a NC capital with multiple Academies. Since it's impossible really to focus each city of an empire like this, games where the empire of each player is larger tend to favor the AI. Each AI city has all the infrastructure, population and flat bonuses of all the others. On top ofthat, the AI will expand to fill all available space, without regards to happiness.

So even on Emperor, a Huge Continents map with 2 players leads to the AI winning pre T300. Huge maps with less than 12 players on Immortal also lead to victory times that are shorter than some settings on Diety. And as one example, a Small Archipelago map with High Sea Level extends AI win times by quite a bit, even on Diety. Maps that produce runaway AI's are going to see very quick victories.

 

[dodgerphreak]

My past few experiences with King have had the AI win right around turns 380-390. Both of those wins were science victories for the AI, the first time beating me to a cultural victory by a considerable margin and the second one beating me to a science victory by 5 turns.... Agh the agony of defeat haha

 

[McBean]

Interacting with the AI is part of the game. However even in continents with the runaway on the other side, I've never seen such times for computer. I played a LOT of emperor games, they never were near to win before T400, even big runaways going SV.I am more inclined to this numbers: Emperor T400-420 Immortal T320-340, Deity T260-280

I've only got 2 emperor games under my belt and both were losses to runaways. I had been playing OCC cultural games on King so I decided to stick with that for emperor.

The first game was a T343 scientific loss to a runaway Russia. I lost it by 4 turns . Would have been a 2 turn diplomatic loss to Siam anyway but Russia swooped in for the science victory. That loss was my own fault as I forgot to shift from artist slots to engineer slots and I forgot to build the hydro plant. Those two things would have shaved off some turns when building the Utopia Project.

The second game was a T330 scientific loss to an insanely runaway Hiawatha. I knew it was gonna be bad when he had 10+ cities around T60-70. My scout was constantly seeing his settlers. He ended up with 70%+ of the map. No other AI could stop him. I don't believe AI's should be able to settle that many cities plus build up a massive army in that short of time.

 

[Hammer Rabbi]

i cant really speak to VC times lower than Emperor, but in my experience this has the been the range of victory by difficulty (assuming you left them alone, relatively speaking):

Deity: 225-275 (diplo/science on the low end, culture on the high end)
Immortal: 250-300
Emperor: 300-350

there are some outliers on occasion but on average these are my experiences on standard, non-archipelago maps with standard settings/speed.

 

[Traejun]

I've never had an AI win, even unmolested, on any level below deity sooner than T325 - and that only happened once on Immortal (Egypt, Earth map, standard/standard - AI got the Americas all alone). I (Rome) was stuck in a never ending war with India, Greece, Russia and China. I'd just gotten Muskets when Egypt hit the Atomic era.

 

[Hammer Rabbi]

interesting. ive seen several immortal AI wins sub-300. mind you they were usually in the 290s and a few in the 280s but i saw a few 260s and 270s sci/diplo wins and one t257 sci win from Iroquois. some of these might have been pre-patch though.

 

[atreas]

I've never had an AI win, even unmolested, on any level below deity sooner than T325 - and that only happened once on Immortal (Egypt, Earth map, standard/standard - AI got the Americas all alone). I (Rome) was stuck in a never ending war with India, Greece, Russia and China. I'd just gotten Muskets when Egypt hit the Atomic era.

That's exactly my experience as well. T325 seems almost a limit in my games for an AI Immortal/sci victory (runaway in another continent), although I have the feeling that it has also to do with the leader traits. Somehow I believe that AI "plays" much better certain leaders than others, or that is the impression I get from my games.

I think that a pre-T325 win by AI on Immortal would require some leader with science boosts, either direct or indirect. But that gives me the idea for another (slightly connected) topic.

 

[LoneRebel]

Glad to see that this thread got some more attention. There does seem to be a good deal of variation in the answers. I wonder if a pattern will emerge.

 

[pilot00]

I think that a pre-T325 win by AI on Immortal would require some leader with science boosts, either direct or indirect. But that gives me the idea for another (slightly connected) topic.

That most likely will be Sejong, Pacal or Nabbi. I wonder though how much can they use their abilities to facilitate things.

 

[WarIsPeace]

I agree with Justice1337. Map size makes a difference.

Map type also makes a difference. I always play large maps with all other settings on shuffle. When I'm given an islands map, there's a lot of DOFs early on because civs aren't bumping elbows with one another. This allows for more RAs. But it also means that civs without naval UUs (i.e. most of them) have difficulty using their UUs to go conquering.

I don't know that the OPs question is all that easy to answer without being more specific about map size, map type, and number of civs.

 

[LoneRebel]

It was supposed to be more of an "in general" sort of thing. I'm sure that map type affects the likelihood of wars, with Archipelago and Continents maps being more peaceful than Pangaeas; however the only ways this can affect victory speed that I can think of are 1. More RAs means the AIs advance faster, and 2. An AI manages to grow to super-runaway size by swallowing several other civs and thus advances faster.

Scenario 1 would probably happen more often on islands maps, while scenario 2 would happen more often on Pangaeas.

As for map size and number of civs, let's assume the standard number of civs for any given map size, i.e. 6 for Small, 8 for Medium, and so on. As Justice1337 pointed out, playing any given map size with less than the recommended number of civs would no doubt speed up AI victory by giving them more room to expand.