Does the AI ever threaten to win?

[RaidandTrade]

Genuine question. I play Deity but, judging by GOTM, not quickly; so I’m decent but not great.

Don’t get me wrong; I’ve lost cities in VII and abandoned an ancient era game or two when I’ve messed up a settle and been dogpiled as a consequence. This is like the sub-20 spawned too close of earlier Civs.

What I’m talking about is end-game actual defeat in the Modern. In earlier Civs, making it out of Ancient was not a guarantee of success; I’ve had actual losses where an AI has sniped a victory by fulfilling one of the VCs.

I’m enjoying Civ VII but when I check the legacies at the end, I’m yet to see the AI close to triggering one of the VCs. There’s therefore no rush or panic to head off an AI like there was in previous Civs. E.g., “X is culturally dominant” with victory in Y turns -> can I take the last cap/launch/convert/etc. in time or do I need to wipe them out.

Thoughts and experiences here?

 

[Isca Dumnoniorum]

I've not lost a game yet but Isabella ran me very close with a science victory just recently. I find the AI only really gets chance to compete if I get bogged down in war in modern. That's particularly fatal to the economic victory where I can't spend my gold building out my rail/factory/port network but I've also had games where most of the world has declared war on me and spammed war support such that my happiness goes through the floor and my yields tank. That makes getting across the line all the harder and while I'm drowning in war weariness, that gives the AI the time and space to make it competitive.

 

[bigjim1]

Like you, I have my doubts. Recently I was early in the modern era working towards flight when my advisor warned me that one of the AI rivals had just built their first Launch Pad! I assumed they'd win in short order, but they never did any of the three Science missions that lead to victory, so I got into space first
Seems the AI needs a few tweaks

 

[vorlon_mi]

I'm playing a more middle difficulty (Viceroy), but will probably step up a notch in my next game.

Military VC: The AI nearly always choose an ideology, but not quickly. In my last game, where I was trying to win by Science, everyone (including me) chose an ideology. The hate built up quickly, neighbors declared on neighbors, AI conquered my city-states. Since conquering any settlement (town or city) counts towards the legacy path, some made progress. If I was not aggressively working towards a different VC, they might have made it.

Economic VC: I've only seen a few AI hook up their factory resources. The distant lands AI struggle, especially on a Terra Incognita map. I'm usually far ahead and end the game before the AI can finish one of their other VCs.

Scientific VC: Every game, one or two AI achieve the first milestone and often the second. I don't know if they would finish all of the milestones, since I usually end the game with the Great Banker or World's Fair.

Cultural VC: This one is where the AI might be a threat. Yes, I'm aggressive in seeking out relics. But what if I weren't aggressive? The AI -- especially distant lands AI -- have built universities in Exploration, so they do the research and get digging. I've lost quite a few relics due to a footrace to get to the site. I expect that they will also, eventually, study Hegemony and dig up Antiquity relics. If the human player ignores relics, I expect one of them could collect enough to start the World's Fair. Having said that, I've not seen them progress to that point. Just call me a relic hoarder... if you must.

 

[bengalryan9]

I've definitely seen the AI threaten to win, especially in more recent games.

In one game a month or so ago I saw Himiko win a cultural victory a few turns before I completed the World's Fair myself. In retrospect I probably could've done things a bit more optimally at the end of my game but I was kind of coasting as I didn't really expect Himiko to do so and it kind of caught me off guard... I knew she had enough relics but there had been past games months ago where I saw the AI just refuse to build the final project over a long period of time so I wasn't really expecting it.

The last game I finished was actually another lost as Machiavelli claimed a science victory a handful of turns before I could finish my World's Fair. In this game I hurt myself a bit by changing victory plans halfway (I originally thought I'd go for science but the AI wasn't really doing well at collecting relics so I ended up getting more than I expected and shifted to cultural). He probably beat me by 8 turns or so. I will also say that the victory screen showed that two AI civs were also about halfway towards finishing the economic path as well so they do make an effort there too.

Even had one game where the map generator started me with 2 AIs right on top of me and me sandwiched between the two... one of them declared early war and while I was able to hold them off for a bit they eventually overwhelmed me with numbers and I ended up just rolling a new map where we were a little more spread out.

As I mentioned above, though, I think this has changed a bit over the last couple of patches, as I know in one of my very first games where I was just messing around with some different mechanics Lafeyette could've won a cultural victory but never built the World's Fair (maybe he didn't have space to build it?). Right now though my experience is that the AI *will* pursue victory, though only 1-2 of them will really make any kind of run at it.

I will say that I play on Immortal (and pretty much have since release), I'm sure I don't play anywhere near optimally, and I mostly play peacefully (so I'm not neutering the AI's performance by capturing its cities which may help its performance). For me personally, though, I've been pretty happy with the level of difficulty in my last handful of games.

 

[Yzman]

I won on the one Deity game i did but ironically lost once in this game in immortal. Isabella was a beast with 2000+ science and culture a turn and she got 14 artifacts. I was worried she was going to beat me by getting another, I was still at 13/15, and then i got the victory screen for science. She launched a spaceship out of nowhere! When I saw the summary screen she had also finished econ...

 

[Eagle Pursuit]

I've lost one game. Himiko (Queen of Wa) won a science victory when I couldn't track down the last capital in the fog while trying to get an economic victory. She beat me by just a turn or two.

 

[RaidandTrade]

Interesting to see the different experiences here - thanks for sharing.

Even had one game where the map generator started me with 2 AIs right on top of me and me sandwiched between the two

I’ve had this and it’s either a post-defeat re-roll or a real grind in Exploration to get settles out!

For those lost games, curious to know how many turns into modern the game goes for the AI to threaten/win?

 

[bumpyglint]

Genuine question. I play Deity but, judging by GOTM, not quickly; so I’m decent but not great.

Don’t get me wrong; I’ve lost cities in VII and abandoned an ancient era game or two when I’ve messed up a settle and been dogpiled as a consequence. This is like the sub-20 spawned too close of earlier Civs.

What I’m talking about is end-game actual defeat in the Modern. In earlier Civs, making it out of Ancient was not a guarantee of success; I’ve had actual losses where an AI has sniped a victory by fulfilling one of the VCs.

I’m enjoying Civ VII but when I check the legacies at the end, I’m yet to see the AI close to triggering one of the VCs. There’s therefore no rush or panic to head off an AI like there was in previous Civs. E.g., “X is culturally dominant” with victory in Y turns -> can I take the last cap/launch/convert/etc. in time or do I need to wipe them out.

Thoughts and experiences here?

I've tried to test this a bit, and honestly, the AI takes forever to actually win. Even in situations where the win condition is obvious, they don't handle it well. For example, I'm 99% sure that even after finishing the Legacy path, the AI doesn't immediately queue up the final project for production. I’ve seen Deity-level AIs take 15–20 turns to win after completing the path. And considering all the bonuses they get at that difficulty, that kind of delay feels really off.

I don’t remember the exact turn they finally won, but it was definitely after turn 100. That’s on Deity — so on lower difficulties, it likely takes them even longer. Meanwhile, a human player without any bonuses can usually wrap things up in under 70 turns once the endgame starts (I've seen wins in 40 turns, 70 is generous).

It really feels like the AI has no concept of "focus" — which is more crucial than ever in Civ VII. The modern era is basically a straight-up race to push one victory type, and the AI seems to just play in a generic, unfocused way. When it wins, it’s almost by accident — more of a side effect than a deliberate strategy.