Can AI wipe out other empires?
- Thread starter Haig
- Start date
Shirotora Kenshin
King
- Joined
- Apr 26, 2014
- Messages
- 642
It happens, but very rarely. Really if you want a better AI in Civ6, do yourself a favor and use RomanHoliday‘s AI Rework Mod.
- Joined
- Aug 3, 2020
- Messages
- 779
You can manipulate the AI into not attacking on deity, and it works pretty well once you game the system that way.
The AI generally attacks based on its military score relative to yours, as a proportion oh how much they're ahead in that score.
There is no hard and fast rule at which point they attack, but generally it seems to be when they are (at a minimum) about 3 times stronger in that department than you.
This is also influenced by what sort of leader you're dealing with (Gandhi usually won't invade until the scales tip far further, while Chandragupta doesn't take much convincing).
This is why I usually open slinger or warrior on deity, both as a way to get my military score up (less likely to be invaded by turn 10), and because it gives me the flexibility of having one more unit for attacking the AI if I choose to go down that route later.
The thing with military score is that its also a deeply flawed system, as it basically just sums up the base combat strength of your units into a combined "military score".
This makes it easy to manipulate and abuse, because you cannot accurately represent the strength of an army by the sum of its units base combat strengths.
In civ 6, combat modifiers work exponentially and not additively.
For instance, a warrior has a base combat strength of 20, whereas a swordsman has a base combat strength of 35.
But since the game calculates military score by adding units together, the AI sees the military scores of 2x warriors as 40 (20+20), and thus as stronger than a swordsman's 35 (which we all know isn't the case, as a single swordsman can plow through multiple warriors on his own).
This means that if you are trying to avoid getting invaded by a technologically superior AI, you can invest in a buch of cheap warriors (to get your military score up) and be relatively safe as long as you keep an eye on the score and dont drop below ~1/3 of the AIs combined military score relative to yours.
You can also use this in reverse, and bait a backwards AI into attacking you.
I had a game once where I played as Gitarja, and I had very few units apart from some rather high tech ones (cuirassiers, artillery and battleships), at about 300+ military score.
Meanwhile Trajan, who had about 1100 military score, decided it was a good idea to declare on me, even though his army consisted almost completely of legions and catapults.
Needless to say, his army got utterly massacred, in part due to how poorly the AI judged military strength when looking at military score as a basis for deciding to attack into cuirassiers, artillery and battleships with classical era tech.
Last edited:
Oh. Good to know. (Always learning new things here!!!)
I did not know about the 1/3rd rule.
Had a game on the weekend where an AI came over, full green - ready for friendship, right?! - tried to friend them, nope... uh oh - can't friend them.
All of a sudden, four units on my doorstep. War. Must've been the ratio of my military to theirs.
For my style of game (culture, sub200) - pretty tight to invest in military early on.
I'm usually really focused on scout/scout/settler/holysite finished ASAP. Just putting in one extra warrior... ugh.
The warrior (20?) + scout (10), scout (10) is my usual opener to try to get huts/meets for golden age.
One warrior for barb camps (which often doesn't get it done in tundra with no CS help for clearing) and the two scouts sent East, West to try to meet the other civs.
I did a check on a morning-game I started... turn10 (one scout out, one being built) and I'm at "30" and nearest AI is at "110".
I guess that's where a 2nd scout getting to '40' pushes the AI military to needing to be 120+... for them to consider coming.
2nd warrior (with one scout) would be '50' and force the AI military to 150+... better cushion.
[A lot of games, if I coastal start, I'll try to do 'builder' and get them into the water just for exploration and maybe island 'hut' or two... but that's probably costed me a few early wars/losses.]
I'm super happy to know about the 1/3rd (3x).
I often run my scouts & warrior at less-than-full-strength bc moving and getting those huts/meets is so key in first age...
Does their "current" strength factor into the military total?
Wounded unit is some partial amount of their healthy strength?
ie - If I clear a barb camp, I may be at some% (say... 50%)... and I usually prefer to keep exploring rather than 'heal'... but that may be something I have to re-think about it if nearby AI is high with their military#.
I did not know about the 1/3rd rule.
Had a game on the weekend where an AI came over, full green - ready for friendship, right?! - tried to friend them, nope... uh oh - can't friend them.
All of a sudden, four units on my doorstep. War. Must've been the ratio of my military to theirs.
For my style of game (culture, sub200) - pretty tight to invest in military early on.
I'm usually really focused on scout/scout/settler/holysite finished ASAP. Just putting in one extra warrior... ugh.
The warrior (20?) + scout (10), scout (10) is my usual opener to try to get huts/meets for golden age.
One warrior for barb camps (which often doesn't get it done in tundra with no CS help for clearing) and the two scouts sent East, West to try to meet the other civs.
I did a check on a morning-game I started... turn10 (one scout out, one being built) and I'm at "30" and nearest AI is at "110".
I guess that's where a 2nd scout getting to '40' pushes the AI military to needing to be 120+... for them to consider coming.
2nd warrior (with one scout) would be '50' and force the AI military to 150+... better cushion.
[A lot of games, if I coastal start, I'll try to do 'builder' and get them into the water just for exploration and maybe island 'hut' or two... but that's probably costed me a few early wars/losses.]
I'm super happy to know about the 1/3rd (3x).
I often run my scouts & warrior at less-than-full-strength bc moving and getting those huts/meets is so key in first age...
Does their "current" strength factor into the military total?
Wounded unit is some partial amount of their healthy strength?
ie - If I clear a barb camp, I may be at some% (say... 50%)... and I usually prefer to keep exploring rather than 'heal'... but that may be something I have to re-think about it if nearby AI is high with their military#.
- Joined
- Aug 3, 2020
- Messages
- 779
Do note that its not a "hard" rule, its just what I've observed over thousands of hours of civ 6 (I have taken notes of military scores when the DOW happens, to get a general picture of ranges where a DOW is likely).
1/3rd military score is usually getting into the "it's somewhat dangerous depending on the AI leader"-territory, and getting increasingly risky beyond that point as the difference keeps increasing.
Usually you won't see an AI invade when it's twice your military score, but I'm not ruling out that it could happen, just that its very unlikely.
Neither does it mean that an AI will invade when it has eight times your score, but it is very likely to happen.
Factors that also influence heavily is how close they are to you.
Meaning that (assuming continents map), an AI that is 8x your military score but on a different continent, is very unlikely to attack you, whereas one that borders you is very likely to attack you, and probably rather soon.
Also the AI will almost never attack you without having their units in position, and this is also a thing that players on deity should look out for when trying to min-max the opening stage.
Usually I use one of my military units as a "forward scout", about halfway towards the nearest AI, between my borders and his.
If he for some reason has more than 2+ military units headed my way, this tells me that I need to be cautiours.
3+ units (and no barbarian camp or city state that he's attacking), and alarm bells are starting to go off (meaning that I should strongly consider switching to unit production instead).
Fortunately the AI tends to go in a straight line towards whatever it plans to attack, so I can be quite sure that he's headed for me and not a CS depending on his unit pathing (when multiple units are involved).
The reason this "forward scouting" is important to do, is that it allows you to "save" production and play more greedy.
Say if you want to rush out a religion (which means going super hard on holy sites at the cost of anything else), if you place a unit forward, you can get upwards of 7-10 turns notice that he is about to invade you, before the declaration actually happens (which gives you time to swap for unit production and getting units out).
This, combined with looking at the military score for the relative difference between you and the AI, lets you play greedy on things like holy sites, builders and monuments, without having to worry on whether you should "invest" in more military (assuming you want to play peacefully for the time being) to keep the AI from killing you.
Yes, unit health also influences your military score.
Usually this isnt a big deal, but if you are getting low on military score early on (you have 2 hurt warriors who add up to, say, 14 score, while he has 100), you might want to heal them up asap to buy you time before the AI decides that it will invade you since their 100 > 14.
Personally I dont actively scout much these days with my military units (scouting far off into the fog), rather, I keep 1 relatively near home to clear out barbarian camps (gold and era score), and the other one to look for the nearest AI.
That one then can scout a little more in the general area, but is mostly placed there as an early tripwire to let me know if the AI plans to invade, which lets me play greedy in the meantime since I know that I'm safe in the short term as I'll get a heads up.
I know that a lot of players that like to play scout opening even on deity (and it can certainly work), but its imo a riskier opening because you let your military score go down (higher risk of invasion), you cant clear barbarians with them properly, and tribal huts are often taken.
And you also want to actively scout with a scout, not passively scout the borders of a nearby AI (which a military unit can do just fine).
1/3rd military score is usually getting into the "it's somewhat dangerous depending on the AI leader"-territory, and getting increasingly risky beyond that point as the difference keeps increasing.
Usually you won't see an AI invade when it's twice your military score, but I'm not ruling out that it could happen, just that its very unlikely.
Neither does it mean that an AI will invade when it has eight times your score, but it is very likely to happen.
Factors that also influence heavily is how close they are to you.
Meaning that (assuming continents map), an AI that is 8x your military score but on a different continent, is very unlikely to attack you, whereas one that borders you is very likely to attack you, and probably rather soon.
Also the AI will almost never attack you without having their units in position, and this is also a thing that players on deity should look out for when trying to min-max the opening stage.
Usually I use one of my military units as a "forward scout", about halfway towards the nearest AI, between my borders and his.
If he for some reason has more than 2+ military units headed my way, this tells me that I need to be cautiours.
3+ units (and no barbarian camp or city state that he's attacking), and alarm bells are starting to go off (meaning that I should strongly consider switching to unit production instead).
Fortunately the AI tends to go in a straight line towards whatever it plans to attack, so I can be quite sure that he's headed for me and not a CS depending on his unit pathing (when multiple units are involved).
The reason this "forward scouting" is important to do, is that it allows you to "save" production and play more greedy.
Say if you want to rush out a religion (which means going super hard on holy sites at the cost of anything else), if you place a unit forward, you can get upwards of 7-10 turns notice that he is about to invade you, before the declaration actually happens (which gives you time to swap for unit production and getting units out).
This, combined with looking at the military score for the relative difference between you and the AI, lets you play greedy on things like holy sites, builders and monuments, without having to worry on whether you should "invest" in more military (assuming you want to play peacefully for the time being) to keep the AI from killing you.
Yes, unit health also influences your military score.
Usually this isnt a big deal, but if you are getting low on military score early on (you have 2 hurt warriors who add up to, say, 14 score, while he has 100), you might want to heal them up asap to buy you time before the AI decides that it will invade you since their 100 > 14.
Personally I dont actively scout much these days with my military units (scouting far off into the fog), rather, I keep 1 relatively near home to clear out barbarian camps (gold and era score), and the other one to look for the nearest AI.
That one then can scout a little more in the general area, but is mostly placed there as an early tripwire to let me know if the AI plans to invade, which lets me play greedy in the meantime since I know that I'm safe in the short term as I'll get a heads up.
I know that a lot of players that like to play scout opening even on deity (and it can certainly work), but its imo a riskier opening because you let your military score go down (higher risk of invasion), you cant clear barbarians with them properly, and tribal huts are often taken.
And you also want to actively scout with a scout, not passively scout the borders of a nearby AI (which a military unit can do just fine).
Last edited:
Been meaning to ask - you mentioned once (I believe) that you play on "Epic" speed, typically. (?)
Do units still get standard (Warrior:2, Scout:3) movement at each turn?
I would think that would help generate more AI-meets before the diplo-costs skyrocket.
I tend to feel this *need* to push two units (preferably scouts, but sometimes the warrior is so far I just say 'keep going')... quickly... to the bounds of the map and meet every AI.
On standard speed, by turn 60-70, the AI all start not selling their Diplo favor on the cheap... which is a major early-game currency-generator.
So meeting them all - for era score AND diplo-raking - is pretty critical (on standard speed).
[*I'm currently trying Bull Moose Teddy... with no early Unique Units/Bldgs... the era score is little tougher than those that have Ancient Era uniques - like Russia Lavra or Babylon Palgum/Sabum*]
I'm assuming the magical Diplo-Increase scales on Epic speed... and that you have 1.5x as long... (turn 60x1.5 or 70x1.5 ???) and you have 'more explore turns' before Ancient Era closes (turn45/50-ish x 1.5???).
This would make golden-age on longer-speed games, more likely - through explore Era points: huts, meets.
It would give somebody a different way to allocate their units.
I was also thinking about this with regard to "Moving Moksha" (different thread).
But does a governor (say Moksha) still move in 5-turns on "Epic"?
That would also mean that you have less downtime to shifting governors over course of a game.
*Great advice - thank you. I will start to track nearby AI military score a bit closer... especially healing my own warriors back up to their +20. This has definitely cost me a bunch of early 'quit'-games!*
Do units still get standard (Warrior:2, Scout:3) movement at each turn?
I would think that would help generate more AI-meets before the diplo-costs skyrocket.
I tend to feel this *need* to push two units (preferably scouts, but sometimes the warrior is so far I just say 'keep going')... quickly... to the bounds of the map and meet every AI.
On standard speed, by turn 60-70, the AI all start not selling their Diplo favor on the cheap... which is a major early-game currency-generator.
So meeting them all - for era score AND diplo-raking - is pretty critical (on standard speed).
[*I'm currently trying Bull Moose Teddy... with no early Unique Units/Bldgs... the era score is little tougher than those that have Ancient Era uniques - like Russia Lavra or Babylon Palgum/Sabum*]
I'm assuming the magical Diplo-Increase scales on Epic speed... and that you have 1.5x as long... (turn 60x1.5 or 70x1.5 ???) and you have 'more explore turns' before Ancient Era closes (turn45/50-ish x 1.5???).
This would make golden-age on longer-speed games, more likely - through explore Era points: huts, meets.
It would give somebody a different way to allocate their units.
I was also thinking about this with regard to "Moving Moksha" (different thread).
But does a governor (say Moksha) still move in 5-turns on "Epic"?
That would also mean that you have less downtime to shifting governors over course of a game.
*Great advice - thank you. I will start to track nearby AI military score a bit closer... especially healing my own warriors back up to their +20. This has definitely cost me a bunch of early 'quit'-games!*
Last edited:
- Joined
- Aug 3, 2020
- Messages
- 779
Yes I play mostly on epic, occasionally on normal and rarely on marathon.
These days I prefer slower speeds because I both like longer games, and because I like the challenge of slower speeds.
Generally speaking, the game is easier (in a way) the faster it becomes, and I genuinely think that the fastest speeds (quick/online) teach very bad habits.
Yes, pretty much the only variables that change with slower game speed is production time, research time, era point requirements and gold/faith costs, that all get ramped up in proportion to the game speed.
Everything else remains the same however, which means that tactical skill on the battlefield matters more (yes, the units behave completely the same, including unit movement speed that does not scale with game speed - 2 movement for a warrior remains 2 movement, regardless of speed).
The reason the game is generally a bit harder (imo) on slower speeds is because you have less room for error when it comes to military gameplay, especially early on.
Since production and research time is slower but units move the same, the "window of opportunity" for military units is open for longer.
With that I mean, the units can cover more distance and do more harm when they're technologically relevant, as the next tech military level takes longer to reach.
Thus you "do more" with those units as you simply get more time to use them before they become obsolete.
Since the AI starts with 5 warriors on deity and the player with 1 (and the AI is about to increase their lead further rather soon in tech/civics/number of cities and possibly units), you need to pay more attention to the military aspect, unless you want to get rolled over by the AI.
To contrast why I think online speed especially teaches bad play:
Say you get jumped by the AI declaring a surprise war, and caught completely off guard.
Well, you can just react and build warriors in 1-2 turns per city, and the AI might not even reach your city until you got units out.
And Iron Working tech is 8 turns around the corner (for swordsmen tech), meaning that as long as you can hold for that short amount of time, the AI deity push will crumble as you get swordsmen out.
Thus faster game speeds favour the defender, and early on that defender is usually the human.
Meanwhile contrast this to marathon speed:
The AI catches you with your pants down, and you try to react by building warrior and getting Iron Working.
Well tough luck, it takes you 20 turns to build a warrior, and Iron working is 60 turns away.
5 turns later you first city falls to the AI, and then the next shortly after, and soon you're dead.
This is what I mean with "bad habits", because the faster the speed is, the more you can neglect the AI threat early and "react" rather than "act" (pre-emptively, on information).
Playing a few games on epic/marathon can be a good idea if you want to improve on defending early on, because you really cannot react on a threat and have to learn to look for the cues in other ways (by acting through such means as looking for relative military score, and having a forward scouting unit to buy you time, in order to buy you time to produce to units).
That being said, epic/marathon speed can also be abused to give you a big leg up (assuming you play properly and dont get killed by the AI), in that things like governor establishment time also doesn't change.
Having Moksha establish in 5 turns is very strong on epic, and absolutely bonkers on marathon speed.
In normal speed terms, it's like having Moksha establish in 3 turns (epic), or 1 turn (marathon), which means you can leverage the district faith buying to a much larger extent.
Its also slightly easier to get era score on slower speeds, even when the era score requirements scale with speed.
The reason for this is mostly that you get more time to clear barbarian camps (they spawn at the same rate regardless of speed), or time to look for other opportunities.
These days I prefer slower speeds because I both like longer games, and because I like the challenge of slower speeds.
Generally speaking, the game is easier (in a way) the faster it becomes, and I genuinely think that the fastest speeds (quick/online) teach very bad habits.
Yes, pretty much the only variables that change with slower game speed is production time, research time, era point requirements and gold/faith costs, that all get ramped up in proportion to the game speed.
Everything else remains the same however, which means that tactical skill on the battlefield matters more (yes, the units behave completely the same, including unit movement speed that does not scale with game speed - 2 movement for a warrior remains 2 movement, regardless of speed).
The reason the game is generally a bit harder (imo) on slower speeds is because you have less room for error when it comes to military gameplay, especially early on.
Since production and research time is slower but units move the same, the "window of opportunity" for military units is open for longer.
With that I mean, the units can cover more distance and do more harm when they're technologically relevant, as the next tech military level takes longer to reach.
Thus you "do more" with those units as you simply get more time to use them before they become obsolete.
Since the AI starts with 5 warriors on deity and the player with 1 (and the AI is about to increase their lead further rather soon in tech/civics/number of cities and possibly units), you need to pay more attention to the military aspect, unless you want to get rolled over by the AI.
To contrast why I think online speed especially teaches bad play:
Say you get jumped by the AI declaring a surprise war, and caught completely off guard.
Well, you can just react and build warriors in 1-2 turns per city, and the AI might not even reach your city until you got units out.
And Iron Working tech is 8 turns around the corner (for swordsmen tech), meaning that as long as you can hold for that short amount of time, the AI deity push will crumble as you get swordsmen out.
Thus faster game speeds favour the defender, and early on that defender is usually the human.
Meanwhile contrast this to marathon speed:
The AI catches you with your pants down, and you try to react by building warrior and getting Iron Working.
Well tough luck, it takes you 20 turns to build a warrior, and Iron working is 60 turns away.
5 turns later you first city falls to the AI, and then the next shortly after, and soon you're dead.
This is what I mean with "bad habits", because the faster the speed is, the more you can neglect the AI threat early and "react" rather than "act" (pre-emptively, on information).
Playing a few games on epic/marathon can be a good idea if you want to improve on defending early on, because you really cannot react on a threat and have to learn to look for the cues in other ways (by acting through such means as looking for relative military score, and having a forward scouting unit to buy you time, in order to buy you time to produce to units).
That being said, epic/marathon speed can also be abused to give you a big leg up (assuming you play properly and dont get killed by the AI), in that things like governor establishment time also doesn't change.
Having Moksha establish in 5 turns is very strong on epic, and absolutely bonkers on marathon speed.
In normal speed terms, it's like having Moksha establish in 3 turns (epic), or 1 turn (marathon), which means you can leverage the district faith buying to a much larger extent.
Its also slightly easier to get era score on slower speeds, even when the era score requirements scale with speed.
The reason for this is mostly that you get more time to clear barbarian camps (they spawn at the same rate regardless of speed), or time to look for other opportunities.
Last edited:
Similar threads
