Which just gets back to the problem I mentioned earlier. You’re looking for patterns. I can counter your observation based on my own observations of the times the AI did suboptimal things. And here we are, neither of us the wiser. We’re just sharing anecdotes. Wouldn’t it be better to get actual data before making claims like you’re making?
I mean, even your comment about the AI understanding flanking is confusing to me. All I know about flanking from the game’s documentation is that it’s a trait for units with a general who’s a Leader of the Commander archetype. In this very specific situation, the flanking trait gives the unit a 20% bonus to attack. And didn’t you claim upthread that you thought the AI doesn’t take into account unit traits?
Actually, from reading your comment, it sounds like you’re saying there’s a bonus to flanking regardless of a unit’s general. The encyclopedia doesn’t mention it beyond the Commander bonus. I did a keyword search in your manual and found two uses of the word “flanking”. First, in the section on Commanders, you say it’s a trait units get from Leader generals. But then later you have a section on flanking where you claim it’s a constant bonus for all units. However, I just tested it and I’m not seeing the bonus listed.
So there are a couple of things going on here. The first is that I’m again confused about how Old World works because I’m getting conflicting information from the documentation. But perhaps more importantly, you’re telling me you’ve observed the AI setting up flanking, yet it seems to me flanking doesn’t even work like you seem to think it works.
In other words, you’re making inferences based on assumptions and anecdotes, and they don’t seem to jibe with the game’s actual rules. If there’s no flanking bonus without a Commander leader as a general, you’re not seeing what you think you’re seeing.
Tom - I’m only telling you what I have seen after literally hundreds of hours of gameplay. Start a new game, play the first 30-50 turns. Start another new game, play another 30-50 turns. Repeat. For months. Fight a battle. Reload. Change a few options. Watch the AI respond to it. Reload again. That’s how I learn. That’s where my conclusions come from.
Over that time, the AI’s behavior has changed markedly. It is still changing. Virtually every update causes a few new behaviors.
My information is imperfect. I made no secret about that, from my very first reply.
I apologize if the information isn’t as detailed as you would prefer. It’s what I’ve got, so…I guess we’ll wait until someone with better information provides it.
I’m not inferring when I say that the AI consistently moves all its units up before attacking. That’s what HAPPENS.
The AI consistently runs for defensive terrain (urban, forest, scrub) when threatened. Not an inference. It’s what happens. Every time.
These aren’t anecdotes. We can quibble about exactly where the line is, but at some point, after the 100th, 200th, 500th test, it stops being an anecdote and starts being a dataset.
What I can’t say is how well the AI understands promotions. I don’t know. I can’t see the code, and even if I could, I wouldn’t really know what I was looking at.
In any case, since I seem to be adding to the confusion, I’ll stop.
But again, just to make myself clear, you’re saying it does that for a flanking bonus that doesn’t even exist. You insisted in your post that the AI understands flanking, when instead, it looks like you don’t understand flanking*.
That, to me, is a perfect example of the pitfalls of discussing AI. :(
-Tom
* The other alternative, which I will gladly accept, is that I don’t understand flanking. Right now, the encyclopedia and tooltips support my understanding. But the section on flanking in your current draft of the manual contradicts the encyclopedia and tooltips.
That sounds like choosing Polytheism (over Monotheism), which I do almost every time - especially for the civs which get that badass adjacent lumbermills gen shields shrine.
The 3MA guys made it sound like one could actually elevate * Paganism to the state religion, which seems to be this button action:
I will take that gold star. Because, yes, that’s what I’m talking about. Monasticism lets you choose Legal Code for a civics boost or Divine Rule for pagan State Religion (also useful for keeping non-State Religions in check). And if you’ve got a pagan state religion, I can’t imagine not using Polytheism’s multiple shrines to then perpetuate your new State Religion.
But, like I said, it’s a big step down from the World Religions and their monasteries, temples, and shrines. Paganism doesn’t get theologies, either, which can be a big deal. But with the right shrines…like I said, now I want to play a Roman pagan nation.
Sorry, on vacation so don’t really have time to digest the whole thread but the Flanking bonus only applies to attacks with the Commander leader. The AI will take promotions and traits into consideration as it’s a simple calculation for the combat damage preview (which the human and AI see in the same way). Will the AI position a unit in anticipation of using Flanking later? Will have to ask Alex about that although it certainly will use the bonus if the opportunity arises.
Given that you guys made the game, I’m not the least bit surprised. Thanks for the confirmation. And stop reading Old World threads while you’re on vacation!
ah! I built the hAgia sophia, which requires a holy city, in the city that founded my paganism But it didn’t count for the ambition of having a holy site. I wonder if it would have counted if I had adopted paganism in this way?
Isn’t a holy site something like the Temple of Solomon (the things a Disciple can build once in the city where a religion was founded) whereas the Hagia Sophia is a Wonder?
This is the right answer. A Holy City is a city where the religion was founded (including paganism), a Holy Site is one of the four religion-specific wonders that Disciples build.
that makes sense, dont know why i was confused. i ended up winning with a double score victory by getting a bonus science point by only defending until turn 120, then taking over Rome when theyunwisely declared war, and getting a bonus science victory point every 2 years for the last 10 (I think) years. This made me happy as it seemed a very “Babylon-like” way to win; I wanted the game to have more of a builder feel than my usual Rome or Greece game. The topography of the map helped a lot, one strong neighbor was blocked by mountains and could only invade by sea.
The other thing I think about this terrific game, is that it’s really wonderfully fighty. Not just because the AI puts up stiff resistance. The balance between building/developing cities and impaling Romans with sticks is just excellent. You need the one for the other etc but it rarely seems out of balance where you’re only developing cities whilst the AI fiddles.
Currently hacking and slashing my way towards victory and having to contest every yard of ground. There’s a really great tight relationship between what you do in your cities with your workers and how many centurions’ heads you get to nail to your city walls.
I have definitely encountered longeurs towards the end of various double victories where economically etc am miles ahead but right now playing as Assyrians its just glorious slaughter all the way. Marvellous. Barbarians are set on some high level and they never let up.
My goodness, can Persia’s unique ranged cavalry do some silly things. They hit like trucks, and ranged rout is sometimes just an absolute beating.
Finished crushing my second game, this time on baby (the Good) difficulty instead of tiny baby difficulty. Learned a lot w/r/t tech ordering and what’s super important, how to manage relationships, and city specialization. Looking forward to applying it to the next difficulty up. Still just absolutely loving the game.
In my current game, I am at Peace with Rome and have been consistently sending my caravans to them (btw, that paid off really well in terms of gold, so I’d highly recommend caravans!). In my current 85 years into the game, I see a lot of Roman armies surrounding the borders of my two cities. Either they are planning a hostile takeover but unable to do so because of the Peace pact between us or the AI is stuck in an infinite loop and does not know what to do. They should be declaring war on my neighbour, the Assyria. This made me nervous and I am investing all of my orders to build an army for an impending attack instead of working on something else.
I have something similar going on with the Assyrians in my current game. They’re in solid second place behind me, don’t actually like me much (I think they’re like -50 after their old prickly asshole king finally died), and have a big pile of units sitting near me but don’t seem interested in invading.
They might be kinda confused because I’m between them and the Greeks and Egyptians, at least one of which they’re probably at war with.
Me, I’m ensuring that King Fuckstick of Greece and his crazy massive stats (seriously he’s like 8/9/3/4) goes to his grave bitter and defeated after he foolishly seized a pretext for a war of revenge after losing two Greek colonies to the Darius the Great in his youth.
(Spoilers: King Fuckstick lived just long enough to sign an embarrassing peace treaty. Bwaha.)
I just figured out that force-marching a unit actually works like this, where for some reason I thought it was basically “you can double move this dude this turn.”
You pay 100 training.
Unit ignores fatigue limit until end of turn.
Moving (but not cooldown actions) take 2 orders instead of 1 for that unit this turn.
So if you have enough orders – say your leader is an Orator and you’re playing Persia – you can absolutely hump armies across freaking continents in a single turn. Good times, good times.
Force moving is so tremendously disruptive to the conventions of these games that it might have destroyed other 4Xs for me. I mean, it wouldn’t work without Old World’s orders system, and the way training is a resource, and maps built to be wide open, and the unit balance, and the fatigue model, and a dozen other variables Mohawk smartly designed into the game. But otherwise, every game should have force marching like this!
I tried a game where I actually emphasized military tech instead of economic boom and it is fun how you can bully the AI. It’s still too hard to capture cities unless you have many units, or siege weapons (I think AI doesn’t do well against siege). Camp + grind down the AI when you get siege, works well. But to be fair maybe it’s one of the scenarios where a human wouldn’t have an answer either.
There are definitely some optimization fruit that Mohawk could pick. I started noticing performance degradation around turn 90 on my monster beast (less than a year old, 3070, etc, etc).
To be fair I have everything cranked, but I’m relatively confident that this machine should be able to play Old World at 1440p with everything on without issues.
Maybe some or all of the improvements involve high enough definition models/maps/shadows that get expensive with a bunch of them onscreen? That’s the biggest difference in terms of what’s being displayed lategame vs early, anyway. But I’m no graphicsologist; I’ll leave that to the pros.
