Could the AI please stop cheating on sieges

[RobRoy3]

you see the AI sieging forts with less than n x1000 men all the time.

You see the AI do it, occasionally, but not often enough. As a player, I do it much more frequently than the AI since, as Stolen Rutters points out, it stops all reinforcement/build progress. It often makes sense to carpet siege every province with a minimal force, even if that only leaves you with a few provinces where your siege is actually progressing.

Frankly, the opposite AI behavior is much more common and much more problematic. The AI regularly leaves a far larger army in the province than it needs to make the siege progress. Presumably they are taking attition and wasting manpower. Always entertaining when your enemy is doing something stupid, but a weakness in the game, nonetheless.

 

[TargaryenBlood]

Why would it depend on garrison strength? If they're not assaulting, they're simply preventing supplies from reaching the fort while they stand off and use artillery to try to knock holes in the wall. The number of men needed to effectively interdict resupply is going to depend on the fortifications, not the guys manning them. I always assumed the required siege army was what was needed to invest the fortress by circumvallation, since that's the model that actually makes sense.

Garrison strength matters because that is how many active duty armed and trained soldiers are inside the fort. If you have less, then they can just sally forth and kick your assaulting group's face in. A fort's garrison usually has some manner of siege weapons inside, so there must be a larger number of troops outside the wall than inside. Basically, if you are coming up on a tier 5 fort with 3 guys in it, there is nothing those 3 can do. If you have 5,000 versus 5,000 then the garrison can still use their forces to poke holes in the besieging force's lines to get more supplies.

 

[BFTeixeira]

Every time I've sieged a fort, I've needed n x 1000 men for it to progress. And that made sense. It's not like the game has useful UI or tooltips for 99% of the rules. I don't think I've ever even thought to run a single unit out to siege a fort unless I was carpet sieging somewhere there were level 1 forts.

The only other paradox game I've played is Vicky 2, which has worse UI if anything, and have not played it much. Not sure why 'the mechanic is used in other paradox games' is a reason why I should know that.

BFTeixeira saying he's never seen anything like that (successfully sieging with less than nx1000 men) doesn't suggest 'no, it's actually based on a totally different number that only relates to that one when the fort is at full strength', but rather that he's never seen an AI sieging a fortress with less than nx1000 men. So you'll forgive me my incredulity when he says that, since you see the AI sieging forts with less than n x1000 men all the time.

Sorry, but i didn't said that. I said that i never saw that happen in any of my games, where "that" = "Could the AI please stop cheating on sieges"

Just like i said in a previous post... Assumption is the mother of all f*ups. Stop making assumptions, and you'll stop f*ing things up, ok?

 

[TheMeInTeam]

You see the AI do it, occasionally, but not often enough. As a player, I do it much more frequently than the AI since, as Stolen Rutters points out, it stops all reinforcement/build progress. It often makes sense to carpet siege every province with a minimal force, even if that only leaves you with a few provinces where your siege is actually progressing.

Frankly, the opposite AI behavior is much more common and much more problematic. The AI regularly leaves a far larger army in the province than it needs to make the siege progress. Presumably they are taking attition and wasting manpower. Always entertaining when your enemy is doing something stupid, but a weakness in the game, nonetheless.

Want some fun? Split off a siege stack that is minimal and siege the AI while it sieges you. It will siege you briefly, but then evaluate that it can easily kill your siege stack and lift its siege to attack. When it does so, threaten to reinforce this stack, such that the AI thinks it can't win because of the reinforcements. It will go back to sieging you instead.

Obviously, if you repeat that ad nauseum you can easily (if slowly) siege out an AI while it siege races you back...except it's running the race in the wrong direction. This is particularly nasty to use on it in heavy mountain areas, where you can't kick it off your land easily, but it's also helpful in areas with a lot of attrition.

 

[ChildeR]

I tagged the OP helpful, because it's a good lesson on jumping to conclusions.

Protip: if you think there's something wrong with a feature, you may want to check out what the wiki says about it, so you can be fairly sure you understand how it ought to work.

 

[GoldenBear10]

I tagged the OP helpful, because it's a good lesson on jumping to conclusions.

Protip: if you think there's something wrong with a feature, you may want to check out what the wiki says about it, so you can be fairly sure you understand how it ought to work.

That is why they made the wiki official... because the original manual means nothing at this point (or it won't after this 2.0 patch).

I am surprised we weren't going to talk about how the humans cheat by being able to click on the sally button when the AI cannot.

 

[Squirrelloid]

I tagged the OP helpful, because it's a good lesson on jumping to conclusions.

Protip: if you think there's something wrong with a feature, you may want to check out what the wiki says about it, so you can be fairly sure you understand how it ought to work.

If the game actually told you what the rules were, and had decent UI, it wouldn't be so aggravating. Sure, maybe I assumed something too quickly, but the game gave me no way to check that *inside the game*, and the thing I assumed described observed behavior 99% of the time and made more sense as a model of what the mechanics represented. The real lesson here is the game desperately needs at least passable UI, like, I dunno, an in-game rules manual that is kept up-to-date at a bare minimum.

The wiki is not a solution. It can be wrong, and not just after a patch changes things. It is not maintained by Paradox. Example: The wiki thought PU inheritance chance's size modifier for the junior partner was calculated on base tax for months (at least), when it was actually number of provinces, despite inheritance chance calculations not being changed since *at least* 1.8 (if not longer). And I only caught that and fixed it because the UI at least tells you what the size modifier is, and I could see it had nothing to do with base tax by inspection.

No one is going to go to the wiki unless they're confused. I wasn't confused - it looked like more AI handholding to me.

 

[BFTeixeira]

If the game actually told you what the rules were, and had decent UI, it wouldn't be so aggravating. Sure, maybe I assumed something too quickly, but the game gave me no way to check that *inside the game*, and the thing I assumed described observed behavior 99% of the time and made more sense as a model of what the mechanics represented. The real lesson here is the game desperately needs at least passable UI, like, I dunno, an in-game rules manual that is kept up-to-date at a bare minimum.

The wiki is not a solution. It can be wrong, and not just after a patch changes things. It is not maintained by Paradox. Example: The wiki thought PU inheritance chance's size modifier for the junior partner was calculated on base tax for months (at least), when it was actually number of provinces, despite inheritance chance calculations not being changed since *at least* 1.8 (if not longer). And I only caught that and fixed it because the UI at least tells you what the size modifier is, and I could see it had nothing to do with base tax by inspection.

No one is going to go to the wiki unless they're confused. I wasn't confused - it looked like more AI handholding to me.

Or maybe you just jumped to conclusions without even investigating why would that happen. Or even better, maybe you could post that situation as a question just to check if something was missing in your analysis. But no... i just went on accusing Pdx, accusing other posters, and now you even accuse that your lack of understanding the game is Pdx fault.

Bravo!

 

[Squirrelloid]

Or maybe you just jumped to conclusions without even investigating why would that happen. Or even better, maybe you could post that situation as a question just to check if something was missing in your analysis. But no... i just went on accusing Pdx, accusing other posters, and now you even accuse that your lack of understanding the game is Pdx fault.

Bravo!

Don't blame me for your response being uninterpretable. Seriously, it's not unreasonable to assume you read the OP, and were responding to the specific description, not just the thread title. If you really meant to be understood, 'That's not cheating, it only needs its forces to exceed current manpower' would have been clear. 'I've never seen anything like that' should mean you'd never observed what the OP described, not just the thread title. I suppose I also shouldn't assume people replying to a thread actually read the OP.

But I apologize for accusing you of not playing. That was rude.

And ultimately, it is Paradox's fault. This is the only strategy game I've ever played which doesn't include the rules with the game. That's a serious UI problem, and it is entirely Paradox's fault. If they don't make the rules available, and I therefore don't understand the rules, whose fault is that exactly?

 

[hwoosh]

Don't blame me for your response being uninterpretable. Seriously, it's not unreasonable to assume you read the OP, and were responding to the specific description, not just the thread title. If you really meant to be understood, 'That's not cheating, it only needs its forces to exceed current manpower' would have been clear. 'I've never seen anything like that' should mean you'd never observed what the OP described, not just the thread title. I suppose I also shouldn't assume people replying to a thread actually read the OP.

Actually, the way I read BFTeixeira, he was assuming in good faith you knew how the game worked and were thus describing an actual case where the AI got siege progress despite not having enough men. Both he and I are guilty, I guess, of skimming your OP a bit and not realizing the misconception about fort level that you were laboring under until later.

And ultimately, it is Paradox's fault. This is the only strategy game I've ever played which doesn't include the rules with the game. That's a serious UI problem, and it is entirely Paradox's fault. If they don't make the rules available, and I therefore don't understand the rules, whose fault is that exactly?

Oh come on. EU4 is worlds better than, say, previous Paradox games or (shudder) Dominions. And while, yes, there are many UI inaccuracies, this is hardly the worst offender. For one thing, it's ridiculously easy to test this out and discover the rule in-game, if you show a little curiosity and initiative; I remember learning how this worked myself in the first few hours of learning the game. I think it was from noticing that some siege stacks taking heavy winter attrition would have their siege stacks halted, while others wouldn't; those that didn't were more advanced sieges, whose garrisons had been reduced below 1960 or 1980 or whatever the attrited stacks were at.

A smart and adaptive player's first reaction, when they see the AI (or another player, for that matter) doing something unexpected, is not "Hey, the AI is cheating!"; it is "Huh, I didn't realize you could do that. I should test it out."

 

[Zelius]

Don't blame me for your response being uninterpretable. Seriously, it's not unreasonable to assume you read the OP, and were responding to the specific description, not just the thread title. If you really meant to be understood, 'That's not cheating, it only needs its forces to exceed current manpower' would have been clear. 'I've never seen anything like that' should mean you'd never observed what the OP described, not just the thread title. I suppose I also shouldn't assume people replying to a thread actually read the OP.

But I apologize for accusing you of not playing. That was rude.

And ultimately, it is Paradox's fault. This is the only strategy game I've ever played which doesn't include the rules with the game. That's a serious UI problem, and it is entirely Paradox's fault. If they don't make the rules available, and I therefore don't understand the rules, whose fault is that exactly?

No, my response would be the same, if I may paraphrase: "I have never seen this situation, where I need 2000 men to siege a level 2 fort with 2000 men while the AI can siege the same fort with 1 unit of presumably any size."

The UI does tell you when you can't siege, but admittedly it doesn't give you a tutorial on exactly what you need to do. Which is basically all of Paradox's games.

But really, have you never come across a situation where you sieged a fort with a depleted garrison? Never had one of your provinces sieged and taken it back? Never had rebels anywhere?

 

[BFTeixeira]

Don't blame me for your response being uninterpretable. Seriously, it's not unreasonable to assume you read the OP, and were responding to the specific description, not just the thread title. If you really meant to be understood, 'That's not cheating, it only needs its forces to exceed current manpower' would have been clear. 'I've never seen anything like that' should mean you'd never observed what the OP described, not just the thread title. I suppose I also shouldn't assume people replying to a thread actually read the OP.

But I apologize for accusing you of not playing. That was rude.

And ultimately, it is Paradox's fault. This is the only strategy game I've ever played which doesn't include the rules with the game. That's a serious UI problem, and it is entirely Paradox's fault. If they don't make the rules available, and I therefore don't understand the rules, whose fault is that exactly?

It's just a basic rule of good behavior: ask before accusing someone.

And it's not me you should apologize. You should apologize Pdx for accusing them (again) before having any kind of proofs that they could be wrong. How much trouble would it be for you to expose the situation and ASK what could justify what's happening? You have hundreds of users in this forum that together have dozens of thousands of hours of gameplay. Just ask when you don't understand something, or it seems wrong to you.

 

[Wizzington]

I would suggest taking this as a valuable lesson about jumping to conclusions about the AI.

Also not blaming everyone around you when you're wrong about something. It makes you look childish.

 

[WhiskyGlen]

well what really defeat the purpose of fort is that in most game where you expend, you level 4 fort will be in the middle of your country where NO enemy will ever reach...
fort cannot be moved, extra manpower means more army, means that it can intercept the enemy before it reach your low level fort...
depending on country fortifying some province is worth it... as France I fortify Dauphiné and Ile de france mainly...

I think the idea of forts is to only put them in strategic locations as the manpower benefit is minimal considering all the MIL you waste on building the forts (soon to no longer be an issue). For example, I had just retaken Orkney Islands from England after they held them for 150+ years. I even lost my core on them (I was Sweden). I decided I would do whatever I could to prevent that from happening again as I knew they would be a war goal if England declared war on me again and I knew it would be a long time before I would be navally powerful enough to gain more English provinces. Especially with coalitions left and right. Since England, or rather, Great Britain, had a powerful navy at the time and I was unsure if I could beat them, I upped the fort in Orkney Islands to a five AND built a march there. It worked quite well.

I am surprised we weren't going to talk about how the humans cheat by being able to click on the sally button when the AI cannot.

I am amazed how many times I forget about the sally forth option when I am trying to liberating a besieged province with an army similar to my opponents.. Always forget I have those expendable garrison troops to fight in the fray.

The UI does tell you when you can't siege, but admittedly it doesn't give you a tutorial on exactly what you need to do. Which is basically all of Paradox's games.

No, but it becomes quite obvious after a short time why a siege will sometimes work when you should be short on men if comparing it strictly to the fort level. I remember when I first played EUII, I spent literally four or five hours just staring at the screen and wondering "WTF am I supposed to do?!' The game made no sense and the fun was figuring out out. Although, figuring out the speed up button a bit sooner would have been ideal...

 

[Aries666]

You see the AI do it, occasionally, but not often enough. As a player, I do it much more frequently than the AI since, as Stolen Rutters points out, it stops all reinforcement/build progress. It often makes sense to carpet siege every province with a minimal force, even if that only leaves you with a few provinces where your siege is actually progressing.

Frankly, the opposite AI behavior is much more common and much more problematic. The AI regularly leaves a far larger army in the province than it needs to make the siege progress. Presumably they are taking attition and wasting manpower. Always entertaining when your enemy is doing something stupid, but a weakness in the game, nonetheless.

This is something well worth doing when fighting a superior enemy who after having wiped their army has the capacity to recruit far more men than you could. A good example is the Najd start, it is easy to go over FL and beat someone like Oman but they have 7 provinces and would quickly become a problem again, better to put 1k on 5 provinces 2k on the other 2 and shut them down for good. This is why the AI needs to learn how to sortie as this wouldn't work against a human.

 

[TheMeInTeam]

Or maybe you just jumped to conclusions without even investigating why would that happen. Or even better, maybe you could post that situation as a question just to check if something was missing in your analysis. But no... i just went on accusing Pdx, accusing other posters, and now you even accuse that your lack of understanding the game is Pdx fault.

Bravo!

Actually, it is PI's fault to a degree. If you want to, through the game's interface, learn what any number of dozens of important rules are...you can't do it. It has improved in recent patches, but a distressingly large number of the game's rules wind up being learned via "trial and error gameplay"...in a strategy title!

What factors into an assault? Nope.
What's your combat width? Can't check that without going into battle or memorizing your base width to sum with tech.
Why can't you CTA your ally at war? Oh, because one of you happens to be in a war with a minor participant and OBVIOUS LOGIC states that you then couldn't have them in both wars!
Back when terrain % were a thing, what influenced those? PI told us half a year later but not in game then took it out.
What are the rules for when rebels change your religion or just have other penalties, without memorizing it/looking it up in external UI?
Which peace deals actually cost diplomatic points, and which ones don't? When can you trust the UI in this regard?
What is the formula for gaining army tradition? Naval tradition?
What kind of peace deal is required to stab hit nations, and why does it change? Which patch notes described when this mechanic was changed?

These are just a sampling, some old, many still relevant. While OP certainly jumped to conclusions about the AI in a semi-inappropriate fashion here, it's hard to fault him. We have historical examples of bugs allowing the AI to cheat inadvertently (like stacking attrition not counting, that somehow didn't apply for humans), gobs of examples where there's no ready way to know the rules of the game you're playing, and even two non-confirmation bias cases of outright wrong percentage displays that have since been patched out (claim fabrication percentage, defensive terrain %...latter removed entirely and the former made accurate as far as testing can tell).

I would suggest taking this as a valuable lesson about jumping to conclusions about the AI.

Also not blaming everyone around you when you're wrong about something. It makes you look childish.

It might be best not to go pointing fingers too much in this regard .

 

[TheChrisD]

The detach-siege-force button always detaches too many units after cannon because it's not counting the cannon regiment's men, so I've pretty much stopped using it.

Detach siege tries to detach exactly the number of men needed, plus one cannon regiment. However, it always detaches regiments at their current strength rather than their full strength, so if your stack has taken a 1% attrition hit from arriving in that province to start the siege, then the detach button will generally pull off one more regiment of both than is necessary (usually it pulls e.g. 3 inf and 2 cannon for 2970 inf and 1980 cannon to siege a Fort 2, or something). Ideally consolidate before you use the detach button, or do it manually.