Static AI positions on walls and towers, Dynamic AI.

Users who are viewing this thread

Aurex

Regular
WB
Hello there.
Given we probably all agree sieges are a big part of the game (or should be) and have never been working properly, and having read many great suggestions on how they could be made more enjoyable, in-depth and immersive, I think I might have had a half-decent idea I'd like to share.

Most of the complaints (mine included) are about the AI and how it reacts (or fails to react) to developing situations, how there's only one phase in any given siege even as the maps (which are quite great, imho) were clearly made with fallback positions in mind, and about the general wonkiness of every single mechanic involved.
Since the AI is not particularly sophisticated and never will be (I fear) why not using a stopgap measure which might make a lot of it much better?

My idea is this: have an internal AI logic which ticks every 10 seconds or so, updating according to how the battle is going (aka: where allied and enemy troops are, how many there are and where they're going) and threat levels (imminent breaches, troops being overwhelmed, morale breaking) and automatically changing the single units' behavior, sending them to pre-determined position according to the developing situation. This, I think, would fix what I believe to be the worst offending issue in sieges: the completely braindead archer AI, leading to difficult defensive sieges and far too many avoidable losses even when micromanaging without however making the AI too performance hungry given the AI checks wouldn't be run every second.

For example, here's your typical siege scenario (forgive me for how grade school it looks, didn't have much time to make something more acceptable today in 4k PaintHD:

BANsiege1.png


The static positions would be present all over the map, of course. They are usually empty and require no checks, acting as precise (or loosely defined if and where needs be) points where single or multiple troops could rally when their AI checks the situation and decides it's time to move to another location given how the battle is going.

In this case the archers, having a clear defined position from where to shoot (and in which direction) would not be idle 90% of the length of the battle. A single check every ten seconds in case no suitable target was found would be enough for a single archer to decide it's better to move to a static position from where he can and probably will hit an enemy soldier.

As the siege tower becomes an active breakthrough point and enemy troops swarm in, the reserve infantry (in reserve since one cannot tell exactly when a breach will occur, especially with siege weapons being present) would send an appropriate number of men to defend from the attacking enemy on the walls, while the archer troops (which would be sacrificial lambs in melee and not able to use their abilities in an appropriate manner) would retreat to other, more defensible static positions on the towers, shooting the infantry *and* the archers as they press on. Already present archers at those positions would adjust as needs be.

BANsiege2.png


In this second phase the breakthrough would have occurred, and the defensive AI would calculate quantity, position and number of the enemy attackers, deeming them overwhelming to deal with in their present static positions. The infantry would split and retreat to easily defensible static locations, allied (defensive) archers would shoot towards the stairs at the enemies' flank *and* towards the advancing enemy archers. In this phase the battle would still hinge on the first defensive layers, although they'd be partially taken over by attackers.

BANsiege3.png

Third phase: the attackers have overwhelmed the defenders on the walls. The attacking archers take some of the same static positions the defending archers took before (while occupying new ones already there for such an event) and shoot towards the second layer of defense to cover the melee attackers' advance: see that as a smaller, wooden palisade. The defenders occupy new static positions to mount a last stand.

Grade-school cutting edge graphics aside, I think using static positions (with a small area around them allowing the unit occupying it to act realistically, for example a crossbowman would reload behind a fortification, peek out to shoot only to repeat until necessary or possible) and a 10-seconds tick to check whereas a single soldier is actually being effective in its role, whichever that might be, would go a long way towards making the battles much more enjoyable, realistic, challenging (probably longer, too) and give the player more time to formulate strategies which could use the existence of different static points "groups" (archer tower 1, archer tower 2, breach 1, fallback 1 and so on) to quickly assign different formations to specific roles in the ongoing battle.

What do you think about the idea? I'd love to hear some feedback :smile:
 

Terco_Viejo

Spanish Gifquisition
Grandmaster Knight
Honestly, there is nothing that can be refuted; your argument is logical and from a tactical point of view, plausible.

It's not unreasonable, it (supposed) should be like this (Archers positioned behind the battlements/arrow loops and infantry positioned towards enemy suposed assaulting/landing area), as the defenders are depleted, they should retreat to the inner defensive rings gradually positioning themselves in the bottlenecks and high ground spots in there.

Morale should not suddenly drop and the defending bots should not immediately flee to the keep as performing in native default nowadays.

Defensive positioning main wall ---> if the defence is breached ---> fallback to inner ring of wall/street choke point----> if the defence is breached ----> fallback into the keep (last defence) ----> if the keep falls ---> the attackers capture the place.
 
These "simple" drawing make it easy to get your point across. I agree with your suggestions.
I would like to make an improvement to the first picture. I hope TW will consider this for scenes that are still in development.

This makes getting on the wall only the first step in getting past them.
I would not make sense to give attacker acces to the streets behind te wall without forcing them to capture the defensible towers.
This way de defender can line te street and the towers with range troops and really punish the attacker.
A handful of melee units should try to block the small acces to the towers. preferable with a destructible door.

i propose that stairs should be limited to the towers. At least in some scenes.
 
Last edited:

Aurex

Regular
WB
These "simple" drawing make it easy to get your point across. I agree with your suggestions.
I would like to make an improvement to the first picture. I hope TW will consider this for scenes that are still in development.

This makes getting on the wall only the first step in getting past them.
I would not make sense to give attacker acces to the streets behind te wall without forcing them to capture the defensible towers.
This way de defender can line te street and the towers with range troops and really punish the attacker.
A handful of melee units should try to block the small acces to the towers. preferable with a destructible door.

i propose that stairs should be limited to the towers. At least in some scenes.
I must admit I didn't draw the destructible doors and the other tower out of pure laziness, but yeah, they'd be there to slow everything down a little. The sieges are very short (after the besiegers have reached the walls) and tendentially chaotic as they are, destructible doors would give the player a little time to reorganize the troops and came out with a new plan. I find it silly how the archers act (or don't act...) during a siege. They basically stare at a single point until 5 minutes have passed and there's a guy slashing their back with a sword... it's kinda sad.

Also yes, the stairs! I agree that having archers shooting at the attacker from the side might be *far* too punishing for the attackers, unless they're already in a huge numerical disadvantage. I'd also prefer if they were limited to specific scenes, it'd make the sieges much more diverse and enjoyable, imho. Thanks for the feedback!
 
Also yes, the stairs! I agree that having archers shooting at the attacker from the side might be *far* too punishing for the attackers, unless they're already in a huge numerical disadvantage
Haha I actually meant that I would like it to be that punishing for the attacker. As you mentioned defensive archers don’t shoot enough en only react when you hit them. I found archers hiding at the back of towers, not doing anything. The RTS mod is great for surveying the battlefield. At the moment tou can win a siege with equal numbers if you have some quality units.
 

Aurex

Regular
WB
Haha I actually meant that I would like it to be that punishing for the attacker. As you mentioned defensive archers don’t shoot enough en only react when you hit them. I found archers hiding at the back of towers, not doing anything. The RTS mod is great for surveying the battlefield. At the moment tou can win a siege with equal numbers if you have some quality units.
Hah, sorry about that! This is what happens when I don't get enough sleep.
Yeah, I'm quite saddened by how easy it is for attackers to successfully siege a more experience, better-equipped force defending a strong castle when the player doesn't intervene by throwing literal bombs on the soldiers standing behind the siege towers. 1000 (mostly) peasants and militia assaulting a castle defended by 150 elite troops should be hardly able to win, and if they do they should be so badly bloodied they should basically leave it unguarded for at least a couple of weeks for the previous owner to retake it.
I mean, historically... sieges were a BLOODY business even when the attacker managed to create multiple breaches *and* was in a 10 to 1 numerical superiority. There are so, so many examples of small, elite forces holding a city or a castle (or a citadel) for years while under siege by a hugely superior (numerically speaking) force and eventually winning the siege when the enemy abandons it due to lack of supplies. Yet in Bannerlord, if you're defending 150 vs 800 it doesn't even matter if your troops are all top-tier and you're a one-man army, you will be swarmed and brought down by literal peasants. That's why I came up with the fixed positions idea, so the AI (which is kinda stupid) might at least behave somehow decently and close the gaps.
 

Mandoble2

Recruit
If we look at the history, there were well commanded defensive forces and quite the opposite as well. If something like "smart" site defense is implemented, it should be "linked" to some level of the site's leader in leadership/tactical.
 

Aurex

Regular
WB
If we look at the history, there were well commanded defensive forces and quite the opposite as well. If something like "smart" site defense is implemented, it should be "linked" to some level of the site's leader in leadership/tactical.
Yes, that would make a lot of sense. Example, you could leave a captain of the guard in charge of the defense of a settlement, there might be quests to give him training, or even equipment. The better he becomes, the better the troops will act. But if you appoint as captain a merchant with no experience whatsoever, the troops will have low morale and be really uncoordinated.
 
How about they make it to where the siege towers are taller than the wall and the ramp drops on the parapet so the defending AI cant rush the tower.....
 
Top Bottom