AI lords should enter in settlements to defend when the settlements are under siege

[Blood Gryphon]

Right now the AI lords do not enter settlements with the intent to defend sieges, it's not a choice that they have right now. But there are many situations that I feel the logical choice for the AI is to enter a settlement to defend the impending siege. For example:

Raganvad's army has been traveling to besiege Caleus Castle, the 500 vlandian army has been keeping close to the sturgian army since Nevyansk Castle, but the sturgian army's intent was always to siege Caleus Castle. You can see that the Vlandian army did not enter the castle to defend and opted to keep hovering around the sturgians despite being outnumber 2 to 1. Personally I think this is the exact situation that a commander would choose to enter the castle to attempt a defense. I also think that sieges would be a lot more fun if they weren't so lop sided.

These kind of sieges are super boring, but end up being half of the fights because border castles trade so often and never have time to build up a militia, due to lack of actual defense by lords. Towns are fine because they typically have 200-400 troops defending them and it's a lot of fun, but castles are always end up being small and boring fights.

I understand that making the AI sacrifice troops to defend a siege could be exploitable, so maybe make it so they only lose troops once a siege camp is completely built? That way you can't just spam sieges. I'm fine with forces closer in size waiting outside the castle, but when they don't stand a chance (greater than 2v1) their best chances are inside a settlement. This also means to successfully siege there will likely need to be a previous land battle that was fought and won by the attacker so that there aren't a bunch of defenders around. That seems like a more logical set of options, either forces similar in sizes fight a battle and the winner gets to siege with less defenders or an attacker shows up to siege against a smaller force that decides its best chances are in the settlement thus evening out the fight.

What do you guys think?

 

[Dr-Shinobi]

Agree that they need to fix the AI regarding that. Problem with most sieges is that they arent that interesting both being the attacker or defending it since the attackers are almost always like 3-1 and having troops enough to siege it or they leave it be. In other terms not as exciting.
The only exiting part strategically is if we save casualties enough to manage to defend it against the outer forces.
However a good defence usually takes 1-2 (depending on the total troop value) to have a chance to be able to fend of the attackers and to be something interesting as well.. so its an hard balance part of which decision you would make. So with that said the AI would perhaps be able to make that fight and would try defending it.

Also another problem is that after lets say AI command takes over the settlement theres an animation that they march in to it then the enemy attack em instead of besieging em inside. Something that has been posted before and debated as another thing that should be fixed. Just sliding it in as an reminder

 

[NPC99]

I agree the local army should defend the walls of its settlement, if it has a reasonable prospect of winning. I wouldn't want it to enter a death trap. If there is no prospect of the settlement holding out despite their added numbers or if insufficient food was stored for a viable defence, the army should hover outside waiting a more favourable opportunity.

 

[AfLIcTeD]

They should also enter settlements to take cover from stronger parties.

 

[mexxico]

Yes if we make AI smarter they should enter settlement before siege starts (in this scenario Vlandian army leader should make cheat and see that army is going to start a siege (maybe not a cheat, player can see this info by mouse over)).

Here problem is if Vlandian 500 men army enters settlement before siege starts then your army leader Raganvad will change target. Because in autocalculation they cannot take a castle if there are more than 0.3-0.4-0.5x of their forces (wall level also important). So you will be bored more because army you are in will change target and travel to another target. Maybe we can make armies with player party more courageous then they can start siege even there are 500 men lord parties inside. I am ok these small tricks but it is generally not accepted here.

I know most of the sieges are similar (attackers are so powerful compared to defenders) currently and because of this we cannot use game's potential fully, there are usually no lords in defensive side and only garrison and militia is at defensive side of siege. However if settlement have lots of defenders (+lord parties) its obvious attackers cannot win that siege without player interaction. So it is also non sense making a siege. Anyway thanks for post, I will think how we can improve these scenarios.

 

[Tryvenyal]

As suggested above, creating a siege camp could make an army rather inflexible and allow for enemy lords to slip though into the town and boost defence. An defending armyleader could consider defending a castle at odds 0.6 - 0.7 as favourable enough. A brave attacker could do the same recalculated consideration and not break the camp - especially not if player is in army(attacking or defending!) to add up for a epic siegebattle! At 0.8 they should break siege immediatly though!

 

[froggyluv]

We need more illustrative factors than just a simple math calculations to give us more narrative to sieges. The greatest battles in History are generally outnumbered troops who for whatever reason - tactical usage of terrain, master siegeweapon defense, traps, fierce religious/nationalistic fervor/a Hero in their midst etc thrive and defeat a far larger numerical enemy. This is what makes reading up on historic battle fun. Theres nothing interesting about reading the romans decided to siege the Barbarian hordes fort because they had a 5:2 numerical advantage and won. Obviously numerical advantage NEEDS to be at the foundation of calculations but what exactly is making any of this an interesting sandbox story that invigorates the players interest in the other intangibles of battle?

 

[Blood Gryphon]

Yes if we make AI smarter they should enter settlement before siege starts (in this scenario Vlandian army leader should make cheat and see that army is going to start a siege (maybe not a cheat, player can see this info by mouse over)).

Here problem is if Vlandian 500 men army enters settlement before siege starts then your army leader Raganvad will change target. Because in autocalculation they cannot take a castle if there are more than 0.3-0.4-0.5x of their forces (wall level also important). So you will be bored more because army you are in will change target and travel to another target. Maybe we can make armies with player party more courageous then they can start siege even there are 500 men lord parties inside. I am ok these small tricks but it is generally not accepted here.

I know most of the sieges are similar (attackers are so powerful compared to defenders) currently and because of this we cannot use game's potential fully, there are usually no lords in defensive side and only garrison and militia is at defensive side of siege. However if settlement have lots of defenders (+lord parties) its obvious attackers cannot win that siege without player interaction. So it is also non sense making a siege. Anyway thanks for post, I will think how we can improve these scenarios.

Great points.

But just because at that moment they would lose an actual siege, as long as they would win a open field battle I think they should stay until the defenders numbers have dwindled (so above 0.5x forces and below lets say 0.8x forces they stay, but above that they GTFO) . The attacking AI army could try to starve them out, along with bombardment (we need a way to bombard the town when the walls are down), the AI could just stall until the castle defender numbers is within the range for them to have a decent chance of success. Also giving the defenders enough time to maybe raise another army to help if it isn't preoccupied. This could seriously help factions like sturgia who have to cross the entire map to defend castles sometimes.

To help with these long sieges, maybe you could develop a way for AI armies to send out their smallest AI party (or fastest) to try and get more food for them (or offer the player a quest to do it if they are in the army, or let the player do it as well). Also would be cool if they sent a party out on scout missions to get vision of enemies and castles then rejoin the army. These kinds of features would lead to cool moments where you could capture/take out an enemy armies scout or food supply, or even devastatingly lose yours.

Also on a side note, I think the AI armies need to act more defensively when their kingdom strength is lower (making them the weaker faction). Wars right now feel like two attackers who defend if they happen to be near by and not already on their way to siege something. Weaker faction armies should patrol and focus on fighting enemy lords/armies (before the enemy sieges something) to try and even out the faction strengths and then once evening out the strength they switch to sieging the enemy faction. It hurts to see armies of the weaker faction walk by losing battles in their territory to go siege a far off target just because that target was recently taken and has no defenders in it.

 

[Anushtegin]

But just because at that moment they would lose an actual siege, as long as they would win a open field battle I think they should stay until the defenders numbers have dwindled (so above 0.5x forces and below lets say 0.8x forces they stay, but above that they GTFO) . The attacking AI army could try to starve them out, along with bombardment (we need a way to bombard the town when the walls are down), the AI could just stall until the castle defender numbers is within the range for them to have a decent chance of success. Also giving the defenders enough time to maybe raise another army to help if

pretty much this. while a larger garrison has more men, it would also run out of supplies quicker under siege.

 

[Apocal]

(we need a way to bombard the town when the walls are down)

The defenders take losses from their siege engines being targeted by trebuchets. I'm not sure if it is the hits or misses doing it, but the effect is the same either way.

 

[Knyghte]

I agree as well.

Let them lose people as they enter, maybe a truncated fight with the siege. Set the AI to have a large portion charge the gates to get in, the rest hold back to stage a front repelling the siege troops.

The AI for siege combat on both sides needs a lot of attention. I hate seeing 80% of my troops at the base of a single ladder on a siege tower while two other ladders on that tower are empty. Defending troops need to be climbing the walls before the ladders or towers get into position. Siege engines (catapults, ballistas, etc) needs have more fine control, able to strike large groups of troops at varying ranges and the loaders need to stay at the engine and not grab their bow and charge forward between loads.

I would also say that (and I will start another thread for this) I hate laying siege to a castle or town, taking it, and then NOT being in the list to be awarded the fortifications I lead an assault on.

 

[aggravatedAardvark]

Agree will pretty much all of this. AI parties should enter settlements that are about to be sieged to help defend (unless too weak to make a difference or not enough food). The besieger should consider starving the defenders out as a viable tactic instead of switching targets, and send raiding parties out to scavenge for more food for themselves (unless there's a lot of enemies nearby). Also other good suggestions.

Also on a side note, I think the AI armies need to act more defensively when their kingdom strength is lower (making them the weaker faction). Wars right now feel like two attackers who defend if they happen to be near by and not already on their way to siege something. Weaker faction armies should patrol and focus on fighting enemy lords/armies (before the enemy sieges something) to try and even out the faction strengths and then once evening out the strength they switch to sieging the enemy faction. It hurts to see armies of the weaker faction walk by losing battles in their territory to go siege a far off target just because that target was recently taken and has no defenders in it.

Kind of on this topic of patrolling and focusing on field battles, I do like that in your OP the 500 Vlandian army was deliberately following the Sturgian army (if I'm reading that correctly). Maybe this behavior could be expanded and given a higher priority? For individual lord parties as well, if they don't already do that.

Basically I'm suggesting that if a party (army or individual lord) spots a stronger, slower enemy party in their territory, they should follow that party for a while instead of just fleeing or ignoring them (I mean really, would you just ignore a big group of enemies in your territory?). This would make it much more likely that two or more small parties are in the same place and can combine forces to take down a large enemy in their territory. This would also make it so that diving deep into enemy territory becomes quite dangerous unless you're a small, mobile party that can outrun most others. Or just too massive to be dealt with.

As the player too, this would make entering enemy territory feel much more dangerous and exciting I think. Because once you pick up a tail that's faster than you, you better GTFO or hope you don't run into another enemy. Could add a lot to gameplay.

Having small mobile parties collect around and then swarm stronger enemy parties in the field could also give weaker factions a better way to deal with invaders. Combine this with entering settlements to defend and you might have a decent basis for a defensive AI. (Would probably also need a few other considerations like prioritizing food production and stockpiling, etc.)