Defending archers do not do enough damage in siege.

[nikolai1962]

build at least one or two siege towers when possible, the tower and archer AI has a fetish for firing at them, and it makes it much more likely you rams will make it

using siege towers as decoys for your rams works well

[Ciaran]

Exactly. Besides, the most important thing is to take the outer wall, you can reach the inner ones from it.

And besides, a castle assault should be a risky and bloody venture. I´m certainly no expert, but I suppose the whole point to build a castle in the first place is to hold off a large number of enemies with rather few men.

[Alexander the Pretty Good]

If 1.2 addresses this in its final form, it's going to be one heck of a patch.

[Didz]

I usually build two seuge towers and two ladders. I find the ladders actually more effective than the seige towers in scaling the walls but I usually send the towers in first becuase they are slow and then run the ladders to the walls after they have dropped their ramps. So I get to choose points where the wall is poorly defended for my ladder assault, and usually run my crossbowmen onto the walls this way.

Once On the walls I capture the nearest gate to let the rest of the army in and then use the walls to gain access to the inner fortess.

[Daveybaby]

Yeah, ladders are great - up until the point where castles become too big to use them. Catapults are my fave - it does slow your army down and thus means you might take more turns to get to the enemy castle, but if you have catapults it means you can attack the very same turn you get there without having to stop and build rams/towers/ladders.

And while a couple of catapults will take up 2 slots that might otherwise be allocated to infantry, they more than make up for this IMO by taking out large chunks of defending units when the walls collapse. Theyre also invaluable for that bit after you take the gatehouse in cities, where you sometimes have masses of enemy infantry marching down the street from the central square towards you. A couple of catapults with flaming ammo, firing straight down the street works wonders in that situation.

But then I'm confused about the archer-on-battlement situation. Does the computer make adjustments to archer hit rates based on such graphical issues as battlement wall height, and required angle of trajectory??? This obviously COULD be done, but it seems a pretty fine-tooth approach, especially considering other calcs, like giving spears depth of rank bonuses, have apparently been removed from the game.

I imagine that things like battlements and forests (and shields and rocks) give fixed bonuses to any defenders - i.e. each arrow has a % chance to hit the battlements/trees instead of its target, as opposed to individually calculating the trajectory for each arrow and determining what it hit.

Probably for a flight of, say, 90 arrows, the game takes into account unit formations, locations, terrain, armour, accuracy etc etc etc etc and then works out how many of the arrows score 'wounding' hits (e.g. 20), how many score 'killing' hits (e.g. 10), and how many miss (60), then plays a load of arrow firing animations on the archers, then animates the arrows appropriately, then finally plays animations for the target unit - wounding animations on 20 soldiers, death animations on 10 soldiers, and 60 arrows hitting the ground.

[gardibolt]

So far I haven't seen any sign that CA is either aware of or cares about this bug. I wouldn't expect it in the final 1.2, or maybe even in the eventual 1.3.

[John_Longarrow]

DaveyBaby,

I'd be very surprised if CA went to all the work and effort of doing as you suggested. Were I the programmer, I'd just spawn the arrows going at a pre-defined angle with a pre-defined speed from each bow. Each arrow would have an "Affected by gravity" property which would simply be incremented every second. This would tell it by how much to deflect the arrow down. Then simple collision detection would ID who (if anyone) it hit and you can then do the “Did it hurt them” calculation when it hits.

That way you don’t have to worry about updating arrow paths as the target moves and it does explain why units besides the one you’ve targeted take casualties. Much easier than trying to figure this all out ahead of time then hope the player doesn’t do something to mess with your predetermined results.

If you tried pre-calculating the results, you’d have a nightmare of a time if you have more than one unit at the same location, let alone what would happen if the target gets hit by a cavalry charge before your arrows land. Much simpler to do it dynamically.

[Daveybaby]

Believe me, the way i described is far less work for both programmers and the CPU/graphics card. And, there's nothing 'simple' about collision detection - particularly when you have thousands of models, thousands of arrows, and complex terrain to take into account.