The "siege engine" phase of sieges works well in some ways (the ability to starve out the garrison, while hoping reinforcements don't arrive, or rush in and just use ladders because the defenders are weak), but it's mostly a broken mini-game that needs lots of improvement.
1. The best way (that I've found) to do serious sieges is the "queue up 4 trebuchet builds, immediately put them in reserve as they finish, and then pop all 4 out at the same time." This is bad for many reasons:
a. it works very well every time, and so every siege is the same. There's no drama, there's no game - it's a puzzle that you solve once. (On a side note, it's maddening that trebuchets occasionally attack the walls while defending siege engines are active, but it ends up not really mattering too much)
b. the AI DOESN'T do sieges this way, so the player essentially becomes a siege-god who can destroy walls and remove defending siege engines with ease in any battle, while the AI will stagnate in most serious siege contests (which is the more realistic situation, but see below).
c. if the player hasn't figured out this puzzle, then sieges are just frustrating affairs of "build an onager, it gets whacked," over and over, "I guess I'll just make some towers and a ram and slog it out."
The solution: The player and the AI should be able to just build the siege engines in the reserve state, and then "move them up/attack" altogether when they're ready to bombard. There should be a whole reserve area at the back of the siege camp (basically where the rams and towers are currently placed) where you can choose to build ranged engines if you want, and when you later choose to attack with them, it doesn't instantly teleport them, but rather there is a little progress bar (maybe even visible little guys) that goes while they're moved to firing range.
To counter/balance the increased effectiveness of the attacking armies with this new system, defending siege engines would get some advantages:
1. The height advantage of defending siege engines simply makes them more accurate, OR it allows them to get a hit or two in on the attacking siege engines before they are moved into effective range - something along these lines. Similarly, you could argue that defending siege engines are simply harder to hit and take out because they have the advantage of their walled/entrenched positions.
2. Defenders should start with siege engines already built. This just makes sense?
3. Wall level could play a more important role in all of these features - level 1 walls would not even support trebuchets, for example, while level 3 walls could even have a 5th siege engine slot (although I realize this would probably require modifying the battle maps, but you could just skip that if you wanted).
The goal would be a siege phase that involves the attackers building up their engines (assuming they don't just rush to a tower/ram strat), and then advancing them into position and battering down the defender's position. A well defended position, with enough men, on level 2 or 3 walls, will almost certainly beat back that 1st attack, but their emplacements will be weakened. Attackers would have the choice of targeting enemy siege engines, or just focusing on punching a hole in a wall. Then it comes down to numbers - a large attacking army will be able to build new siege engines quickly, a commander with lots of engineering skill will get additional advantages, etc.
Honestly, if all of this is too complicated, then I think a better solution than the current system would be to just have it all be an automated process that compares army and commander stats and produces an outcome over time. Put the player on the same playing field as the AI, give them a few choices to make about the goals of the siege, and then tweak the system until you get the outcomes you want. The current system just feels like exploitation or even cheating - it's not the immersive, visceral experience that makes this game so amazing.
1. The best way (that I've found) to do serious sieges is the "queue up 4 trebuchet builds, immediately put them in reserve as they finish, and then pop all 4 out at the same time." This is bad for many reasons:
a. it works very well every time, and so every siege is the same. There's no drama, there's no game - it's a puzzle that you solve once. (On a side note, it's maddening that trebuchets occasionally attack the walls while defending siege engines are active, but it ends up not really mattering too much)
b. the AI DOESN'T do sieges this way, so the player essentially becomes a siege-god who can destroy walls and remove defending siege engines with ease in any battle, while the AI will stagnate in most serious siege contests (which is the more realistic situation, but see below).
c. if the player hasn't figured out this puzzle, then sieges are just frustrating affairs of "build an onager, it gets whacked," over and over, "I guess I'll just make some towers and a ram and slog it out."
The solution: The player and the AI should be able to just build the siege engines in the reserve state, and then "move them up/attack" altogether when they're ready to bombard. There should be a whole reserve area at the back of the siege camp (basically where the rams and towers are currently placed) where you can choose to build ranged engines if you want, and when you later choose to attack with them, it doesn't instantly teleport them, but rather there is a little progress bar (maybe even visible little guys) that goes while they're moved to firing range.
To counter/balance the increased effectiveness of the attacking armies with this new system, defending siege engines would get some advantages:
1. The height advantage of defending siege engines simply makes them more accurate, OR it allows them to get a hit or two in on the attacking siege engines before they are moved into effective range - something along these lines. Similarly, you could argue that defending siege engines are simply harder to hit and take out because they have the advantage of their walled/entrenched positions.
2. Defenders should start with siege engines already built. This just makes sense?
3. Wall level could play a more important role in all of these features - level 1 walls would not even support trebuchets, for example, while level 3 walls could even have a 5th siege engine slot (although I realize this would probably require modifying the battle maps, but you could just skip that if you wanted).
The goal would be a siege phase that involves the attackers building up their engines (assuming they don't just rush to a tower/ram strat), and then advancing them into position and battering down the defender's position. A well defended position, with enough men, on level 2 or 3 walls, will almost certainly beat back that 1st attack, but their emplacements will be weakened. Attackers would have the choice of targeting enemy siege engines, or just focusing on punching a hole in a wall. Then it comes down to numbers - a large attacking army will be able to build new siege engines quickly, a commander with lots of engineering skill will get additional advantages, etc.
Honestly, if all of this is too complicated, then I think a better solution than the current system would be to just have it all be an automated process that compares army and commander stats and produces an outcome over time. Put the player on the same playing field as the AI, give them a few choices to make about the goals of the siege, and then tweak the system until you get the outcomes you want. The current system just feels like exploitation or even cheating - it's not the immersive, visceral experience that makes this game so amazing.
