Yeah, the siege is somewhat ugly in that there is no reason for why it takes exactly N turns to siege a city. It just happens to be so because, err, we say it is so And I can see it slowing down attacks needlessly (why must I siege 5 turns this city containing one peasant???).

Now that I look at my ideas above they are needlessly complex (as in coding, and also in game mechanism). My favorite from that list is magical units which are powerful in your own land. It is easy to understand, probably simple to implement, allows for faction / magic school / sovereign differentiation and it would fit into the game's world. It is also easy to tune, just change how powerful they are and the attack/defense balance will change.

But, maybe the solution is just making the AI a better defender...

I strongly disagree. It is quite possible to add tons of strategic depth just making small changes such as this. Regardless of how sieges "should" be a wait time for sieges would add lots to the game on the strategic level, and help prevent players from just blobbing. If nothing else a wait time for sieges would allow the AI to counter-mass if you try to steamroll it, since it seems to have trouble with that.

Don't forget there is several levels of fortification improvements for cities, obviously the ability to defend should be tied to the level of fortifications you build in the city. Something like 3-6-9 turns? Also as Horemheb said fortification improvements should give the defending player advantages, like some walls to channel the enemy to one point and some conveniently placed good defense boosting tiles.

Really tough capitals wouldn't really add much to the game. It's not about prolonging defeat but rather rewarding strategy and allowing players to come back, so a single defeat doesn't always snowball into a complete game over. If you have only one city left then you won't be coming back.

Also since we have an unhappiness system it would be weird if citizens couldn't revolt, especially after being conquered. Steamrolling would be much harder if you had to leave garrisons in freshly conquered cities.

I've always thought a way to add epic battles would be to allow tactical battles to occur over multiple turns.  X number of rounds caps how much fighting occurs each game turn turn.  This allows Brad's goal of shorter battles to be met, but still allows battles to take longer and, through the addition of some subsystems, to be more epic. 

Imagine a surprise attack against one of your cities where your goal is imply to survive through X rounds.  During the intervening strategic turn you scramble to get reinforcements on scene while the opponent sends additional troops to join their attacking army.  Over the course of a few seasons the battle builds beyond what either side originally expected and becomes the defining battle of the war.  Both sides know after a few seasons that losing this battle may mean losing the entire war as their army is likely to be decimated and they have little left in the way of reserves.  Everything is committed already.  Or do they keep some troops back as reserves in a fortified area hoping to inflict enough damage in the current battle that they can stave off any further progress by the opposing army? 

Obviously this adds complexity to the balance and AI requirements of tactical battles.  But it also opens up all new possibilities for the interaction between the strategic game and the tactical one.  It would also add more possibilities in the realm of true medieval combat.  Empires could spend heavily on fortifications that would allow them to keep lighter garrisons and be more flexible in their fielded army, but at the cost of greater upkeep and resource investment.  They would have the decision of just how light a garrison could survive long enough for reinforcements to arrive (as one example, there are all sorts of other potential scenarios).

I realize this likely isn't implementable in FE 1.0 (and really belongs on one of those pie in then sky wish lists), but I've always thought it would add depth and grandeur to the tactical battles in the Elemental world.  With the right set of supporting subsystems it also adds whole new levels to the strategic game. 

Frogboy, I'm curious how you steamrolled the AI when all the indicators showed that they were competitive.  What did they do wrong? 

Not trying to stir trouble, just curious what the AI got wrong that caused such a drastic change in the picture so quickly. 

Yeah, the city militia thing has me intrigued.

Militia could be a cool feature. I'd like it if you could design them yourself, but that might be a little complicated. Also if you choose how many you want to have through using level-up bonus buildings.

I'm going to throw my wait behind the siege mechanic idea though. Imagine what GalCiv would be like if you could invade planets without researching planetary invasion or building transports. That's how elemental plays now. While Heavenfall has a point that simply applying a turn limit before you can attack is rather arbitrary (in the the total war games, the siege time is the time it takes the enemy to starve, at the end of which the have to sally out; you have the option to attack the walls earlier if you choose), this feature would be so useful that it would be a benefit to have any kind of siege system, even if it isn't up to the standards of the TW games.

I always thought of magic as a sort of get-out-of-jail-free card. Perhaps what is needed is some sort of "defcon" system where mana consumption:accumulation is decided by how serious a situation the AI believes itself to be in?

Woot! Militia is great news. This will greatly improve the game. 

@DsRaider, What are you talking about? Militias are just a representation of the able bodied men and women of a city taking up arms to defend it. I would imagine they are the pitchfork and torches variety. It is illogical to upgrade them. They are there to prevent early game steamrolls as each city will have a natural defense. This is much better than allowing one unit to to be able to capture and then raze a town just because there are no guard units present. I don't think they can leave a city, but are likely added to the roster when entering tactical combat.

This has been a long time coming. Great decision guys. 

Ya the huge change to unit scaling in FE slipped my mind. They already mentioned a militia random improvement, it was called the belltower. Not sure what it will do now, maybe more militia?

I imagine it will be something like that because they also have mentioned guardian statue improvements. I would be very surprised if it wasn't very easy to create improvements that spawn any creature you want through xml.

One suggestion: gold rush should perhaps cost some pop like it does in AOWII?

Love city militia, helps builder players- and Stardock games are fun as builder games.

I'd like to take the opportunity to turn the scenario around. This sort of focused army invasion is exactly what the AI should be doing to you. Now, obviously I don't know the full circumstances, but in E:wom a big problem was that the AI never seemed to amass that "killer army" themselves and invade you.

My point here is that this isn't just an AI weakness, it's an "issue" with game design. And just as the AI has to counter it, the AI should be using it to its own advantage.

A few theories I just thought up, though I don't know if these are realisitc additions.

Hmm.  Occupation forces are a good way of whittling down killer stacks, and SOD is the issue right. 

It's not so much Brutus administering the city, it's Brutus keeping the city from revolting. 

One idea: conquered cities have a high number of malcontents.  Champions stationed in the city reduce the malcontent rate and can convert them to workers based on CHA.  Occupation forces just drain the excess malcontents (who would be a 4th pop group, but otherwise function same as other malcontents, but you gain them by event or occupation)  Certain civ techs and spells could also convert the malcontents, or allow malcontents to become workers (a nerve stapling type of spell perhaps?)

If the 4th group gets too high- they might leave the city and become enemy militia outside the city if enough militia can be generated to hurt an occupation force.  Champions would help to prevent this, but not occupation forces.

Personally, I think the problem with AI stacks of doom, is that they can be countered by two half-stacks.  Having limits to size of force that can be generated I think is better for AI, but you could game it in GC2 as well. 

One other suggestion: allow defending armies to pin attacking armies for a turn if they can hold them off for x number of turns, with the battle continuing into the next turn.  This would allow decent-sized defender stacks to hold off stacks of doom for a while until reinforcements come in.  The ROTK game for the Saturn did this.

On the next turn, the attacker has the option of retreating or continuing the attack.  (any reinforcements are added on)  You will likely need to save positions of forces, which could be memory problematic though. (probably less then I'm thinking though)

I thought the answer to this question revolved around how you Brad can develop the AI to counter a human SOD, not how the mechanics do not replicate occupied city administration...?

(Btw history has been littered with able bodied people on both the attacking and defending sides who could administer a city after occupation)..