So, I was scanning the new posts today, and came across yet another recent discussion about city defenders, which made me think of one of my pet peeves in Elemental.  How the blazes did the attacking armies get in the city already?!?  Especially if you have added walls to the city?

So, following this train of thought, I spent a few minutes (re)envisioning some tactical vs strategic map changes that would make combat a little more interesting IMHO.

Maps drawn based on terrain and other features in the square you are attacking

   - Right now, Elemental chooses 'generic' maps for combat, with no direct consideration of what the underlying terrain looks like on the world map.  What could be more interesting is an algorithm that takes the strategic map square in question, analyzes the terrain in that square as well as the terrain in the surrounding squares, and generates the tactical map from that information.  So if there is a mountain range to the north, and a forest to the east, and the square you are entering has scrub terrain, well the tactical map would reflect that.  If you are in a mountain pass (mountain terrain to the east and west), well you will be fighting in a mountain pass with the mountains to the east and west.

  - Armies would be placed according to the direction they entered from.  Essentially, if you entered from the southwest, your armies would be placed in the southwest corner of the map.  Defenders could be placed in the middle of the map, or on the opposite corner.

  - Rivers would be a bigger deal - does the river have a bridge?  If not, how will your army get across that river to fight your opponents?  This gets a little tricky r.e. defender placement, which brings up my next point...

  - Why are rivers in the middle of a square?  This makes it a little more difficult to explain where the city defenders are.  Rivers by definition are natural barriers - hence placing them on the edge of a square/hex makes a little more sense, and more easily defines where the defenders are.  After all, the defenders aren't going to camp IN the river...  Bridge placement suddenly becomes VERY important, as do fords.  Also, if you are wading through a ford, what effect does that have tactically?  Maybe a penalty to dodge, for example?  Certainly there would be a movement penalty.

  - Some terrain features destructable - If there is a forest on the strategic map, the tactical map would have trees.  Said trees could be lit on fire... tornadoes might uproot said trees and scatter them across the map.  Units near these terrain features may take some residual damage (i.e. caught in the forest fire, etc.).  After a number of turns, well the trees burn down, and you just have smoke and ashes to deal with.  On that note, said smoke might cause an accuracy penalty for ranged weapons firing through it.  Or, if someone casts a 'flash flood' spell, suddenly the river doubles in size, washes out/unindates the bridge, and you have to wait a few turns for the river to subside before yo can cross a ford.  You get the idea...

Hmmm, this brings up fog of war... what if visibility was taken into account?  I.E. if none of your units have a line of sight to the defender, they 'disappear' from the map, underneath a 'fog'.  Could make things interesting if you are trying to sneak around another stack... perhaps if you are able to mange to exit on the other side of the map, you'd be placed on the other side of the square you just attacked?

  - City building placement and features should be reflected on the map.  Right now, a generic 'City Map' is chosen.  But if the above algorithm used the strategic view of the city in question to determine how the buildings were placed, that would automatically create varied maps.

  - City buildings and walls 'destructable' -  In order to enter a city, well ya gotta find a way in.  If defensible features are added to the city (walls, gates), well you'd need to traverse said walls or gates, or breach them to get in.  Also, maybe you'd like to mount a raid to destroy that great foundry, in a hit and run raid, rather than try to occupy the entire city... Hence walls and buildings would be attackable, and any resulting destruction would be persistent until fixed.

  - Block by block fighting in cities - Instead of one 'generic' fight to take a city, cities would be placed on a 'larger map' which shows the entire city.  And city defenders would show up in the city squares they entered, placed in the middle, placed near the attackers (observing their movements and reacting accordingly), or some other such.  Attackers would need to move through the city to eliminate all defenders if they intend to keep it.  As the city is fully represented on the map, this could get interesting.  Perhaps the defender has archers placed on rooftops?  LOTS of interesting options here.  This also ties in to the above, i.e. having to fight your way TO that great foundry on the other side of the city before you can set it on fire...

Here's an illustration of this idea, that I did back in the 'old days'...

If you think of the big squares as a number of smaller squares (say 8x8), then you can visualize displaying those structures on a tactical map.  Here the monsters are attacking northward, and the game could automatically assign initial defender deployments favoring that particular square.  Other defenders would be scattered nearby.

Also, note that dragon.  He'd probably occupy 7 or 8 tactical squares (7 diagonally, 8 horizontally/vertically)...  And Mr Stone Elemental there would of course occupy 4 squares.  Physical unit size hence becomes more important.  Groups/Squads of units could similarly occupy multiple squares, to reflect their size.  Say, 1 square for 3 or less, 2 squares for 4-6 individuals.  Mounts would occupy 2 squares, and units with multiple mounts would occupy 4 or more squares...  Or perhaps individual units could occupy even more squares, if the tactical square count was increased further (say 19x19 instead of 8x8)...

With actual walls represented that need to be breached, the city defender count of course may need to be adjusted downward accordingly.  But this does raise an interesting situation.  Say that the attacking player has multiple stacks.  In the first battle, the city defenders are automatically/pseudo-randomly placed near the first set of attackers.  THEN, when the next stack attacks, the city defenders are placed based on where they were in the last attack...

This also incidentally causes an interesting downside to 'snaking'.  Sprawled cities are inherently harder to defend than more compact arrangements.  If the defenders are in the northwest 'snake' of a city, defending against attacker stack A, well when attacker stack B in the southeastern 'arm' of the city initiates their subsequent attack, well those defenders will need to run all the way to the other end of the city, OR, the player/AI may 'leave/send' some defenders at/to the other end of the city instead of using them to repel the first stack.  Keep in mind that the entire city is represented on the larger tactical map, so the defender would 'see' that second stack lurking beyond the city walls, even though it is not active yet...

OR, once one attacking stack initiates the attack (clicks on the city to attack), the attacking player/AI has tactical access to ALL the of adjacent stacks he owns, for one super melee... In which case, it would make more sense to assign more dedicated city guards (a la what happens in Elemental now), which are initially assigned/placed by the game based on adjacent enemy stack placement: i.e. two enemy stacks = city defenders split in half into two groups near those two enemy stacks, with additional defending units (sovs, heroes, etc.) placed in between the two enemy stacks.

Catapults become more interesting in such a situation, as they can 'lob' attacks over city walls.  Archers could also shoot 'blind' over city walls, but with a huge accuracy penalty I would think.  In which case, they'd be more effective if they were stationed 'on the walls'.  Not to mention the fact that catapults are uniquely suited for city wall and building destruction.

One other note: With such an arrangement, and taking hit and run raids into account, retreat WOULD need to be an option.  Essentially if the attackers, or perhaps even defenders, move off of the the edge of the extended map, they leave the map, perhaps with no strategic movement points left.  Which is important because this gives the defenders the opportunity to 'chase' them on the defending player's game turn, and also because city fights SHOULD take longer than just some skirmish in the open.  Or perhaps city fights would soak up 3 additional movement points instead of just 1 additional MP...

With the above ideas in play, I would think you'd want units moving a little faster tactically (say twice as fast), to offset the larger maps that would be required for adequate representation of this concept.  In that case, units could be placed initially twice as far from each other, so that's easily accounted for.  Also, weapon ranges would become more important - assuming that the AI issues with weapon ranges are solved (we used to have ranges on bows, but currently we do not due to some AI issue).

Sure, city fights may end up taking longer with this arrangement, but cities are MUCH more important than open ground, so I think the time investment on the player's part is worthwhile.  Besides, if city counts are kept low, you won't be investing a lot of game time in city fights in the first place, as you need to accumulate an appropriately sized attacking force beforehand...

One final note, buildings that are destroyed should be able to be rebuilt more cheaply than if they were built from scratch.  The foundation is already there... Similarly, wall breach repairs would cost considerably less in time, money and materials than building the walls in the first place (obviously!).  But until they are repaired... well holes in walls tend to lure attackers.  Especially wandering monsters!

The city combat ideas expressed above are a little more radical, but I really do think that the next Elemental iteration needs to take it to the next level.  Other ideas have been thrown out as well (more interesting diplomacy, faction specific tech trees, etc.) which have been discussed elsewhere.  Feel free to share your own ideas here!

Thoughts?