I like the idea of forts outside cities, but I haven't found a game yet that implements this well. Outside forts and castles were meant as early warning posts and also to pin down armies, as attackers would have to surrender and siege the castle before been able to move on. So, the attacker would have three choices:

- Storm the castle. Usually quite improductive (lots of casualties and very low chance of success).

- Siege it until the castle surrenders. Improductive too as armies could only war some months of the year (peasants had to plant and care for their fields, and no one makes war in winter), so you couldn't waste so much time on those types of fortifications.

- Split your army and leave enough army behind to avoid any defenders coming out and raiding your supply lines and getting a fight in two fronts.

So far I don't know a game that has made that process well (as the attacker can just ignore the castle and move on). So unless there's a really strange geographic configuration, they are useless in games. And if it's implemented in a way that stops attackers, I'm not sure it will be fun for them (or encourage defensive play too much).

While nice for Kohan, I'd rather forts not be magical army producers, and they'd need to produce a good sized army to do anything against another army.

In RTS games, combat doesn't happen in a turn, your fort either sends guys out and they're utterly destroyed with little effect, or they win.  In either case, the action takes time that allows you to send an army to deal with the threat in the mean time.  It's a delay, one that shrinks with the advantage the enemy has, but still a delay.  In TBS, a turn is a turn is a turn.  Delaying an army will at best, use up a point of movement.  If you deplete movement with an attack, then armies can be perpetually stalled by suicide runs of single squads.

I'd much rather have a fort function as an actual fort did in real life.  It's a place to watch the borders from.  Walled, relatively well defended, and of zero value to the enemy unless they plan on holding the area.  If you need a serious force to assault a fort, they will make excellent watch towers from which to scry your border and prevent incursions.  You can then keep actual armies further back in a more convenient location for guarding a larger area from major threats.  The forts themselves can ignore any minor incursions, as they will be easily defended against with minimal troops.  Without forts, your border watch would be constantly wiped out by small parties, easily blinding you to attacks.

I do like the intent of the topic cretor, which is (if I'm reading correctly) to make the player choose between multiple city types to prevent the player from spamming an identical "optimal city" build over and over.

It would kind of be like in Medieval 2 Total War. Every settlement could either be a "castle" which produced the best troops and was difficult to seige, or a "city" which didn't train as good units but was an economic center. You needed to find the right balance with mostly citys, but with enough castles spread around to be able to defend your borders.

Now, the TC's idea isn't nearly as cut and dry as that, but the concept of having cities that have strengths and weakneses is still there. However, I not sure I agree that it should be the central "castle" of a city that determines that.

I think galciv2 worked fairly well with it's capitals. I would usually have one planet full of banks and an economic capital, a research planet with only research centers and the research capital, etc. It led to my core planets being highly specialized and valuable, and my border planets to be a bit more generic. If they could expand on that somehow, it would work to add diversity to cities.

Perhaps you could even have a generic bonus as well. eg. One barracks in a city produces troops at 100% speed, two produce troops at 205% speed, three at 310% speed. I would very much like, for the sake of feel, visual signals for this as well. Many forges in a town would lead to a cloud of smog above the city, lots of magic labs would cause the city to glow faintly, and runes to float about, zooming in on a military city would treat you to marching drums and seeing tiny soldiers drilling.

The fort doesn't have to create units every turn.  If being attacked depletes movement, you can tie an army down with normally produced suicide units.  The game then becomes a perpetual stalemate whenever someone uses the tactic.  There isn't any significant delay a fort can do to a passing army without depleting it's movement.

I'd love to have irregular troops that can raid, but I think you guys are seriously over-estimating the damage potential you're looking at when a small force attacks a large one, especially if it's cavalry.  Setting up some sort of ambush system wouldn't be much effort since you can have hidden and stationary requirements that are triggered by the enemy moving past.  Balancing any sort of raid system that doesn't use the normal combat mechanics would be hell though.  It will almost have to be pointless.

The point of the raiders is not to attack the army, but the supplies and supply lines. You need very few men to do that in a very effective way and you need a tooon more of soldiers to protect against it (as if the raiders are mounted they can nearly hit wherever they want). Of course that doesn't work most times in games because armies don't have or need supply lines, so the only choice for raiders is to suicide against the blob of units, which is pretty pointless.