How many fortresses to build?

[SafeKeeper]

I also think it's perhaps a good idea to build a fortress if you conquer a particularly wealthy province, one that allows you to recruit particularly useful units, one with an important magic site, etc. If a province is especially important to you, it might be an idea to fortify. Anyone else agree?

[Teraswaerto]

For many nations building forts in provinces with the most resources is more important than chokepoints or useful indies.

When you can't spend all your income each turn on recruiting mages and national troops, you don't have enough forts. Might not be true for all nations all the time, but it's the general guideline I try to go by.

[Actuarian]

I agree with the previous posters in that forts are usually a more important economic decision than a military one. It can also be a good thing to build a fort in a place with good indie troops, like longbowmen.

If you do build castles defensively and you think an enemy might invade in a turn, make sure you choose if you want your troops to defend the province or hide in the castle. The default is to hide in the castle. I can't remember the defend province command (Is it patrol?), someone help me. Personally, I have a hard time remembering to choose and I invariably find province defenses destroyed and my fort beseiged.

[TwoBits]

Patrol is the only command available to defend a province with a fortress, unfortunately.

Personally, it sounds like some people are building too many fortresses. The original poster had 7 for 18 provinces. Wow! That's a lot of gold expenditure that could have otherwise gone into, well, a lot of things - like researchers, commanders, or more conventional forces to take more land and go on the offensive.

Remember, a fortress only stops an enemy who wants to besiege you. They're not forced to stop there - they can bypass the place and keep on attacking elsewhere. The AI probably won't do that very often (ever?), but an SP player would certainly consider doing it...

[Taqwus]

Very situational.

An army inside a fortress -- especially a flying army -- threatens lines of communication of anybody who attempts to bypass it. Whether or not this is a problem depends on whether the invaders have enough supply items, have another way to bring in reinforcements as needed, and whether there's a risk of being wiped out wholesale due to fleeing while surrounded (ex. by a surprise Wailing Winds and other heavy fear attacks).

Fortresses used to be much more of a no-brainer, 'tho; refer to the D2 "mad castling" approach of taking cheap fortresses esp. Watchtowers and spamming them, partly because of the availability of Ghost Riders (which is now D6 instead of, er, D4 if memory serves). The changes make it both far more difficult and less productive to do this.

[Etaoin Shrdlu]

Might it make sense to throw a few site-search spells at the proposed fortress province? Just in case there's one hiding next door already? It would be mighty annoying to discover you've built a castle and find that the Crystal Citadel or some other high-defense fort was just waiting for someone to find it...
Failing that, I'd put forts:
1. On high-resource provinces (especially if I'm using heavy forces still)
This makes it easy to reinforce chokepoints, even if the castle is not on the chokepoint per se...
2. Places with hard-to-build units (why oh why are Knights only available in resource-poor lands?)
3. Where I was going to need a lab anyway (but for interesting mage types only. If there's three Libraries within range I'd likely put one in the middle and pull the Sages into the castle from next door...)

[Xietor]

In Dom II i built more castles than dom III. Now that the Ai builds castles, i typically just take theirs and use my gold elsewhere: Pans are expensive, as are most of the Pangaea's troops. So I would rather just take castles away from my enemies.

I will build a castle at a clear choke point, as others have pointed out, and on squares that make rare troops that i really like.