Idea for improving AI's use of forts

[glider1]

People over at Civ5 strategy have mentioned that the AI will occasionally build forts and man them, but usually at a poor strategic location. I realize that it is a big computational load for the AI to calculate best placement for forts, but what about if fort placement was pre-computed during the initial map generation based on terrain features, and the fort location points stored in a database, so that when the AI get's the urge to build a fort during the game, it simply looks up the database and checks whether the optimal fort location is in it's territory, and then places it there. This would be a fair bit of work to code and it would mean modifying Lua map algorithms to add fort location logic, but could it be worth it?

First question is whether the AI's fort placement even can be intercepted in Lua during the game. Second question is what is a good fort location that can be pre-computed? I think that it must be possible to compute terrain based choke points that fill in a gap in a line of mountains for example. We don't know where cities will be located during pre-compute, but there are positions on the map that are good choke points merely by terrain feature I think. Things like peninsulas etc.

Cheers

 

[tlaurila]

Maybe.

On the other hand, the AI fortifying cities with a manned fort next to each city towards the general direction the enemy civ is (if adjacent) could already be a relatively useful obstacle to taking cities. AI usually does have the worked time to spare, even to build/tear down those forts upon war/peace.

 

[Sneaks]

I hardly expect the AI to ever want to build forts as I never do it myself.

 

[Ahriman]

I hardly expect the AI to ever want to build forts as I never do it myself.

Agreed. I do build some, but very rarely. Most of the time there isn't a fixed defensive position within my territory where I'm willing to give up tile yields.

 

[Jorlem]

I think I would use forts more often, if the build time wasn't so long. IIRC, it was around five turns the last time I built one, and by the time its done, the war isn't anywhere nearby usually.


This probably isn't under the purview of VEM, or even possible without access to the core game, but it would be neat if forts could be used as part of a territory capture mechanism. Something along the lines of, if a fort is garrisoned by a player for a certain number of turns, it generates a culture bomb effect for that player. This way, you could end up with paired forts on both sides of the border, similar to how many castles were paired by an opposing castle in real life, and you could have wars over territory, and resources that are on the border.

 

[tlaurila]

I was thinking the same, if forts would take the same time to build as roads, say, they might be actually useful. Also the loss of yield bonuses should be harsh enough, no need to make them cost gold per turn in addition.

And even though humans don't really need forts, doesn't mean the AI couldn't slow down human advances by fortifying near cities. Especially before artillery comes around.

And perhaps forts could give cover bonus against bombardments, in addition to the defenses they do now. If forts aren't used now (I certainly don't), then maybe they should be beefed up some to make them a useful option, enriching the game? Downside is that such would increase the AI/human discrepancy.

 

[Sneaks]

General question to all:

Would you use forts if they upgraded into Citadels after 20 turns?

 

[Dunkah]

I would love to see a mechanic that makes AI's build forts and play a defensive game. SOunds like a good idea, but without knowing where the cities are being placed it seems a bit fruitless.

I my current One city challenge, 250 turns in I have a row of forts beyond the buildable tiles that do not contain any resources. My 4 unit army is holding strong in that line. Just waiting for the first backstab.

I also build forts late game if I am lagging in science to a nearby rival civ. They make all the differnece when my Riflemen are being attacked by Infantry.

@Sneaks: Citadels may be a bit Op. Perhaps if there were other conditions to the upgrade, many turn garrison/ Gold/ possibly the presence of a Great General (doesn't expend)... Perhaps the option to upgrade only one...

Instead I would like to see another version of a fort. Quicker to build and not quite as strong... perhaps call it "Front Line" "Out Post" "Defensive Position" "Trench Line" "Fox Hole"... perhaps that could be upgraded to a fort? Should be able to build these in enemy territory once DOW is declared.

And would like to see any unit able to build them as well.

I would love to see units stationed in forts have the ability to gain experience from training. And give them some other significance to peaceful times. Like influence boost to rival civs.

 

[tlaurila]

Instead I would like to see another version of a fort. Quicker to build and not quite as strong... perhaps call it "Front Line" "Out Post" "Defensive Position" "Trench Line" "Fox Hole"... perhaps that could be upgraded to a fort? Should be able to build these in enemy territory once DOW is declared.

It wouldn't be bad, but likely unnecessary complexity. And no, I don't it should be buildable in enemy territory. You have "fortify" for that.

 

[Ahriman]

It wouldn't be bad, but likely unnecessary complexity. And no, I don't it should be buildable in enemy territory. You have "fortify" for that.

Right, this.

And no to upgrading to Citadels; if you want a Citadel, use a GG.

I think forts are fine; they are situational but are occasionally useful. The AI is never going to use them very well. So I wouldn't want them to be any more useful.

The fortify bonus already accounts for regular digging in/entrenchement, and the AI can use that fine.

 

[Thalassicus]

I think forts should simply be a mini-city.

A fixed border radius of 1, plus a garrison slot, but no other city capabilities. Great General citadels could be forts with more hitpoints, a 2-tile border radius, and bombard capability. The AI would understand this well since it already knows how to attack and defend cities. It'd just require some intelligent fort placement algorithms.

 

[Ahriman]

I would be very leery of anything that messes up AI tactical movement by adding extra ZoC or reducing movement; AI tactical movement is already very weak.

Slowing enemies is super powerful because it makes ranged attacks incredibly strong.

 

[Thalassicus]

Yeah, I changed my mind right away when I thought of a better idea.

 

[Sneaks]

Cities founded on forts get free walls and barracks? Again, ai wouldnt get it

 

[Thalassicus]

Sneaks, who are you replying to? Your last sentence was a little confusing.

 

[Sneaks]

Myself!

 

[Dunkah]

It wouldn't be bad, but likely unnecessary complexity. And no, I don't it should be buildable in enemy territory. You have "fortify" for that.

To me Fortify represents a unit taking a defensive posture by making the best use of the cover that is available in that given terrain. Fortify in a farm tile would mean that they are ducking into farm buildings, hiding behind fence lines, behind haystacks, farm equipment etc.

Forts represent a permanent building that would act as miliatary center, waypoint, strong point in your own territory for some future calamity that may take place.

A lesser version of Fort would be a way for units to change the terrain they are in by digging a trench, foxhole, build an outpost, strong point within an enemies terrain even under fire, that would be a way for you to take a defensive posture in an enemies terrain. They they could Fortify themselves within that terrain improvement to get a further bonus at the expense of attacking.

Imagine if there was a choke point in the enemies terrain where you wanted to hold out for a couple of turns. Build a trench line and hold until your units can attack from another direction finish off another civ or simply get from one side of the country to the other..

Even if you needed a worker to build the improvment, (or better yet an engineer unit), it would be cool.

 

[Sneaks]

Ideally, once I can figure out the AI logistics, I am thinking of making a Forward Camp sort of fort that is buildable by Vanguard units, and will essentially be a fort that will destroy itself after 5-10 turns.

 

[Dunkah]

Why would you have it destroy itself?

Could you make that happen after it had been abandon? 5 - 10 turns of non-occupancy?

 

[Ahriman]

To me Fortify represents a unit taking a defensive posture by making the best use of the cover that is available in that given terrain. Fortify in a farm tile would mean that they are ducking into farm buildings, hiding behind fence lines, behind haystacks, farm equipment etc.

To me, fortification means much more than this; it means constructing foxholes or ditches or trenches or cover, or building the type of encampment used by legions at the end of each day. It means putting up stakes in the ground to deter cavalry.
If means positioning the AT guns in appropriate positions and setting up MG positions with sandbags.

Basically, fortification is everything that isn't constructing a serious long-term defensive structure like a castle or star-fort or major defensive works (concrete bunkers, tank traps, mines, Maginot line/Atlantic wall type stuff, etc.)

There is already a sizeable bonus for fortification. I don't really understand why we would want anything larger than that. The defensive bonus belongs to the unit, and is constructed by the unit, it doesn't need civilian workers to come and build on the tile (replacing the existing improvement).

Basically; we have 4 levels of defense; nothing, fortified, fort, citadel. That seems pretty broad to me.

What is the design goal here? Complexity? Do we really think that defensive screen + ranged attacks (which the human can do well but the AI can't) isn't powerful enough? Do we really need to be about to be super-good at defending while inside enemy territory?

that would be a way for you to take a defensive posture in an enemies terrain

We can already do this, it is called fortification.

They they could Fortify themselves within that terrain improvement to get a further bonus at the expense of attacking.

I don't see why we need a terrain improvement. And we already have defensive units that get no bonus attacking; the various levy unit promotions, and the defensive bonus on spear/pike/AT gun.

Imagine if there was a choke point in the enemies terrain where you wanted to hold out for a couple of turns

Then fortify your unit. That is exactly what existing fortify command is for.

Even if you needed a worker to build the improvment, (or better yet an engineer unit), it would be cool.

It would be MM that the AI would not understand. I don't think its worth it.