Paradrop ability as a promotion ???

[The Great Apple]

The plan was to have the AI attempt to congregate it's paratroopers in cities, and then check to see if there are any squares worth jumping to. If they are then the paradrop occurs, otherwise it doesn't.

 

[dh_epic]

"Worth jumping to" is tough to determine. But with enough work, you can get it. Something important to consider is that if the AI is extremely predictable with choosing the "best tile", then the player only has to keep an artillery or two to take them out.

Much harder is teaching the AI to protect itself from paratroopers. With the ability for the player to 'teleport' several tiles, the AI will have a hard time predicting when and where an attack will occur. Not only that, but the AI has a difficult time dealing with "chokes" as is. A paratrooper on a forested-hill would be a nightmare for the AI, especially if it leads to some pillaging.

Starting with multiplayer couldn't hurt. If you make it so MP doesn't degenerate into paratrooper wars, then maybe you can teach the AI how to deal.

 

[The Great Apple]

I don't think the AI will have problems defending against it - they defend quite deeply anyway.

For "best tile" analysis there are alot of factors you can take into account. Proximity to enemy/friendly units/cities, tile defences, resources and improvements, how well the area is known (unmapped plots/FoW/visible plots). The more things that can have an effect the less predictable the AI will be with them. If it can be scripted so as to send scout aircraft in advance of a paradrop, then this would be even better. They would be able to land on those chokes just as well as you can

 

[dh_epic]

Sounds like you have your head on your shoulders. Be extremely careful. Getting the AI to master the basics is only the half of it. Once players understand the algorithms at work, they'll try to milk it and exploit it for all it's worth. Then you have a whole new set of coding to do.

Best of luck.

 

[scottcstoness]

Starting with multiplayer couldn't hurt. If you make it so MP doesn't degenerate into paratrooper wars, then maybe you can teach the AI how to deal.

This is a serious concern; spamming paratroopers could be a major problem. Ideally they would be less powerful than just about anything on the map; their job is to jump in and take out a resource, jump in and take out a supply line, jump in and interdict railroads, etc. Certainly WWII paratroopers were next to useless against a prepared enemy (witness the drops on Arnhem, Crete). When used mainly for interdiction (Normandy, Nimejen/Eindhoven), they were pretty successful.

I wouldn't worry about a paratroop promotion that would be given to a combat 3 or 4 veteren unit; pretty hard to spam those. I'd suggest that the paratroop promotion also come with some additional benefits; perhaps +10% base strength like a combat promotion, or something: after all, most paratroop units have traditionally been used as elite infantry. Units that start with paratrooper should be weaker, like Str 16 + free paratroop for base str of 17.6; weaker than infantry, but still useful. Might be nice to have paramobile artillery, too, perhaps like the poison gas (str 6) in the WWI mod; essentially only good for bombarding or as throw-away units.

Might be nice to have an air transport unit that needed to be in the same square as the paratrooper with a movement point; paradropping would use up the movement point of the paratrooper and the air transport; I think that might be possible in python. That would solve the basing problem (can only paradrop from somewhere with an air base), you could allow interception of the air transport unit, which would allow fighter interdiction of the paradrop, and you'd need to commit a few more resources to the drop, which would help prevent spamming.