I am building a fairly large scenario which will involve almost all aspects of CMO warfare, ship/sub/air/land. The theatre is approx 1500nm across and it involves many-day naval travel, the player will be assaulting a country with a very strong carrier group.
What I would like to ask about is the degree of complexity to give to the programmed enemy.
Do you "over equip" the enemy? (for example, with few enemy submarines, its quite possible in a large scenarios to not encounter them at all. if you add more enemy subs, you increase the chance of an encounter. Or, adding large numbers of aircraft to support the missions needed for an engaging scenario - or a realistic number of aircraft with LUA scripting to swap them between missions?)
Generally speaking, how many missions are required to simulate a decently aggressive enemy, ballpark?
How much to rely on unit AI and how much to "micromanage" mission planning? (eg: many strike missions with independent characteristics, or fewer missions with large numbers of assets assigned.
Many strike missions naturally require a/the target to be located/spotted first - how much effort to put into enemy recon? (otherwise many missions could go untriggered making a pretty boring scenario.)
At the moment I am planning quite extensive recon, involving airborne maritime patrols, many drones and a handful of smaller naval groups on patrol. May potentially add in civilian cooperation with the enemy so that they can spot for them. Given the pretty long-range nature of the players CAP capabilities and the generally quite shortish-range of available surface-search radars(max 225nm), it may become quite difficult for the AI to get a bead on the player group. Or at least to maintain one - it appears long-range strike really need eyes on the target at all times, at least until missile launch. Non-Pinpointed targets (area contacts) make poor targets for AI strike missions, making ELINT and sonar less useful in this role.
Long range CAPs and patrols are fairly easy to manage with in-flight refuelling, but I am having trouble efficiently extending the range of AI strike missions with AAR - especially when the targets (the player group) will not arrive in a fixed location, any pointers?
How much LUA code does a good scenario use?
I have been able to figure out LUA to add units to the scen, which is quite useful with the event system. Is it going to be essential to learn more complex LUA functions to make an interesting AI?
I apologise for the vague nature of the questions, at the moment I believe I have placed sufficient ground/naval units to constitutes a decent enemy, and am just starting to add the missions and aircraft to support the various AI defensive strikes, CAPs and recon.
Any input welcome, I think a few pointers could make the 1st iteration much better and require much less reworking, many thanks in advance
P

