haha, you can tell if I'm coding or if I'm doing something creative (GUI) based on how often I respond. Coding = heads down, full concentration - doing anything else I look for a distraction about every hour.
Initial Release - PC, we have officially added support for Linux as well because we have tested Linux builds and expect to have little difficulty with the full port and I also conceptually like supporting free software. We've used open-source equivalents whenever possible to develop the game, with the notable exception of Unity (and Windows). Mac... is another issue - I've never used a Mac or ported to a Mac - if significant interest exists we will explore it after release but I wouldn't be too optimistic given the amount of "after release" things we want to get to.
Yeah, DLC of that sort would simply be able to be worked around - and I hate that approach as well. We want to be open and let people build the world they want, not restrict anyone.
"How does the AI work" is quite the question, I FULLY intend to write a huge technical writeup on Tigsource at some point - but it is mostly a decision tree with strong state mechanics. There are really three "interesting" (somewhat) mechanics that allow the AI to work the way it does and I'll go over them briefly here.
One is the idea of desires - leaders and heroes have drives that they want to pursue, but they will put them aside for other matters given the state of the world - but keep that desire bottled up long enough and they may make a more impulsive decision along that route as well as losing spirit (expressing their cracking will). It's a unique way to "suppress and frustrate" heroes and leaders, and to allow the AI to operate smartly but also still express its personality. Put a king who just wants to have a peaceful reign under constant pressure of war and he might crack, torment a hero who loves his friends with their death and injury and see him give up the sword.
The second is that the AI is "reactive" in that it maintains its state without recalculating the state of the world. We wanted to simulate that most days are just status quo for a leader - not every King is concerned with expanding his realm. They have a value that triggers how often they "recalculate" their position in the world and what they should be doing, and also some other actions can trigger this recalculation. How far back up they move in their decision tree depends also on their adaptability - an excellent leader will, on even a slight provocation, return to the base "what is the state of my kingdom" while a lesser leader may go from ->Increase Military->Raise Levies to ->Increase Military and then recalculate from that, possibly to ->Hire Mercenary Company. Only a much greater stimulus will force him to rethink his approach. This allows your agents to have some leeway in their actions, while also making it "interesting" who is ruling each nation.
The third is that Leader Available actions aren't "peer equivalent" to the player, meaning since the game is highly asymmetric we had some freedom to build a more interesting pool of actions for leaders. Instead, they have a list of possible actions based on their stats, their culture, and their nation. These actions are a good blend of mechanics and theme, allowing a King of A Feudal Kingdom to "Call up the Banners" and raise his barons to war, let a Slaver State equip its Slaves for Battle, or have a Lord whose Family is Cursed seek a Sage to End Their Troubles (which of course can be one of your agents for more fun). By blending the necessary traits between personal ,cultural, and government we ensure that you can't quite be sure how resourceful a given Leader will be, but you can have a general idea of what resources he can call on.
On the technical side we handle it like a "deck of cards" - each leader has his list of actions (each one being a class) - and he checks the "state of the world" vs "his desires" and finds a few likely options, he then proposes them internally (or externally given his state) - if you have infiltration you can vote on/influence the result, otherwise the event occurs based on who has the highest influence (advisors tend to be rational, the leaders tend to favor their personal traits).
No custom AIs as in changing the operation of logic - we did not go the Lua scripting route that is so popular - I'm going to say we "should have" as we were originally going to but I declined because I worked extensively with Lua on a military simulation project and found that while it was extensible it actually lowered our efficiency at producing quality results when it came specifically to AI. But then in the Editor we've replicated so much of LUA when it comes to exposing and manipulating variables that I should have gone that route - but too late for that now.
Customizing AI instead consists of modifying traits and adding events (as noted above). Want your AI to raise a horde of Chaos Warriors from the Abyss? Add an event for a specific culture/trait, set it to a cooldown or conditional, have it create an army of X type units - and the AI will grab that whenever possible. Make it 100% desirable with a 4 turn cooldown and then every four turns the leader will summon his chaos legions. More... subtle... events are also supported.
Author
Topic: That which sleeps- Kickstarted! (Read 370655 times)