Announcement

Collapse
No announcement yet.

Events for Freeciv 2.0

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    What AI behaviours would it be desirable to be able to script?

    (Please do not go overboard on this one. The AI will not be scripted, that would be way too slow, but some goals and behaviour may opened up for modification by events.)

    Right now I'm thinking:
    - being able to set the 'AI personality'
    - allow/freeze diplomacy for each AI player
    - add desire to achieve a certain tech
    - add desire to attack a certain city (for all units)
    - add desire to build a certain building (for all cities)

    Comment


    • #17
      I am under the impression that much of the AI's diplomacy with other players can be modified in civ2. How is this done?

      More importantly - how can we do it better? (This can theoretically be scripted, since speed is not really an issue.)

      Comment


      • #18
        Originally posted by Per
        I am under the impression that much of the AI's diplomacy with other players can be modified in civ2. How is this done?
        Civ2 AI diplomacy isn't anything special to begin with. In a normal civ game, at a certain size will always hate or dislike you. Although bribing them with techs and gold will restore their opinion of you for a while, but not for long. Allied civ's are even more bothersome in that respect because an ally, which was friendly at first, would break his alliance the moment you wanted to make contact again.


        Modification of AI behaviour isn't really the case here, at least not with Civ2, the ToT AI seems a bit more friendly but not much. For a scenario you can set it to war/be at peace/allied with other civs and set a "no negotation" trigger (certain civs cannot have diplomatic contact with each other) but that's about it. Some of the scenariowriters might be able to tell more.
        Skeptics should forego any thought of convincing the unconvinced that we hold the torch of truth illuminating the darkness. A more modest, realistic, and achievable goal is to encourage the idea that one may be mistaken. Doubt is humbling and constructive; it leads to rational thought in weighing alternatives and fully reexamining options, and it opens unlimited vistas.

        Elie A. Shneour Skeptical Inquirer

        Comment


        • #19
          Originally posted by Per
          I am under the impression that much of the AI's diplomacy with other players can be modified in civ2. How is this done?

          More importantly - how can we do it better? (This can theoretically be scripted, since speed is not really an issue.)
          Basically you have to assign a "no negotiation" event for each possible civ to civ diplomatic tie. The event prevents talking between the two civs. Unfortunately this takes up a lot of memory as you have to have an event for each diplo tie.

          Check the already posted Civ 2 event language for the No Negotiation event.

          As for modifying the rest of the AI, well, this is accomplished through none other than good old workarounds that exploit kinks in the Civ 2 engine. I could explain them all, but I'd reach the character limit for this post.

          If you want us to keep it simple I suggest implementing the basic events that Civ2 has, expand upon the core events by offering more options and bigger everything, and create a base for Scenario deisgners to create their own actions through Event workarounds.

          Comment


          • #20
            One of the many weaknesses of CIV2 events is in the CREATEUNIT command for AI civs. I don't know if the command creates the unit(s) during the AI civ's turn or at the beginning of a player's turn. In either case, AI units are created stillborn, with no movement points. This deprives scen designers of the ability to launch a real AI surprise attack.

            In his Rising Sun scen, Patine has tried to simulate the surprise Doolittle raid on Japan. To do so he has used events to created a stack of US naval units and bombers offshore of Tokyo. A Japanese player can locate and destroy the stack before the raid can be launched.

            Another example are the German units in Red Front which are spawned on the pontoon bridge west of Stalingrad. They are immobile and subject to Soviet attacks before they themselves can attack the city.

            1. Do SL scen designers consider it desireable to be able to create AI units with full movement, or have full movement as an option?

            2. If the answer is yes, can Freeciv developers do the necessary programming?
            Excerpts from the Manual of the Civilization Fanatic :

            Money can buy happiness, just raise the luxury rate to 50%.
            Money is not the root of all evil, it is the root of great empires.

            Comment


            • #21
              Originally posted by Per
              What AI behaviours would it be desirable to be able to script?

              (Please do not go overboard on this one. The AI will not be scripted, that would be way too slow, but some goals and behaviour may opened up for modification by events.)

              Right now I'm thinking:
              - being able to set the 'AI personality'
              - allow/freeze diplomacy for each AI player
              - add desire to achieve a certain tech
              - add desire to attack a certain city (for all units)
              - add desire to build a certain building (for all cities)
              Per,

              All those AI suggestions are in the right on.

              One thing I would add is the concept of a scenario deadline: the ability to "hurry" or "slowdown" before the final turn is reached. In a scenario you usually play until a certain number of turns. The player knows that he's playing against a deadline, but the AI doesn't. Making it "aware" of that deadline could be worthwhile. So if an AI is starting to lose toward the end it makes more sense to stall for time ("slowdown") and switch from playing agressively to defensively. Likewise the AI could be instructed to "hurry" if the victory condition is almost in reach.

              N.B
              The trouble is I'm used to thinking along the lines of Civ2 scenarios. A good Civ2 scenario will be scripted to follow one main path for the AI to make a good challenge against a human player. However, in Civ2 you can't fully script the AI to take care of any situation. This usually becomes apparent during the midgame. The AI at that point either falls apart or the player has to endure a long slugmatch against heavily fortified AI cities that he knows he will win eventually. If we could make the FC AI act more dynamically in the mid- and endgame this would be a major step forward.


              The main goal should be that for a FC scenario we must have a way to coach/point the AI into the right direction.
              Last edited by CapTVK; May 7, 2005, 17:40.
              Skeptics should forego any thought of convincing the unconvinced that we hold the torch of truth illuminating the darkness. A more modest, realistic, and achievable goal is to encourage the idea that one may be mistaken. Doubt is humbling and constructive; it leads to rational thought in weighing alternatives and fully reexamining options, and it opens unlimited vistas.

              Elie A. Shneour Skeptical Inquirer

              Comment

              Working...
              X