Announcement

Collapse
No announcement yet.

AI and the game

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

  • #16
    I too agree that the AI should not cheat, with one exception... Intuition.

    Human players greatest advantage is that we can make educated guesses... we can guess where the greatest unit concentrations are, we can guess who's more likly to attack, we can guess which cities are the main military producers.

    The AI, can only use information that it knows... thus I think the AI should have a % chance each turn to make correct guesses about info that it dosen't posses... the harder the difficulty, the greater the chance. This will allow AI to behave smarter, as they will be better informed on occassion, something that humans (at least civers) already can do. Additionally, the AI should never guess wrong... just not be provided with extra info.

    The other important thing is to give the AI a memory, just because a unit moved out of line of sight dosen't mean that it should forget that it's there. This can be done by estimating force strengths within various regions of the map... this is a much more complicated matter, but is essitial to having a smart AI.
    [This message has been edited by Trachmir (edited February 08, 2001).]

    Comment


    • #17
      quote:

      Lets face it for a relatively complex game like Civ 3, we will never have an ultra- challenging human-like AI, so lets not base decisions on what features to include solely on this aspect of the game.


      This is wrong. It's entirely possible to have a competent AI. Today's desktop computers are certainly up to the job. The question is whether Firaxis will put the resources into the project.

      ------------------
      -----
      Chris Shaffer
      shaffer@uic.edu

      Comment


      • #18
        Very nice idea concerning the threat value and alliances dependent on the difficulty level. I like it very much and it is so simple. Why haven't I thought of that? In that case AI cheating would probably not be required, though I am not completely sure.
        Rome rules

        Comment


        • #19
          quote:

          Originally posted by Trachmir on 02-08-2001 07:44 PM
          The AI, can only use information that it knows... thus I think the AI should have a % chance each turn to make correct guesses about info that it dosen't posses... the harder the difficulty, the greater the chance. This will allow AI to behave smarter, as they will be better informed on occassion, something that humans (at least civers) already can do. Additionally, the AI should never guess wrong... just not be provided with extra info.

          The other important thing is to give the AI a memory, just because a unit moved out of line of sight dosen't mean that it should forget that it's there.
          [This message has been edited by Trachmir (edited February 08, 2001).]


          These two ideas are both very good, though I still think AI will also have to cheat to compete effectively.

          Rome rules

          Comment


          • #20
            The AI needs to be easily editable. The problem with the AI is that while at first it may appear chalenging, it wont learn. It will always be the same AI as on day one. On the other hand, the player learns from his/her mistakes, the player finds the optimal paths to victory and exploits them. Unless the AI can change (or be changed) it is NEVER going to be up to the challenge of beating a good human opponent.
            - Biddles

            "Now that our life-support systems are utilising the new Windows 2027 OS, we don't have to worry about anythi......."
            Mars Colonizer Mission

            Comment

            Working...
            X