Announcement

Collapse
No announcement yet.

Fixing the AI's 3 movement folly

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

  • Fixing the AI's 3 movement folly

    3 movement units have always been a problem for the AI when they launch their attacks. You might have an AI attack for example, where 18 tanks move next to your city to end their turn, and 3 cavs (with 3 movement) attack. Naturally, your next move would be to blow the tanks away with everything you have, even though under human control, those 18 tanks would be quite devastating. The short of it is that the AI has a problem of showing their hand (target) with poorly planned unit movements that allow players time to counter.

    After observing enough games, I've come to the conclusion that a potential (easy) fix might be to simply tell the AI to prioritize 3 movement units with a preference towards targeting cities /targets that they can reach in the same turn.

    Attacking important targets on the same turn, can really blow our socks off by. (ie: I declaration of war followed by an all out assault on cities within reach by the fastest unit, followed by the 2nd fast and so on). This means the mobile 3 move units go in first. And the 2 movement unit can follow. Defensive units then move last, finding the path that allows them to stop on top of the 3 movement or 2 movement stacks. The idea here is for the fastest units to go first, the slower units will then try to find alternate paths that will allow them to end the turn stacked with the faster units (if possible). This is of course, a common human strategy.

    Alternatively, if targets are too far for a 3 movement unit to attack in their first turn, the stack can simply move like a turtle, with the defensive unit (slowest) determining the pace of the rest of the stack. I believe this is done by the AI currently, but i'm putting it out there just to cover all bases.

    An implicit advantage here is that this also partially solved the combined arms issue as movement based attack priorities would imply that the AI would go through a set order and send units of a certain type. Later in the game for example, cavalry charges will precede a tank attack. I don't know about other players but during the transition from tank to cavalry, I like to send my expendable cavalry to soften up a target first then move in with my tanks. pre-tanks, Cavalry go in first, which make some sense, and assuming some sort of tech parity, cavalry being the fastest unit in the late middle ages tend to be used first anyways, because of their Attack power and range.


    I think this prioritizing is possible to do with the current engine. I doubt we'll see it, but any comments are welcome.
    Last edited by dexters; February 24, 2004, 09:24.
    AI:C3C Debug Game Report (Part1) :C3C Debug Game Report (Part2)
    Strategy:The Machiavellian Doctrine
    Visit my WebsiteMonkey Dew

  • #2
    Nice can of worms!!

    Of course the issue is ACCESS to AI routines et al.

    Let us pray to the errant Demi-god Firaxis for these sublime privilages and with this add 'events'.

    Many Thanks and er adorations...
    The Graveyard Keeper
    Of Creation Forum
    If I can't answer you don't worry
    I'll send you elsewhere

    Comment


    • #3
      This is not as big of a problem as the AI not using defensive works or artillery en mass. If they built half as many artillery's as they do ironclads then the AI would be much more of a challange. Artillery is devestating on defense and formidable in attack it is by far the worst problem with the AI.
      * A true libertarian is an anarchist in denial.
      * If brute force isn't working you are not using enough.
      * The difference between Genius and stupidity is that Genius has a limit.
      * There are Lies, Damned Lies, and The Republican Party.

      Comment


      • #4
        I think the arty issue is far more difficult to solve.

        having the AI build lots of them isn't a problem. I think the problem is making the AI use it. Which for some reason was a promised and hinted fix in C3C that never really materialized.


        I still think making the AI use its 3-4 movement units better by prioritizing them and focusing on targets they can hit on the same turn will make it hard for human counters to work, at least in the initial stages of a war. As I noted in my original post. most AI invasions consist of poorly planned movements that see big expensive (potentially lethal stacks) ending their first turn going for a city or target that would normally not be on a human hit list in the first turn of war.

        Most of us want to declare war and use our stacks to hit cities in the same turn. Rarely do we send expedition forces flat hatting through enemy territory unless the force is expendable or we have overwhelming superiority. The AI's problem is, they almost always seem to send these type of stacks into our territory beggin to be bombed and mamed.

        eDIT: I make note of this in hopes that there is still time for Firaxis/Breakway to at least consider this and try out my ideas before the final patch is released. I'm not expecting a user mod from this suggestion, which I concede is probably far beyond our powers.
        AI:C3C Debug Game Report (Part1) :C3C Debug Game Report (Part2)
        Strategy:The Machiavellian Doctrine
        Visit my WebsiteMonkey Dew

        Comment


        • #5
          Addendum: Another potential benefit.

          Fast movement units hit cities within range. Some may fall/razed pushing back borders. Once this is done, the slower units can now use the terrain/roads/rails as without border penalties.

          The order of movement is important. Moving a stack of 18 tanks as opposed to attacking with cavalry first could mean the difference between leaving those 18 tanks out in the open and vulnerable at the end of the turn vs. having the option of hiding some of the tanks in the newly conquered city or hitting another juicy target in the now reduced borders.

          This also allows slower 1 or 2 movement defensive units to catch up and end their turn stacked on top of the faster moving units since they don't suffer the border penalties.

          IMO, we all do this. order of movement is important and prioritizing fast movers to go first, has this added benefit. Of course, they may not take any cities, but it has that potential to push back borders. Ideally of course, we want the AI to move their units in such a combination that allows for maximum effect (ie: a mix of fast and slow movers on a target). But being realistic about it, I think having it attack by simple movement priority can allow the AI to capture some benefits of what human players do.
          AI:C3C Debug Game Report (Part1) :C3C Debug Game Report (Part2)
          Strategy:The Machiavellian Doctrine
          Visit my WebsiteMonkey Dew

          Comment


          • #6
            Heck, I'd just be happy to see the AI use high ground better.

            Good ideas though.
            The greatest delight for man is to inflict defeat on his enemies, to drive them before him, to see those dear to them with their faces bathed in tears, to bestride their horses, to crush in his arms their daughters and wives.

            Duas uncias in puncta mortalis est.

            Comment


            • #7
              Originally posted by Theseus
              Heck, I'd just be happy to see the AI use high ground better.

              Good ideas though.
              Hmm...I have always thought that the AI uses the high ground pretty well, maybe thats why I am stuck playing Regent.
              * A true libertarian is an anarchist in denial.
              * If brute force isn't working you are not using enough.
              * The difference between Genius and stupidity is that Genius has a limit.
              * There are Lies, Damned Lies, and The Republican Party.

              Comment

              Working...
              X