Announcement

Collapse
No announcement yet.

Somthing's wrong: Repeatable and predictable

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

  • Somthing's wrong: Repeatable and predictable

    I played Civ3 at Warlord level a few times and got to the point where I felt man enough to try Prince level (not that I had won any Warlord level games). At Warlord level, I had the frustration of loosing 20 or 30 Modern Armor tanks to a single Mech Infantry unit when trying to take over a city , but that frustration didn't compare to what I ran into at Prince level.

    Early in my first Prince-level game I sent a veteran archer and a veteran warrior to take over a city with no walls, a population of 1, and a single regular warriar guarding it (I was German, enemy was English). I couldn't take over that city to save my life (after several re-loads and retrys). But then one of their regular archers took over one of my cities with a population of 3 that had a veteran archer gaurding it. The regular archer beat me whether my veteran archer attacked him or he attacked me. That prompted me to do some testing:

    Problems Defending:

    Situation: An English regular archer (2 attack, 1 defend, with 3 hit points) attacked from grasslands into my city (population of 6, but no walls) guarded by a veteran German spearman (1 attack, 2 defend, 4 hit points)

    Results: He took over my city SIX TIMES IN A ROW WITHOUT SUFFERING A SINGLE HIT!

    Problems Attacking:

    Situation: I created an ARMY of 3 veteran spearmen (combined for 3 attack, 6 defend) and attacked a regular warrior (1 attack, 1 defend, 3 hit points), I attacked from a forest into forest.

    Results: SIX TIMES IN A ROW he took two damage, but always managed to wipe out my entire army!

    I'm sure I could have reloaded and retried this over and over with the exact same results because it always did the exact same thing. So unless the English have some built in advantage over Germans, I would say that something is seriously flawed here.

    I agree with other people who say there's no way a swordsman can destroy a tank (unless you want to take into account that the swordsmen sneak up on the tank crews while they are sleeping and slit their throats -which I think actually could happen, but the problem is how often is that likely to occurr?) More problematic is a wooden ship taking down a battleship. I don't think it would so much as put a dent in it even if every single shot was right on target.

    I've had a lot of fun getting whooped at Warlord level, but 5 minutes into my first Prince level game I am calling Civ3 unplayable until a patch is released, or could it be that I just need to reboot my computer? Or is it the fact that I'm using Windows XP?

    I guess I'll resume my life without Civ3 until an answer (or better yet a patch) comes.

  • #2

    I believe your "repeatable" results are due to a saved "random seed." Reloading a game will not change the results of the next "random" action. Thus, if you lose a battleship to a caravel, and reload and try again, you will lose again. and again. and again. Hence, the "inbalence" you saw was truly an isolated event, not a repeatable phenomenae.

    IF you really think the AI gets a combat bonus, you should record ever battle that occurs for a few turns, then to a statistical test of the results versus true random results.
    -----------------------------
    -JsA

    Comment


    • #3
      I think what you fail to realize is that the outcome of the combats in question had already been generated by the game before you ever saw it. The little animations you are seeing are merely the game showing you, the home viewer, the outcome based on an imbedded random number the game already knows.

      In other words, you could reload the game and retry the combat a million times, or until Armageddon, and the outcome wouldn't change, because the outcome is no longer random - it's predetermined.

      The only thing you can do if you save and reload a game is try something DIFFERENT, to "waste" that number - and return to the original combat in a different turn.

      Comment


      • #4
        This is absolutely correct. Reloading and trying the same battles over and over again should yield the exact same result.



        I predict when the patch is released a lot of speculation about what bonuses the AI receives will be put to rest...




        Dan
        Dan Magaha
        Firaxis Games, Inc.
        --------------------------

        Comment


        • #5
          I appreciate the comments...definitely good advise, but I can still assure you that it looked like every battle was biased SERIOUSLY against me, which really started to tick me off. I'm mean I'm there flipping off my monitor and shouting things like "yeah, right...sure I believe that could happen", cussing and then quickly looking around and feeling happy that I work for a bank and got Veteran's day off while my kid was at school and my wife was at work.

          Comment


          • #6
            Originally posted by Dan Magaha FIRAXIS

            I predict when the patch is released a lot of speculation about what bonuses the AI receives will be put to rest...




            Dan
            Perhaps but WHY wasn't this resolved BEFORE the game was released?

            Didn't you do any playtesting at all?
            Any views I may express here are personal and certainly do not in any way reflect the views of my employer. Tis the rising of the moon..

            Look, I just don't anymore, okay?

            Comment


            • #7
              No, Infogrames made them release it before it was ready, to avoid the deluge of big name games in the first quarter of 2002 and hit the christmas market.

              Comment


              • #8
                Originally posted by Dan Magaha FIRAXIS
                I predict when the patch is released a lot of speculation about what bonuses the AI receives will be put to rest...
                Dan
                I would guess from this that the patch will add the combat odds to the screen? Good idea, if true.

                Comment


                • #9
                  If you'd really like to test whether the game is doing some sort of funny bias against you in combat, you could try this: save your game right before combat. Record the combat environment:

                  - the unit types
                  - current unit damage levels (hit points)
                  - whether the defender has fortified itself
                  - whether the defender is across a river
                  - defender's terrain
                  - if the defender is in a city, record city size, and whether the city is walled

                  Then fight, and record the results, including the number of hits scored by each side. When you're done, reload, and this time, engage in diplomacy with someone, ask to make a deal, then cancel. Then perform the combat exactly as before. Then reload, make a deal and cancel twice. Then three times, then four, and so on. Each time, the AI should be saying something different at random. I have a hunch that this uses the same seed as the one used for combat, and so that should result in "random" combat outcomes.
                  gamma, aka BuddyPharaoh

                  Comment


                  • #10
                    I'll show you my patch if you show me yours...

                    Originally posted by Dan Magaha FIRAXIS

                    I predict when the patch is released a lot of speculation about what bonuses the AI receives will be put to rest...
                    Dammit Dan don't just leave it like that, we need INFORMATION. Don't be a patch tease!

                    Venger

                    Comment


                    • #11
                      Originally posted by Alexander's Horse


                      Perhaps but WHY wasn't this resolved BEFORE the game was released?

                      Didn't you do any playtesting at all?
                      I doubt he means that AI "bonuses" will be removed, but rather that something will be added making it clear they never existed in the first place.

                      What I want explained is how things like the inabilty to move unit stacks, the fact that you are forced to watch AI moves, slowness on the game for some systems, and, most glaringly, malfunction of air superiority passed playtesting. Complaining about AI advantage is chasing ghosts when real daemons exist.

                      Comment


                      • #12
                        If your losing 20-30 modern tanks against a single mech infantry your doing something TOTALY wrong buddy.

                        Try Blasting the city to the ground with bomber/artillery. THEN you attack with Modern Armour.

                        In my campain vs the Germans (in warlord) I took out 5 of their Mech infantry in a size 25 city with 3 modern tanks, 1 mech inf. 6 artillery and 8 bombers. (not to mention the handful of cheesy riflemen/infantry garrisoned they hadnt bothered to upgrade yet).

                        Ive noticed that especially in modern warfare you must BOMBARD the HELL out of the enemy to pull off a successful attack.

                        I Really enjoy this. No more clean sweep, destroy a civ in 1 turn warfare. You gotta really lay siege to the enemy!

                        Comment


                        • #13
                          Originally posted by Dan Magaha FIRAXIS

                          I predict when the patch is released a lot of speculation about what bonuses the AI receives will be put to rest...

                          Dan
                          Thanks, Dan for giving hope on the game and us.

                          I don't want to put the blame on Firaxis too much on the playtesting deficiency. It happens in the PC gaming industry. But don't try making it as a norm or habit.

                          Comment

                          Working...
                          X