Combat system - statistical stuff from >20 games

Originally posted by Lord Thorium
Whether the AI gets an unfair advantage or not, one thing is clear to me. The game should not rely on randomness so much. Doesn't it just make you want to delete the game when you decide to attack a weaker civ only to see your tanks slaughtered by their cavalry because you are "unlucky" this time around?

Where was your infantry support?
 
Originally posted by Lord Thorium
Whether the AI gets an unfair advantage or not, one thing is clear to me. The game should not rely on randomness so much. Doesn't it just make you want to delete the game when you decide to attack a weaker civ only to see your tanks slaughtered by their cavalry because you are "unlucky" this time around?

Tanks will often lose to attacking cavalry in the open. Tanks may look like they are invincible, but their defense rating is only slightly higher than the attack rating of cavalry. I will not restart a game for this reason. It would be, after all, my own fault for underestimating my opponent. A learning experience for you I hope.
 
The Civ3 combat calculator shows that a veteran Cavalry has 46.7% chance of defeating a regular. Since your tanks will often be damaged after their attacks in your turn, the odds will often be even higher. Nothing special here.

Also note that the AI will often concentrate on the easy battles - put an undamaged infantry on the tank and they won´t attack at all. This could be one reason people complain, because people are used to AI:s that make mad rushes against any enemy at any odds. The AI won´t do this in Civ3.
 
The Germans acknowledge the vulnerability of their armor in France 1940. After the war General Paul von Kliest wrote about the French defences during the Western Offensive.

Along the Meuse there was a moderate amount of fortification, in the way of pillboxes, but these were not properly armed. If the French troops here had been adequately equipped with anti-tank guns we should certainly have noticed it, as the majority of our tanks were of the early Mark I type, and thus very vulnerable! The French divisions in the sector were poorly armed, and of low quality. Their troops, as we repeatedly found, gave up the fight very soon after being subjected to air bombing or gunfire.
 
LaRo or anyone else who can help... Is there anyway we can download the combat calculator that was referenced?

I'd like to use it while I'm playing, but I have dial-up access and only one phone line in the house... could be mutiny by my wife if I tie up the phone too!
 
The only hole in that Zachriel is that Civil War Cavalry (which is what the Cavalry look like and which I assume they are modeled after since you don't get tanks for a full age's worth of tech later) did not own Anti-Tank guns. Nor could a person use a WWII Anti-Tank gun while mounted on a horse. IIRC the early AT guns were very heavy and had problems with barrels melting and what not. I'm not a war buff nor a history major, so I'm basing this on vague recollections of the History Channel shows.

But despite my lack of historical knowledge, I still know that since tanks weren't around during the Golden Age of Cavalry, that therefore Anti-Tank weapon could not have existed. Unless someone had the bright idea to invent AT guns before Tanks. Which I severly doubt.

The reason I say this is because your quote says that IF the French had ANTI-TANK GUNS, then they would have noticed. So Cavalry taking out Tanks is somewhat unlikely. However, I will readily admit that game mechanics speaking, an attack of 6 versus a defense of 8 can win since the odds aren't too bad.
 
There`s been much whining here about the combat results and so...

Great thread Killer!

I think Civ3 is a great improvement on Civ2 and a real advancement of the genre, but this *^&%#! combat issue infuriates me each time I play. No combat function is ever going to achieve absolute realism and we're all prepared to handle the occasional pikeman-defeating-the-tank. It happens, we got over it in Civ2, and we'll get over it again in Civ3. No big deal. It's when inferior AI units consistently defeat my numerically and qualitatively superior units, often when I have the terrain advantage, that there's a problem. This did not happen in Civ2.

This issue's been kicked around a lot already and others have said that they don't even see this problem in their gameplay. Is it possible that some copies of Civ3 are more prone to this than others, a sort of random bug?

I see Dan's exasperation in his reply to this thread and I want to reiterate that I think Civ3 is a great game with additional functions and features I'm very happy about - but this combat issue is very real. I half-considered doing what Killer did, doing a statistical tabulation of the outcomes of all combats in a single game, but he beat me to the punch. (Great job Killer!) I'm certain my results would be similar to his, if not even worse.

As for those who want to play an entire game without a war, aside from the fact this would be completely unrealistic historically and not much fun (for my gameplay, anyway), about the only thing you can do I imagine is play a game without any other Civs...
 
Originally posted by DarkwingGT
did not own Anti-Tank guns. Nor could a person use a WWII Anti-Tank gun while mounted on a horse.

The reason I say this is because your quote says that IF the French had ANTI-TANK GUNS, then they would have noticed. So Cavalry taking out Tanks is somewhat unlikely. However, I will readily admit that game mechanics speaking, an attack of 6 versus a defense of 8 can win since the odds aren't too bad.

The game is merely an approximation. The U.S. 1st Cavalry Division in WW2 fought dismounted, for instance. Cavalry units are not going to use a saber charge against tanks, but can develop tactics appropriate to the challenge. As far as being able to carry anti-tank weapons while riding, horses were used to pull artillery and other weaponry throughout WW2.
 
I didn't mean just tanks, I meant modern armor as well. And that was an example. Sometimes it happens when the odds are 75% or more in your favour, but you still lose a few times in a row. In my opinion randomness should not play that big a part. It should be reduced to the alpha centauri level, where you could generally predict which way a battle would go, if the units were unequal.
 
Originally posted by Lord Thorium
In my opinion randomness should not play that big a part. It should be reduced to the alpha centauri level, where you could generally predict which way a battle would go, if the units were unequal.

This is a matter of taste. I hated the SMAC approach, where you could feel pretty confident that you´d win a battle if you had 52% chance of winning, and lose it if you only had 48%. Civ3 is much more random, but don´t expect Firaxis to change it. So I suggest you change it in the editor or download some of the mods around.
 
Originally posted by Hurricane


This is a matter of taste. I hated the SMAC approach, where you could feel pretty confident that you´d win a battle if you had 52% chance of winning, and lose it if you only had 48%. Civ3 is much more random, but don´t expect Firaxis to change it. So I suggest you change it in the editor or download some of the mods around.

I agree. To be "realistic" there has to be some chance of failure. Plus any degree certainty comes from proper planning and battle tactics.

Bombard, bombard, bombard.
Like this: http://www.crowncity.net/civ3/Attack.htm
 

Attachments

  • aroundicon.jpg
    aroundicon.jpg
    7.3 KB · Views: 107
Originally posted by Lord Thorium
I didn't mean just tanks, I meant modern armor as well. And that was an example. Sometimes it happens when the odds are 75% or more in your favour, but you still lose a few times in a row. In my opinion randomness should not play that big a part. It should be reduced to the alpha centauri level, where you could generally predict which way a battle would go, if the units were unequal.

The change between SMAC and CivIII is the number of hitpoints for units. Boost up the number of hitpoints for all units, and you will reduce the chance of getting an improbable outcome (since the outcome is now even more improbable). The reduced number of hit points is the only reason why CivIII appears to be more random than SMAC.
 
Originally posted by etj4Eagle


The change between SMAC and CivIII is the number of hitpoints for units. Boost up the number of hitpoints for all units, and you will reduce the chance of getting an improbable outcome (since the outcome is now even more improbable). The reduced number of hit points is the only reason why CivIII appears to be more random than SMAC.

TRUE! I nominate this to the best post of the day. :king:

Now, try to convince the conspiracy believers of this simple fact. :(
 
I don't want this to turn into an arguement so I'll keep it simple.

Zachriel, of course I didn't think cavalry would sabre charge tanks. That would be Horsemen and Knights. It's assumed by Cavalry that they would have rifles. But regular rifles are not capable of damaging a modern tank. Plus, I didn't say horses couldn't be used to pull wagons and what not, but a person would be hard-pressed to even aim an anti-tank gun (like the .50 caliber Browning M2) which weighed in about 50 lbs. (or something like that or more) while on horse back.

Ok, back to the matter at hand. Originally, I was somewhat skeptical of the Spearman defeating the Modern Armor. That was until Pikemen started fending off my Modern Armors. I couldn't believe it. And since a Spearman has a defense of 2 and pikeman a defense of 3, that 1 point can't make that much difference against 24 attack, I now believe it. My gripe was always on much smaller scales, like Immortals never being able to kill Spearmen. But I'm a believer now. Hallelujah, I have seen the light....and it's kind of scary.
 
etj4Eagle wrote: The change between SMAC and CivIII is the number of hitpoints for units. Boost up the number of hitpoints for all units, and you will reduce the chance of getting an improbable outcome (since the outcome is now even more improbable). The reduced number of hit points is the only reason why CivIII appears to be more random than SMAC.

Is this a viable solution? It appears Firaxis is in denial over this, so Civ-ers will have to develop a patch or a solution themselves. Where do we find these files?
 
Than you all for replying! (especially Dan at Firaxis!)



I still get the distinct feeling that the random number application to the combats or random number creation is somewhat buggy. But that seems to be limited to incredibly long threads of high or long numbers. That means you`ll "always" loose for a while (took me 2 1/2 month to get through that), then you might "always" win (have been in that phase for the last 10 games :D ).

Here`s the same calc as before for the last 10 games:

520:386 on +/- 50:50 battles.
only change: I did this in modern times by blocking iron and rubber and all that and disabling Riflemen. Thus i was able to chek for invisible City walls. They are pretty common; actually almost every time a town with walls grows to size 7 and then goes below 6 the walls dissapear from view in my games. Thus the last outcome was probably heavily influenced by this!!!


Firaxis: PLEASE fix the No-Map-Update-Bug! it simply sucks when there`s roads i don`t see, roads aren`t there. that i see, and town walls work that aren`t displayed!!!!



Aside from that I still feel:

1: it is highly advisable to multiply HP by 3
this will keep unlikely results rare and the waiting down to acceptable. (I actualy switched animations off; sometimes I miss the results from a fight but usually it doesn`t matter)


2: the random number generator produces "strings" that are strange
even if they level out it is pretty irrealistic when you get rare strings one way a lot, then rare strings the other. I`d actually like to see more combat where the strings are more in the middle (as in theory they should be).
 
First, I'm one of those players who has not had any sort of problem with combat results. Maybe that taints my view. I'm also the author of the CivII combat guide, which most certainly taints my view... Anyway, I realize some players just have had bad luck, don't take what I say too harshly!

Civ3 combat does not give buggy results. With firepower gone and hitpoints reduced, combat is simplified. Mathematically, that just means that the unpredictable result is more likely than players were accustomed to in CivII. Remember, with even A/D values (e.g. archer vs. spearman), the attacker can expect to win less than 1/3 of the time. It's simply how it works out. To win, you need an advantage, often a significant one.

Civ3 combat requires the player to use bombardment, numbers, and in general more strategy than in the previous civ games. I happen to think this makes for a better game. But then, I don't mind the occasional unbelievable result - if I knew ahead of time how everything ended, I wouldn't waste my time playing.

Will a spearman defeat a tank? It can, and once in a blue moon will. Do you remember the 4326 times it loses to the tank? No, the one exception is what stands out in somebody's mind. Besides, your pc does not compare tank vs. spearman, it compares 16:2. It is only unintuitive to a human mind. If you need a realistic explanation, just imagine that those spearmen were led by that civ's greatest strategic mind, and your tank was led by a bumbling fool who was arrogant enough to think his tech advantage would sidestep the need for careful tactics. He deserved to die! ;)
 
hey zachriel! go buy a lottery ticket! right now! You must be the luckiest SOB in the world! :)

sodak: I trippled my HP!!!!! And if you`re so tired of it, why are you here????? I`m currently enjoying results way left of the good side of good - and i don`t like it! It`s not about whining but about finetuning CivIII so that high-level play is fun!

Don`t you see that: 1 unrealistic combat on deity geives a huge advantage to the guy who produces faster. sicne that`s the AI the games get rather frustrating!
 
Back
Top Bottom