The fix for the warrior defeating the tank

[lorddread]

How many people who play the game have had this happen to you:

You are in the middle of a war, and a warrior (armed with a flint axe), defeats an infantry unit or worse a tank. The straw that broke the camel's back was when I lost a advanced armor unit to a warrior.

The fix for this is to reset the offensive and defensive scores to keep this from happening.

Let's face it, a Abrhams M1A2 main battle tank is the Sh#t! I set the defense to 60 and offense to 50. Using this as the top end of the scale, I moved all the other units offense and defense, ending with the warrior with a 1-1. This way no matter how damaged my tank is, it cannot be defeaed by a unit that can never even scratch it.

The game creators should have done this to begin with, but hey!

Let me know what you think!

[steelehc]

True, a dude with an axe or club wouldn't cause much (read: any) damage to a tank, but
1)This has been b1tched about many times already.
2)Increasing the HPs for each experience level helps.
3)So does increasing the attack and defense values for each unit.

There are a number of mods available that deal with this (and other) issues.

Steele

[Inverse Icarus]

the best way to revamp the combat engine is to add more hitpoints to the modren units and to bring back firepower.

it's been beaten to death. don't re-open this war.

[Zachriel]

Originally posted by lorddread
I set the defense to 60 and offense to 50. Using this as the top end of the scale, I moved all the other units offense and defense, ending with the warrior with a 1-1. This way no matter how damaged my tank is, it cannot be defeaed by a unit that can never even scratch it.

That is incorrect, even with 50 on attack, there is a small but nevertheless real possibility of the combat still having an unexpected result. The better strategy is to assume when there is war that there will be the possibility of losses.

BTW, it's never happened to me. I never see warriors in the modern age. But I use combined arms, and plan my attacks carefully, so I wouldn't expect it to be much of a problem.

[Thrawn05]

Has anyone thought that perhaps some of the tanks in the tank unit broke down and so warriors were able to get in and kill the pilot?


This whole ranting about Tanks or what getting killed by older units is getting dull IMHO (I side with Uber). I have yet to see this happen. The most I've seen was my Tank getting killed by a Riflemen, but that was because the Tank had 2 health points left and the Riflemen was elite.

[kimmygibler]

"This way no matter how damaged my tank is, it cannot be defeaed by a unit that can never even scratch it."

Not true. It could still happen. Don't you understand random numbers?

[Ethelred]

Most people don't understand random numbers or statistics.

Thats why they think there is something perverse going on when random weird things happen. May even be the main reason for religion. People are built to try to understand why things happen. Sometimes things are random though an no amount of trying to understand the reasons will change it.

In fact the universe is statistical to the core. However assuming something is random when it may not be is a sure way not figure out the cause. So human have evolved to see patterns even where there are none.

Then again when a spearman in the real world beats a bunch of Brits armed with Martini-Henry rifles with men armed with assegi it wasn't exactly a random event. Someone on the Brittish side blew it. In that case though the numbers were sufficient that the statistics reached high probabilities. In Civ we are dealing with only one unit vs another each time which allows for apparent anomalous results. Put a big stack of tanks against a equally big stack of spearman and the tanks will win every time.

No I haven't seen the legendary anti-tank speaman myself either. Though I thought I might last week though. My veteran tank got knocked to one hit but it was no ordinary spearman it was a hoplite in a city on a hill.

[Thrawn05]

Originally posted by Ethelred
In fact the universe is statistical to the core. However assuming something is random when it may not be is a sure way not figure out the cause. So human have evolved to see patterns even where there are none.

You should read that Steven Hawking's book.

There're a few chapters that talk about that. If the universe was not random, then everything would be symetrical, meaning each star is equaly far apart from one another.

But since you don't see that, there is some form of randomise to the universe. Chaos is the only constant in the universe. Well, that, and Murphy's Law.

[Ethelred]

You should read that Steven Hawking's book.

If you are talking about A Brief History of Time I have it. Not only that I read it.

Try John Gribbons books for more depth without a huge quanitity of math. I did find his books a lot more work to read though. Twenty pages at a time. Then sleep on it. On the book sometimes or the couch or the chair. Occasionally on the bed. Where ever my mind decided it had enough and was going to shut down like XP does to protect the system with a BSOD.

[SpencerH]

I believe that the physicists definition of "randomness" relative to the Universe and "random" from a statistical perspective are not the same thing. Its been a while, but I believe that Hawking was pointing out the existance (on some level) of forces that make the universe "non-random".

As I understand it, combat in CIV3 is not controlled by a random number anyway. It was pointed out in another thread (it might have been you Ethylred) that there is an assigned number for each turn that is used to generate the "random number" for combat. From a statistical perspective the randomness of the "random numbers" can only be assessed with a large population such as over the course of a game. But that doesnt change the observations that some very anomolous combats occur that are unlikely with completely random numbers.

[Thrawn05]

A month ago I wanted to test the limits of the combat system by making a 100/100 unit (just to see if attack/defend number REALLY matter). What I found was that those edited units still died on the field, but the problem was how long until they meet their end.

Those 100/100 units killed most units and until riflemen, were undefeated. By the time MechInfantry game, those units died on occasion. I do believe that there is SOME higher power working here in the game mechanics.

[Ethelred]

Originally posted by SpencerH
I believe that the physicists definition of "randomness" relative to the Universe and "random" from a statistical perspective are not the same thing. Its been a while, but I believe that Hawking was pointing out the existance (on some level) of forces that make the universe "non-random".

Same for me regarding the time. I read it shortly after it came out. Gravity might have been what he was talking about. Its an organizing force. The universe would have needed to be VERY uniform for gravity not to have caused clumping. Very as in less than 1 part of non-uniformity in 10,000 which is about what COBE found.

Don't quote me. I didn't look that up so its likely wrong.

As I understand it, combat in CIV3 is not controlled by a random number anyway. It was pointed out in another thread (it might have been you Ethylred) that there is an assigned number for each turn that is used to generate the "random number" for combat. From a statistical perspective the randomness of the "random numbers" can only be assessed with a large population such as over the course of a game. But that doesnt change the observations that some very anomolous combats occur that are unlikely with completely random numbers.

Not me. Well I wouldn't have said it wasn't random in that sense. I may have said something that was close to that. Here it is in more detail as I understand it.

The game stores a SEED number for each save to start with. The initial seed is generated by using something that is sorta random. I thinks its a timer that overflows very quickly from the time Windows, perhaps the machine, starts up. It was a hardware timer on the Apple ][ and that the last thing I programmed on.

Apple by the way had a busted random number routine courtesy of Bill Gates. Bill's first comercial Basic was Applesoft and there was a problem that caused the numbers to become in someway unrandom after around 10,000 numbers were generated in a row. I think they started repeating the same sequence no matter what the intial seed was after that. Simply getting a new starting SEED would work around the glitch.

There is no way to get random numbers like a roulette wheel or better yet a bit of radioactive material in a PC. An algorithim is run that generates numbers that are not predictable. The only way to find out what the next number will be from a given seed is to run the algorithim.

The Random Number Generator functions in PC's generate numbers that are pretty darn near indistiguishable from true random numbers. Each number that is generated after the start become the seed for the next.

There are two reasons people think Civ III is not using random numbers. One is that they just don't like the results of some battles and fixate on what they think is an anomalous result. This is normal human nature and it can be hard to avoid thinking that way. The other thing is when people try the Civ II reload cheat they get the same results they had the first time.

That same result is due to the SEED for the Random Number Generator. In Civ II and the original a new seed was obtained from the timer function each time you reloaded. Now in Civ III the SEED is stored with the Autosave. Apparently if you do a reboot you will get a new SEED.

Psuedo Code

Start Civ III

SET initial SEED variable with a number from timer function

DO
Run Random Number Routine using SEED variable
Save generated RN as SEED
Use RN for game function
LOOP

(the first language I learned was Fortran and DO LOOPs are still whats in my head)

IF save game then include SEED in save

Gamer hates result of RN

Gamer LOADS old save from start of turn

DO

Run Random Number Routine using SEED from the save.

THEN
Cheating Gamer rants the game is cheating

So the number is as random as it can be on a PC. But it is the same at the start of a turn even if you reload it because you are reloading the same SEED used the first time.

Simply changing the order of battle gets you a diffent number. So if you want to use the reload cheat you must change the order of events. Given the same seed the same sequence of numbers will be generated every time. After all it is running on a machine that is supposed to get the same results when it does the same thing. Only the SEED can be different and it won't be in Civ III from a reload.

I don't know if it was really a good idea or not for Firaxis to do this but just changing the order works around it. Of course if you only have one thing to do that requires RN to be generated you will have to add something that uses one. Make an attack with a unit that wasn't involved the first time for instance.

Keep in mind that loosing a unit may have involved many RN's and if you lost with each of them they must all be used up. Its not a stack of numbers that is saved just the intial seed but the sequence will not change. What you need is a very different place in the sequence especially if its a tank VS a spearman as that means the numbers were very bad indeed so not unit can be expected to win that particular battle. Also if the battle went on for a while with both units missing a lot trying to attach with a weak unit isn't going to clear out the unlucky sequence. More than unit may be needed to use the sequence.

This last may be why some people think there is more going on than a SEED save. They try using up the unwanted results with a unit that has one hit. That may use only only one number out of a sequence of many badly losing numbers.

[SpencerH]

I wonder if we have "all our ducks in a row" with regard to the combat calculators? i.e. Is there a little "checkbox" somewhere that we've missed (such as weakens the barbs against other units) that kicks in every once in a while?

Heres the observation:

In a combat between elite warriors (delete the 10% defender advantage) with random results you would expect to see HP being deducted from one unit then the other. Of course you can get two to five point HP reductions in a row on one side or the other but to loose 5HP in a row should be less rare than what seems to occur. Of course it could be perception but I think there is some "clustering" of random numbers that allows these events to occur more often than is purely random. Argueably, there could be a reason for this. In combat the odds for success change as hits are scored (as you wound your opponent your chances of victory increase while theirs decrease). I wonder if something like this is present in the CIV3 combat calculations? If so, its faulty since I most often observe this combat result clustering when a unit beats its opponent down to 1 or 2 HP then losses 5 or 6 rounds in a row (and dies).

[lorddread]

Unfortunately, there is no such thing as a random number using a modern day computer. The US Government found this out when they tried to use a random number generated on a modern computer for cryptography.

The programers need to make the following fixes to the game to ensure problems like these go away.

1. Obsolete units need to go away. I do not want the option to make a warrior everytime I want to build a F-15.

2. Ancient units, (if they are still around) should not be able to inflict ANY damage on modern units, NO MATTER IF THEY ARE ELITE OR NOT!

3. Some serious thought needs to go into evamping the random seed. Is it necessary? IMO no!

[Thrawn05]

Originally posted by lorddread
Unfortunately, there is no such thing as a random number using a modern day computer. The US Government found this out when they tried to use a random number generated on a modern computer for cryptography.

/* RAND.C: This program seeds the random-number generator
* with the time, then displays 10 random integers.
*/

#include "stdlib.h"
#include "stdio.h"
#include "time.h"

void main( void )
{
int i;

/* Seed the random-number generator with current time so that
* the numbers will be different every time we run.
*/
srand( (unsigned)time( NULL ) );

/* Display 10 numbers. */
for( i = 0; i < 10;i++ )
printf( " %6d\n", rand() );
}

[Ethelred]

Unfortunately, there is no such thing as a random number using a modern day computer.

I said that. Well I thought I did anyway. Its a psuedo random number. Close enough to the real thing for most uses. I would like to know what kind of ignoramous in the government tried useing it for random numbers in cryptogrophy.

We are not talking about the average person there. Cryptogrophy helped create the computer. The people that do it understand random numbers and they are one of the main groups that came up with methods of testing just how random a set of numbers was.

Psuedo random numbers have slightly different properties than the real thing and even if I knew what those were I am pretty sure I couldn't explain it. Number theory was an exceedingly arcane area of mathamatics untill people noticed it had use in cryptography and communications. The Russians didn't even stop it from being published for a long time. One of the guys I used to play D&D(not to be confused with AD&D) with was specializing in number theory because it went with all the programming he was learning.

Frankly I don't see the option for warriors when I have F-15s. Perhaps if you installed the 1.17f patch this apparition would go away. Some others will remain though. I don't notice them anymore.

Perhaps its because I use a modern mouse with a scroll wheel. Handy dandy things scroll wheels. I am going to wear out my middle finger with them.

I disagree with your second point. They barely scratch anything as it is most of the time. Even in Civ II they could harm modern units occasionally.

Besides obsedion axes are really sharp. Eye surgery is done with glass blades. Really. Glass is sharper than steel. It doesn't exactly hold an edge well though. Frankly I don't think a hand axe is going to do much to a tank unless you get the driver to meet your sister and then slit his throat.

Have you ever noticed that the combat system is intentionaly abstract. Haveing helped write D&D rules I know just how much useless detail can be added. I once ran an major battle that took hours to adjudicate. One of the players noticed but carefully neglected to mention that the odds of the Kobalds hitting them was actually one percent and that I could save a lot of die rolls by just using two ten sided dice and ignoring all the other stuff I was doing. I think he didn't want me to adjust the odds. Anyway I could have saved a lot of time if the system had been a bit more abstracted.

To make the D&D thing clear.

D&D is generic. AD&D is trademarked. In other words I did not work on any official rules. The original rule set was near useless and fans had to change things. My group changed things frequently and with a great deal of depth. Page after page.

I see by your third point you are unrepentent reloader. You must change your ways. Or hair will grow on your palms. Its growing in my ears allready.

[dubwai]

A tank can be destroyed by a single soldier. A soldier with a rocket, molotov cocktail, or a .50 caliber rifle can take out a tank. Of course, a stone ax and a .50 claiber rifle are very different things, but because there is no way for civs to buy or otherwise acquire weapons, this isn't easy to represent in the game.

[Grey Knight]

I don't know if it got posted here (it was posted to www.civfanatics.com a few weeks ago), but some enterprising fellow hooked a debugger up to the Civilization3.exe, and extracted some large sets of random numbers.

The numbers *are* random (more on this later).

It's our misperception of the combat model which is at fault. Try this -- which of these battles does the attacker seem more likely to win:

A2 vs D1
A8 vs D4

It's like an optical illusion. We 8 is much bigger than 4, and we think "8" should take the day. 2, only being 1 more than 1, doesn't really have a big advantage, right? Wrong of course - the odds for both battles are exactly the same, since it is the *ratio* between the two numbers which matters, not the distance between the numbers. As one poster pointed out, even a 100/100 unit will loose on occasion.

If you want to see this in action, set the unit HPs to 9, 10, 11, and 12, in the editor and start a game. You'll see the unit with the numerical advantage (after defense bonuses) almost always win, and the "winner" will almost never come out unscathed (loosing their ratio of HP based on the A/D values). For example, an archer attacking a fortified spearman:

A:2 D:2, +35% for 2.7

Overall odds (from the calculator):
25% 75%

Likely HP left for the spearman: 4

Cheers,
Shawn

[ALPHA WOLF 64]

from my experience, my fortified spearmen lose far more often than 25% of the time. I believe that there are factors coded that we havent figured out yet. I think there s a moral type influence, because as another thread once commented on, the AI fights much better earlier in a war than it does after suffering some defeats. I started a war with a stack of veteren knights against swordsmen. All the AI cities were about the same size and all defended with 2 or 3 swordsmen. There were no other noticeable factors such as fortified vs non fortified or terrain bonuses. The first few cities I sustained heavy casualties. But as I worked my way into the inner cities, those same swordsmen were causing far less damage. many of my knights didnt even take a single hit. I tried replaying those turns by altering attacks and diplomacy to get a new set of "random" numbers, but the result was the same, hard won early victories followed by pushover victories.

[Ethelred]

About the same size is not quite the same as if they actually were the same. There is a break point between town and city. One pop difference is enough to significantly change the defensive modifier. Population six is much easier to defeat than seven. Same for 12 vs 13.

[ALPHA WOLF 64]

Originally posted by Ethelred
About the same size is not quite the same as if they actually were the same. There is a break point between town and city. One pop difference is enough to significantly change the defensive modifier. Population six is much easier to defeat than seven. Same for 12 vs 13.

I know this, and was trying to avoid mentioning that the cities were all in the 6-8 range because I'd increased town size to 8 and city to 16. All the towns were classified as towns for defensive purposes .

[Venger]

Re: The fix for the warrior defeating the tank

The best fix is to go to File -> Run -> civ2.exe

Venger

[Thrawn05]

Originally posted by dubwai
A tank can be destroyed by a single soldier. A soldier with a rocket, molotov cocktail, or a .50 caliber rifle can take out a tank. Of course, a stone ax and a .50 claiber rifle are very different things, but because there is no way for civs to buy or otherwise acquire weapons, this isn't easy to represent in the game.

So you're saying a warrior that killed a riflemen, never bothered to put up some rifles? You would think that they would pick up. Or is firaxis saying that warriors are inferior people and therefore stupid?

[The Rook]

Originally posted by dubwai
A tank can be destroyed by a single soldier. A soldier with a rocket, molotov cocktail, or a .50 caliber rifle can take out a tank. Of course, a stone ax and a .50 claiber rifle are very different things, but because there is no way for civs to buy or otherwise acquire weapons, this isn't easy to represent in the game.

The way I would have did it would be to change the graphic. Kinda like the workers do. Flint axe warriors should look like rag-tag soldiers or partisans. At least when the 1.1.1 partisan kills your modern armor, you can picture the guy building a homemade booby-trap or something, and destroying the tank. The guy with the fint axe, just doesn't cut it.

[Grey Knight]

I think I have the answer. I'll make a mod. I'm not sure what to call it though. "No spearmen killing tanks here!" is a bit long.

I'll rename all the ground combat units to "Tank [fill in the blank]". A/D/M etc. will all be the same.

This way, since every unit is a tank, you can't have a spearman killing a tank any more.



Shawn

[lorddread]

Originally posted by Thrawn05


/* RAND.C: This program seeds the random-number generator
* with the time, then displays 10 random integers.
*/

#include "stdlib.h"
#include "stdio.h"
#include "time.h"

void main( void )
{
int i;

/* Seed the random-number generator with current time so that
* the numbers will be different every time we run.
*/
srand( (unsigned)time( NULL ) );

/* Display 10 numbers. */
for( i = 0; i < 10;i++ )
printf( " %6d\n", rand() );
}

Alas even this is not truly random. If you did this enough times (as the Feds did) you would find that a pattern would emerge. But I digress.

The point I was trying to make was that instead of a random seed generator, why not use straight attack-defense ratings for combat and throw in initative as a multiplier. This was a tank (which can move 30 miles an hour and has 75mm if armor plating can vaporize an elite hoplite.

[simwiz2]

Originally posted by The Rook


The way I would have did it would be to change the graphic. Kinda like the workers do. Flint axe warriors should look like rag-tag soldiers or partisans.

Partisans: We already have them. They are the conscript units.

"At least when the 1.1.1 partisan kills your modern armor, you can picture the guy building a homemade booby-trap or something, and destroying the tank."

A tank unit is alot of tanks. So unless they manage to build a LOT of sucessful booby traps...

[Ethelred]

Originally posted by lorddread


The point I was trying to make was that instead of a random seed generator, why not use straight attack-defense ratings for combat and throw in initative as a multiplier. This was a tank (which can move 30 miles an hour and has 75mm if armor plating can vaporize an elite hoplite.

And the hoplite steps to the side evading the tank. Or the tank is attacking a tank. Everything is fixed yet it mustn't be so or you too achieve unreality. Or the hoplite steps aside of the iron spike filled trap it was standing in front of for that matter. Even the warrior can manage that one.

Wanna hear about incendiary pigs? Or Samurai that committed suicide en masse. 200 of them. For real.

You can't have a combat system with fixed results for all possible opponents and be realistic. Probability is inherent in combat. Things ALWAYS go wrong or even better than expected. They nearly never go exactly as planned.

I am not saying a better combat system can't be devised. Just that one without probabalistic results won't be a better one.

[SpencerH]

Originally posted by dubwai
A tank can be destroyed by a single soldier. A soldier with a rocket, molotov cocktail, or a .50 caliber rifle can take out a tank. Of course, a stone ax and a .50 claiber rifle are very different things, but because there is no way for civs to buy or otherwise acquire weapons, this isn't easy to represent in the game.

Only by a very very brave soldier!

A .50 cal machine gun would do nothing to any tank built around the time of WWII or after. Armoured car or APC yes but not anything recognizable as a tank.

You can't have a combat system with fixed results for all possible opponents and be realistic. Probability is inherent in combat. Things ALWAYS go wrong or even better than expected. They nearly never go exactly as planned.

Theres a truism for y'all. Its the level of the "imponderable" (as they used to say in the Avalon Hill days that I've (we've) had to fix with the editor.

[Zachriel]

Originally posted by SpencerH
Only by a very very brave soldier!

A .50 cal machine gun would do nothing to any tank built around the time of WWII or after. Armoured car or APC yes but not anything recognizable as a tank.

Theres a truism for y'all. Its the level of the "imponderable" (as they used to say in the Avalon Hill days that I've (we've) had to fix with the editor.

Brave soldiers, willing to die, do exist. The tanks in Civ3 are certainly not as advanced as panzers, so they are vintage 1920's or 1930's technology.

Traps and explosives are most effective against tanks, unless you have "proper" anti-tank weapons.

And yes, there is always the imponderable, unpredictable element in warfare.