WTF: no firepower?

[Grumbold]

Originally posted by korn469
KrazyHorse

ok basically this is a statistics problem correct?
the larger the number of rounds the smaller the cofidence interval and all of that jazz right? so if the hitpoints are high enough (like if the units had 30k, 40k, and 50k hitpoints) then when a spearman attacks a tank it would 99.9% of the time inflict 32% damage for example (the 32% damage number was a made up number)

however, after playing a thousand games the number of times a spearman loses to a tank should be about the same no matter if they had .3, .4, .5 hitpoints or 30k, 40k, 50k hitpoints right?

No, sorry Korn. Every hitpoint that needs to be removed reduces the chance that the less likely event will occur. A 60-40 chance is pretty risky with 1 hp. Every time you repeat the chance, the more likely it will be that the 60% has come up more times than the 40%. So 1000 1 hp fights will come out appriximately 60/40 but 1 1000hp fight should come out in favour of the stronger opponent ~99% of the time, the winner having a spread of results centered around 400 hp left.

[Deathwalker]

I personally feel that the Firepower unit option was never fully used in Civ 2. After all there is not way a Phalanx no matter how unlikley is going to damage a tank. So the fire power should ahve been a lot higher in comparisaon.

[Bleyn]

Originally posted by Grumbold


No, sorry Korn. Every hitpoint that needs to be removed reduces the chance that the less likely event will occur. A 60-40 chance is pretty risky with 1 hp. Every time you repeat the chance, the more likely it will be that the 60% has come up more times than the 40%. So 1000 1 hp fights will come out appriximately 60/40 but 1 1000hp fight should come out in favour of the stronger opponent ~99% of the time, the winner having a spread of results centered around 400 hp left.

Grumbold, Thank You! Thank You, Thank You, Thank You!

I knew there was a flaw in his reasoning, I even sort of knew what it was, but as how to explain it properly, I wasn't sure I could do it. I think your explaination sums it up very well.

The way things are now, there is that freak chance that a much weaker unit will win. And rightly so. I've said it before and I will say it again, history has shown far too often that superior weapons do not equal victory.

Increase the hitpoints enough and that starts to go away. Increase them too much and in a one on one fight the stronger unit will always win, and wheres the fun in that? The game would just devolve back into the arms race that Sid and company are trying to lift it out of.

[Bleyn]

I have a question on the combat system that make me wonder if we really have all the data we need to evaluate it properly. I've looked, here and elsewhere, and I can't find any good answers on them.

Are the attack-defense numbers in combat actually calculated the same way as in Civ2? Or has it changed? Let me explain what I mean, and maybe someone has some good info on how it will really work.


In Civ2, combat was calculated by taking the modifed attack number of the attacker (A) and the modified defense number of the defender (D) and using the formulas:

(A + D) = X
A/X = attackers chance of winning each round of combat
D/X = defenders chance of winning each round of combat

With Civ2 combat, every round during combat had a winner, and damage was assigned to the loser until one or the other died(or in rare cases retreated).


With what little I'd see on the mechanics of Civ3 combat (and having forgotten how it worked previously), I'd been thinking it would work slightly different in Civ3. I'd thought that the calculations would work like this:

Aa = Attackers attack rating
Ad = Defenders attack rating
Da = Attackers defense rating
Dd = Defenders defense rating

Aa/(Aa+Dd) = Attackers chance of hitting the defender each combat round
Ad/(Ad+Da) = Defenders chance of hitting the attacker each combat round

And with these formula, I was seeing four possible results each round of calcuations, attacker hits, defender hits, both hit and neither hit.

To give an example of the difference between the Civ2-style system and the system I was envisioning lets take a battle between a Cavalry and a Musket using the Civ3 stats stated earlier in this thread. Cavalry 6.3 and Musket 3.3

Civ2-style:
6 + 3 = 9
6/9 = 66% chance for Cavalry to win each round
3/9 = 33% chance for Musket to win each round
2/3 of the rounds Cavalry scores damage
1/3 of the rounds Muskets score damage

What I thought:
6/(6+3) = 66% chance for Cavalry to hit in a round
3/(3+3) = 50% chance for Musket to hit in a round

In 1/3 of the combat rounds, both units would hit for damage.
In another 1/3 of the rounds, the cavalry would hit and the musket would miss.
In 1/6 of the rounds, the muskets would hit and cavalry would miss.
And in the final 1/6 of the rounds, both would miss.

The Musket has a somewhat greater chance of damaging the Cavalry, but in the long run, Cavalry would still win most of the time.


To me this second system made sense. It fit with what was given to us about combat and hitpoints in the GameSpot article. It would extend battles more than the older method would with what we know right now about hitpoints. Admittedly, there is the slight chance that a combat could take an extremely long time to finish if both sides kept missing, but I don't think that would happen very often. Eventually someone will win, or someone will retreat.

And it made sense historically. Battles do not always have a clear winner and loser. Sometimes the battle ended in a draw. Sometimes it ended with both sides so shattered that neither could be said to have a cohesive fighting force anymore and no-one really won.

[Grumbold]

Originally posted by Deathwalker
...After all there is not way a Phalanx no matter how unlikley is going to damage a tank. So the fire power should ahve been a lot higher in comparisaon.

Please read my post at the tail end of page 1 of this thread. There absolutely are many ways that tanks can be defeated by hopelessly underequipped infantry, it just takes the correct approach. Now in broad daylight on a flat plain with unopposed vision for 2,000 metres, I'll give you the 100% tank win scenario.

[Jason Beaudoin]

Grumbold: You keep talking in terms of one phalax against one tank... but if you want to get technical on the subject, I think each unit represents a plattoon or a group of many tanks or many phalaxes.

Now... if you were to put a real live scenario in place and pitch a group of tanks against guys holding spears, I don't think the spears would do very much against tanks.

Sure... people can set tanks on fire, put things in their tracks, etc, etc, but I believe that in most cases... indeed, in almost every case, a person holding a spear in no match for a tank.

And if you want to get real crazy... how about if we were to bring a group of phalaxes magically from the past into the present, and put them up against some tanks... hummm... I'm not sure they'd be able to stand their ground, let alone fight after seeing a group of those things coming at them. Even if they got used to the idea of seeing a tank, I'm not sure they'd know how to fight one.

Regardless of real life scenarios, I think the point of all this is: If a tank could be defeated regularly by a phalax unit, what would be the point of investing all of the resources needed to build a tank. Why not just build phalax after phalax and be done with it.

In my opinion, I don't think a phalax should have a snowball's chance in hell of winning a battle against a tank, and I believe that Firaxis has made the odds quite improbable... and rightfully so!

[KrazyHorse]

Originally posted by korn469
KrazyHorse

ok basically this is a statistics problem correct?
the larger the number of rounds the smaller the cofidence interval and all of that jazz right? so if the hitpoints are high enough (like if the units had 30k, 40k, and 50k hitpoints) then when a spearman attacks a tank it would 99.9% of the time inflict 32% damage for example (the 32% damage number was a made up number)

however, after playing a thousand games the number of times a spearman loses to a tank should be about the same no matter if they had .3, .4, .5 hitpoints or 30k, 40k, 50k hitpoints right?

Nope. Your post is self-contradictory. If you have a giant # of HPs, then it's true that you almost always come up with the "expected" result. This in and of itself proves that the number of upsets (spearman wins) is much higher with proportonally lower HPs.

BTW, fractional HPs don't make any sense. Lose a round, lose a HP.

[SlowThinker]

Originally posted by Bleyn
(A + D) = X
A/X = attackers chance of winning each round of combat
D/X = defenders chance of winning each round of combat
...
Civ2-style:
6 + 3 = 9
6/9 = 66% chance for Cavalry to win each round
3/9 = 33% chance for Musket to win each round
2/3 of the rounds Cavalry scores damage
1/3 of the rounds Muskets score damage

This is wrong.
The cavalry takes a "bonus":
9 + 3 = 12
9/12 = 75% chance for Cavalry to win each round
3/12 = 25% chance for Musket to win each round

See Info: Combat (GL) thread.

---

I suggest the Boco's combat calculator to count combat probabilities:

http://apolyton.net/forums/showthre...&threadid=25493

[Marquis de Sodaq]

ST, you beat me to it!

I just browsed this forum for the first time in months - a combat discussion, woo hoo! PLEASE FOLLOW THE LINK IN SLOW THINKER'S POST if you have any questions about how CivII combat works! That explains in great detail how it all works in CivII. This thread is full of misinformation and guesses. Some of you are correct about the basic problem, tho. The formula from the manual is hopelessly inaccurate.

Hit points in Civ3 are most likely x10, just as in CivII, as reducing them to 3,4,or 5 would greatly increase the likelihood of improbable outcomes.

Attack and defense strengths were all multiplied by 8 in CivII. Greater numbers mean fewer improbable outcomes.

It is purely a statistics problem, namely the likelihood of unit A winning X rounds against unit B. Attack and defense strength figure most heavily into the equation, followed by hit points, and firepower was the least influential element.

[korn469]

grumbold and krazyhorse

ok i see the flaw in my thinking i was assuming that unit losses would be the same when in fact in the long run only hitpoint losses would be the same

if you had 1 1000 battle with 6/4 odds of victory and 1000 1 hitpoint battles with 6/4 odds, then the number of hitpoints lost will be the same correct? except that in the first case, the stronger unit will almost always win taking 40% damage, and in the second case the stronger unit will win 60% of the time, so it should lose about 400 units correct?

ok now is there any mathematical system that firaxis could adopt that would mean in a 6:4 battle that the first unit would win 60% of the time, yet that the bombard system wouldn't have the flaws i pointed out earlier nor would the healing system have the flaws Haphazard pointed out?

Bleyn

i liked your system, it seems like a better way of handling combat than in civ2

also realize that it is possible for two elite four unit armies to face off against each other, so the combat rounds in civ3 have to be longer, at least when armies are involved

[Dinner]

Re: WTF: no firepower?

Originally posted by izmircali

. Perhaps they combined firepower (since it appeared to be a multiplier to attack or defense) into the two stats but I have my doubts on this new combat system.
Anyone else know better? did I read this preview wrong?

I also have my doubts about the new combat system. I fear that in their zeal to deinfasize combat Fraxis may have totally neutered the combat system. I was hoping that Sid would use the "Ranged vs frontline" way of classifying units from CTP2. Activision did a mediocre job executing it but the idea was great.

[Grumbold]

Korn, I don't see the bombard problem in quite the same way as you do. You seem to be saying it is unfair that after a significant number of bombards both regular and elite units will be only one hit point from death and that this disadvantages the elite unit because it is a lower percentage of its total hitpoints than the regular unit. I look at it that after any number of bombards the chance of the elite unit being better off is significant but diminishing as more strikes are made. The effort applied to kill an elite unit is still considerably higher than would be needed for a regular. If you can waltz around with huge stacks of artillery and not care how many hits it takes then you've already won the game so the exact combat mechanics are irrelevant. Hypothetically giving all elites +1 defence rating might do what you ask but this is more significant in the early game when attack numbers are much lower.


Jason I was not talking 1v1, but typically an infantry formation has 10-20 times the number of troops as an armoured formation has tanks. The successes the Afghanistan resistance had against the Russians were not usually down to antitank weapons so they could equally be equiped with spears as rifles. It was down to ambushes, night infiltrations and attacks on the supply columns. If that is not good enough there is an example just before WWII of African tribesmen literally sprinting up to the sides of tanks and throwing railroad ties into the tracks. Naturally they suffered heavy casualties but they won the battle. The tribesmen were effectively bronze age infantry but having grown up in the modern world they learned how to combat modern weapons.

The problem I have with Civ, I suppose, is that the terrain factors are universally applied and no account is taken of tactical manoeuvre. Armour should rule the open tiles but be disadvantaged in cities and rugged terrain where infantry can get within assault range far more easily. There is a very good reason why modern armies still have far more infantry than tank formations. I'd quarter tank combat strengths in jungle and mountains (enterable only by road), halve them in forest and cities. If the enemy are dug in defending a city you shell them with artillery and then send in your infantry. You don't order an armoured assault.

[HalfLotus]

[SIZE=1]
The problem I have with Civ, I suppose, is that the terrain factors are universally applied and no account is taken of tactical manoeuvre. Armour should rule the open tiles but be disadvantaged in cities and rugged terrain where infantry can get within assault range far more easily. There is a very good reason why modern armies still have far more infantry than tank formations. I'd quarter tank combat strengths in jungle and mountains (enterable only by road), halve them in forest and cities. If the enemy are dug in defending a city you shell them with artillery and then send in your infantry. You don't order an armoured assault.

I like your reasoning, but having unit specific terrain advantages is just too much. Trying to remember ADM+Hitpoints for all units is plenty. Add in different values for each terrain and unit and thats a lot of numbers. Remember, Firaxis shoots for fun and accessibility in games.

[korn469]

Korn, I don't see the bombard problem in quite the same way as you do. You seem to be saying it is unfair that after a significant number of bombards both regular and elite units will be only one hit point from death and that this disadvantages the elite unit because it is a lower percentage of its total hitpoints than the regular unit. I look at it that after any number of bombards the chance of the elite unit being better off is significant but diminishing as more strikes are made. The effort applied to kill an elite unit is still considerably higher than would be needed for a regular. If you can waltz around with huge stacks of artillery and not care how many hits it takes then you've already won the game so the exact combat mechanics are irrelevant. Hypothetically giving all elites +1 defence rating might do what you ask but this is more significant in the early game when attack numbers are much lower

grumbold my problem is solely with the limited number of damage states...now if a regular unit can take 5% or 10% damage from a bombard then i cease having a problem, but if a single successful bombard will always inflict 1/3 damage on a regular unit then i find that a shame, especially since with two bombards the regular unit would be as weak as it is going to get, that seems to trivialize the entire system to me...i am worried that without subhitpoints that combat mechanics in civ3 could cause many problems

[Grumbold]

Terrain differences for certain types of unit were not a problem even in gamer friendly games like the Panzer/Allied General series. Having very simple one-size-fits-all terrain effects plus very limited hitpoints per unit are in keeping with the Firaxis aspiration to move back toward the Civ1 simplicity. Lets hope they've guaged it right and all our different reservations are groundless.