Attack and Defense VS 1 Number

[deanej]

Civ4 has as many rounds of combat as needed for a unit to die or withdraw. The strength of each unit and its bonuses create a percentage chance to win (which isn't just there to tell you something, that's the actual probability the game uses when the random number generator determins if a unit scores a hit on another and how much damage is done). First strikes are rounds of combat where one side can do damage to another without fear of damage itself.

The percentage modifiers act strangely. Everything but the modifier from the combat promotions applies to the defender, and the percents adds. This means that a city garrison promo and a city raider promo cancel each other out perfectly. I'm not sure exactly how stuff like -100% vs. armor is done, but I know it does not result in a unit that has no strength against armor units. -1000% is good enough to ensure a reasonable amount of losing, however it is not guaranteed.

 

[KrikkitTwo]

But for a real consideration warriors or basically ANY unit not armed with anti-tank weapons (infantry can be presumed to have some anti tank weapons, they are a modern unit) is doomed to lose completely, tanks are immune to small arms fire,

Tanks are not immune to IEDs, small arms v. commanders sticking their heads out, tiger traps, or someone slitting the throats of the crew when they are in their sleep.

Remember this is not a 'cage match' it is an abstraction of all the combat that occurs over several months or several decades (depending on the time period) over an area larger than many European nations, ie Belgium, Denmark.

So with modifiers the problem of spearmen vs tanks can be solved, spearmen have -100% str against tanks, they have 0 chance of winning.

Modifiers should probably not be that specific... for Spear v. Tank, a
1.Significant Strength difference (Spear should be much weaker than a tank)
and
2. Range difference (tank should have greater range than a spear)
should be enough

But anyway, someone please reiterate for fun, what is the Civ 4 combat system exactly?

Each unit has a Strength and hitpoints
a 200 v. a 100 gives the same odds as a 2 v. a 1 the ratio is all that matters

Each round of combat one and only one unit takes damage*

The stronger a unit is compared to its enemy, the more likely the enemy unit will take damage and it will not

The amount of damage a unit takes when it is hit, is more when the enemy unit is stronger and when it is weaker*

This is repeated until one unit is dead


*IF the unit taking damage has 'first strikes' then it takes 0 damage

How do you factor in SKILL? (how good shots the men are or how good they are with a sword)

unit with more experience get a bonus to strength

How do you factor in MORALE? (how brave the men are or how much fighting spirit they have)

Units are easier to produce or get extra experience when produced in certain social situations... giving them a bonus to strength

How do you factor in FIREPOWER?

Units with better weapons have a bonus to strength

How do you factor in MANPOWER? (some units have 10.000 men, some have 15.000 etc.)

Units with more men and the same type of weapon would have more strength
or
you would build multiple units

How do you factor in ranged fire and melee? (first you shoot at range then you have melee, first strikes and combat phases, of course there could be exceptions like a forest where ambushes are a possibility, go straight to melee, even swordsmen can beat riflemen in a jungle or forest)

Strength and bonuses to it from different Terrains and unit matchups
You also have range where one unit can attack and the other not attack back because it has better range

I think there could be COMBAT TYPES, ranged combat (archers and riflemen) melee combat (swordsmen and musketeers in bayonet attack, a swordsman will probably win a musketeer in melee, a musket with a bayonet is NOT that good a weapon in a melee and a musketeer has no armor.)

There should definitely be Unit types, but probably not Combat types... Combat situations (I out range you so I can attack and you can't)

My point is, ALL these real life factors have to be represented in the model, that´s the kind of mathematical combat system I´d like to see.

They all are... just not separately... they are all part of strength (or attack v. defense)

 

[Jopo]

Ok, here´s why I would want to have TWO values, HITPOINTS AND STRENGTH (in civ 4 we have only strength, I have looked at the unitinfos.xml file and there is no mention of hitpoints, there is only strength)

Let´s say you have a HUGE horde of barbarians a unit of 50.000 men (one unit) and a legion of 10.000 men (one unit), the legion has higher strength (more attack/defence power because of better organization) but the barbarians have more men (more hitpoints).
The barbarians are brave warriors and fight well individually but the legion has better organization and fights as a unit so it has more attack strength per man, of course attack strength and hitpoints would be related so when you have more hitpoints you also have more attack strength, so the units might be even in odds.

If civ 4 has hitpoints then WHERE IS IT? where is the info of it?

Cheers!

 

[Earthling]

So in your model, every unit is guaranteed to take damage every combat, no matter how much higher its strength is?

No, that would not be necessary, you appear to be missing how this really could be simpler and easier.

In fact, I'll say that while I'm very apprehensive of their whole combat model, something like what Krikkitone proposes is exactly the way it could and should work.

Randomness needs to be decreased and combat results made more intuitive. They could have various modifiers, even different attack and defense strength, but when it comes down to it having a single round of combat that the user can easily understand makes a lot of sense.

For example let's say two units have full hit points equaling 15 and are going to do batter. If a Sword has 7 Attack and 4 defense, a Spear has 4 Attack and 5 defense (using units in civ 4, but it doesn't really matter), all modifiers are then applied and presented.

The end result, say the Sword comes out ahead here, is the Sword will do 4-7 damage to the enemy Spear unit when attacking, and the Spear will do 2-5 damage. And that's it, and weights could be given to make the higher or lower ends of these ranges less likely in an RNG, but the player and AI know what they are getting, while still preserving some chance.

Such a system could easily scale to the point if you have a 24/16 strength tank or something, the result is the tank does 13-15 damage to the spear unit (with a strong chance of the full 15 damage to wipe them out) and the Spear does 0-1 damage. It wouldn't have to be guaranteed damage, 0 or 100% of an enemy unit's health are both possibilities in any combat.

 

[deanej]

Ok, here´s why I would want to have TWO values, HITPOINTS AND STRENGTH (in civ 4 we have only strength, I have looked at the unitinfos.xml file and there is no mention of hitpoints, there is only strength)

Let´s say you have a HUGE horde of barbarians a unit of 50.000 men (one unit) and a legion of 10.000 men (one unit), the legion has higher strength (more attack/defence power because of better organization) but the barbarians have more men (more hitpoints).
The barbarians are brave warriors and fight well individually but the legion has better organization and fights as a unit so it has more attack strength per man, of course attack strength and hitpoints would be related so when you have more hitpoints you also have more attack strength, so the units might be even in odds.

If civ 4 has hitpoints then WHERE IS IT? where is the info of it?

Cheers!

Which is suprising, because I could have sworn I saw it. I dug around and found it in GlobalDefines:

	<Define>
		<DefineName>MAX_HIT_POINTS</DefineName>
		<iDefineIntVal>100</iDefineIntVal>
	</Define>

So you could still give different units different hitpoints, it just requires the SDK in addition to XML.

The best way to represent more men would be to have more units. Too bad civ5's combat system will make this more difficult.

 

[KrikkitTwo]

Let´s say you have a HUGE horde of barbarians a unit of 50.000 men (one unit) and a legion of 10.000 men (one unit), the legion has higher strength (more attack/defence power because of better organization) but the barbarians have more men (more hitpoints).
The barbarians are brave warriors and fight well individually but the legion has better organization and fights as a unit so it has more attack strength per man, of course attack strength and hitpoints would be related so when you have more hitpoints you also have more attack strength, so the units might be even in odds.

In that case they would just be the same
Horde of 50,000 Barbarians= Strength 6, 100 hit points if undamaged by combat
Legion of 10,000 Trained Soldiers= Strength 6, 100 hit points if undamaged by combat


No reason for any distinction. besides graphics... they would probably have the same cost (there are more barbs but they are more poorly equipped and don't have to spend the same time training.)

the actual # of men involved is irrelevant.. you could also say
5 Mutant Super soldiers=Strength 6, 100 hit points if healthy
10,000,000 Mutant Army ants=Strength 6, 100 hit points if healthy

All that would change would be the graphics.

Now certain Types of units would have different abilities ( ie Archers, would have Range, Horsemen would have greater mobility, Some units might have Flanking bonuses, or bonuses v. other Types of units, or for fighting in certain types of terrain)

 

[PieceOfMind]

Civ4 has as many rounds of combat as needed for a unit to die or withdraw. The strength of each unit and its bonuses create a percentage chance to win (which isn't just there to tell you something, that's the actual probability the game uses when the random number generator determins if a unit scores a hit on another and how much damage is done). First strikes are rounds of combat where one side can do damage to another without fear of damage itself.

The percentage modifiers act strangely. Everything but the modifier from the combat promotions applies to the defender, and the percents adds. This means that a city garrison promo and a city raider promo cancel each other out perfectly. I'm not sure exactly how stuff like -100% vs. armor is done, but I know it does not result in a unit that has no strength against armor units. -1000% is good enough to ensure a reasonable amount of losing, however it is not guaranteed.

Just FYI, when a defense modifier is overall negative, all you do is.... wait, it's probably best if I just do a really obvious example.

Spearman with 25% defense bonus is strength 4 * (1 + 0.25) = 5.
Spearman with -25% defense (i.e. a penalty) is strength 4 / (1 + 0.25) = 3.2.
Another way you could put it is the unmodified strength is 25% more than the penalised strength.

Also, a strength ratio of 1000 to 1 is enough to guarantee an attacker victory because the combat die used has 1000 sides and it "rounds" in favour of the attacker. It is of course extremely rare in civ4 to have such a situation occur!

Back to the discussion,
I'm sort of opposed to what Sid did with civrev. He takes advantage of the psychological problems people have with random numbers to make people feel better in their games. In order to do that, they needed to avoid having any sort of odds display. Note that I always talk about odds in a way where I assume you can work out the exact probability from them. What civrev does, and what very early vanilla civ4 did, was only show something like a strength ratio - not the calculated odds. After a while, people just get used to that system and will be similarly annoyed when they lose high odds battles. However, I suspect what has been done is that above some threshold (I saw 85% mentioned earlier), you simply cannot have probabilities of battles in there. In a game like civ5 though, where hopefully it is not as dumbed down as civrev, having such an arbitrary discontinuity point will simply lead to people exploiting it to the fullest. I would much rather help people to understand the nuances of elementary probability theory rather than just protect them from a nasty RNG result that lets them lose a 99% battle. Games don't have to be constantly pleasing the player, something I think Sid focuses too much on. Getting angry about a bad RNG result is just part of the experience

I favour making anything UI-related more transparent (not literally, lol).

 

[Ahriman]

Too bad civ5's combat system will make this more difficult.

A major part of Civ5's 1upt system is to limit how much you can concentrate power in any one tile. This is a deliberate design decision. So I doubt they're going to let you have one huge unit with more "men" to let you create implicit stacks.

Just FYI, when a defense modifier is overall negative, all you do is....

Well, there can be a difference, right?

Imagine if the enemy is strength 5 with +50% modifier (eg from promotions combat 1-5).
And I am strength 4.

If I get a +25% modifier to me, then the strength ration is 7.5:5 = 1.5:1
If my enemy gets a -25% modifier to them, then the strength ratio is 6.25:4 > 1.5:1.
[Because the enemy is now 5*(1.5-0.25) = 6.25]

So there are some differences, because of how modifiers affect the base strength, not the current strength.

Similarly, a unit that is strength 4 and +25% vs unit type A is not the same as a unit that is strength 5 with a -20% penalty vs everything that is not type A.

Because if the first unit gets a +10% promotion, this increases its strength by 0.4, whereas if the second unit gets a promotion this increases its strength by 0.5.

I agree that this is a very unfortunate way of how Civ4 does modifiers.

I would much rather help people to understand the nuances of elementary probability theory rather than just protect them from a nasty RNG result that lets them lose a 99% battle.

Hear, hear!

I favour making anything UI-related more transparent

In general I agree, but I don't think this necessarily means that combat result probabilities have to be something that you could easily calculate in your head (eg by only have 3-4 different outcomes from any fight as in Krikkitone's model).

I think its fine to still have a slightly more complex outcome of possibilities as long as these are transparent to the player (eg: Battle for Wesnoth)

 

[EditorRex]

This is a good question.

I'm generally fairly agnostic on this point, I can see the arguments for the one-value system, but I also really liked when FFH re-introduced separate attack/defense values.

Still I would pose the question; how would one feel about the strength 18 machinegunner who can't attack vs a machinegunner that is 6/18?
Is it really fair that an entire full strength division can't attack even an almost dead unit? Its not like every last guy in the squad has only a machinegun.

Purely defensive units no longer really have much of a place in a 1upt system.

I think "Machine Gun" has you confused. In theory, probably most of your modern troops from Marines to Paratroopers to Infantry to Mech Inf are carrying some type of automatic rifle. But "Machine Gun" hear refers to the big machine gun nests that we saw in WW2. These are not highly mobile and are thus nearly useless for attacking. Obviously, you could dismantle and relocate one, but probably not under immediate battlefield conditions. That's why Civ4 treats these as Siege Units that can only defend. Personally, I thought this was a very logical decision and would like to see more units with this type of feature -- perhaps a gun turrett artillery unit for use in fortified sites.

 

[Ahriman]

I'm not confused.

Some machineguns are definitely able to be relocated under battlefield conditions. That happened all the time. Watch any WW2 movie. A gunner and loader grab the ammo and gun, charge forward to the next cover, and set up the weapon, while under fire.
"Submachine guns are designed to be portable automatic weapons for personal defense or short range combat, and are intended to be fired while being hand held. Submachine guns use small pistol caliber rounds. A proper machine gun is often portable to a certain degree, but is generally used when mounted on a stand or fired from the ground on a bipod"
http://en.wikipedia.org/wiki/Machine_gun
http://en.wikipedia.org/wiki/M1918_Browning_Automatic_Rifle

Also, these represent units of soldiers with a mix of weapons, not just a single weapon.

Its ridiculous that a whole brigade of WW2 soldiers primarily armed with machineguns couldn't finish off a heavily wounded group of ancient swordsmen.

If nothing else, the guys have side-arms.

I'd be very surprised to see defensive only-weapons in Civ5, because these have much less strategic value in a 1upt system, since there is no longer any such thing as a stack guard or city defender.
We might have some units that were better on defense than offense though.

I think my ideal implementation for a machinegun unit was one that gained a double fortification bonus. This would allow defensive weapons that could still attack, and still had only a single strength value.

 

[deanej]

I'd be very surprised to see defensive only-weapons in Civ5, because these have much less strategic value in a 1upt system, since there is no longer any such thing as a stack guard or city defender.

This is compensated for by the change to hexes, allowing for an easier defending of the border.

 

[EditorRex]

I'm not confused.

Some machineguns are definitely able to be relocated under battlefield conditions. That happened all the time. Watch any WW2 movie. A gunner and loader grab the ammo and gun, charge forward to the next cover, and set up the weapon, while under fire.
"Submachine guns are designed to be portable automatic weapons for personal defense or short range combat, and are intended to be fired while being hand held. Submachine guns use small pistol caliber rounds. A proper machine gun is often portable to a certain degree, but is generally used when mounted on a stand or fired from the ground on a bipod"
http://en.wikipedia.org/wiki/Machine_gun
http://en.wikipedia.org/wiki/M1918_Browning_Automatic_Rifle

Also, these represent units of soldiers with a mix of weapons, not just a single weapon.

Its ridiculous that a whole brigade of WW2 soldiers primarily armed with machineguns couldn't finish off a heavily wounded group of ancient swordsmen.

If nothing else, the guys have side-arms.

I'd be very surprised to see defensive only-weapons in Civ5, because these have much less strategic value in a 1upt system, since there is no longer any such thing as a stack guard or city defender.
We might have some units that were better on defense than offense though.

I think my ideal implementation for a machinegun unit was one that gained a double fortification bonus. This would allow defensive weapons that could still attack, and still had only a single strength value.

You are making the assumption that's what this represents. Based on what evidence? On the other hand, the machine gun unit operating as it does in Civ IV makes sense if it ONLY represents the an immobile nest of machine guns. Think about it. Why would they have made it a defense only unit if this isn't what they had in mind?

 

[Ahriman]

This is compensated for by the change to hexes, allowing for an easier defending of the border.

I don't understand what you're saying here. How do hexes make border defense easier than squares?
Yes, defensive units will be useful on choke points - but they'd be totally useless everywhere else.

Based on what evidence?

Based on the fact that units in Civ have always been representative of large groups. Its not just three guys. Its hundreds or thousands.

There has never been a military unit in human history armed only with machineguns. They're deployed in support with other units.

On the other hand, the machine gun unit operating as it does in Civ IV makes sense if it ONLY represents the an immobile nest of machine guns.

It made sense in Civ4 as a gameplay abstraction, because it was still useful to have a unit that was a stack-guard or city defender.
Those combat roles no longer exist in Civ5.

 

[deanej]

How do hexes make border defense easier than squares?

You can't slip through the diagonal any more.

They're deployed in support with other units.

This is represented by stacks. And I think the machine guns in civ4 represent the purely defensive ones in WWI that led to trench warfare.

 

[EditorRex]

I don't understand what you're saying here. How do hexes make border defense easier than squares?
Yes, defensive units will be useful on choke points - but they'd be totally useless everywhere else.


Based on the fact that units in Civ have always been representative of large groups. Its not just three guys. Its hundreds or thousands.

There has never been a military unit in human history armed only with machineguns. They're deployed in support with other units.


It made sense in Civ4 as a gameplay abstraction, because it was still useful to have a unit that was a stack-guard or city defender.
Those combat roles no longer exist in Civ5.

I know I'm arguing a realism issue that doesn't directly go to gameplay, but I disagree with the idea that Civ units represent mixed unit types and that it would unrealistic otherwise. Plenty of militaries have had single-type units within larger mixed-unit armies. You might have a unit of 40 spearmen -- and only spearmen, fighting alongside a unit of 30 archers, and only archers although they also probably have some decent melee equipment on each archer. In modern warfare, and really only from the mid-20th century, there are more mixed tactics, but a unit still has a primary purpose and makeup. There might be machine-gunners and snipers mixed in with regular infantry. But that would be different from the special machine-gun units (that probably also had some infantry and grenadier support personnel) we saw in defensive positions during phases of World War II. These machine gun units had heavier fire power than the mobile version and were dedicated defensive position units, usually in highly fortified positions, such as Normandy or Iwo Jima. These units could advance and retreat with the front to some extent, but what they didn't do was participate in a direct offensive assault on the enemy, except as covering fire (and we have nothing to duplicate the idea of covering fire in Civ games).

 

[KrikkitTwo]

Been debating a lot without putting oout what goals there would be for the combat system. Let me put out mine (based on the 1upt, small# of units assumption), and see if there are better ways to do it

1. Better/Better positioned/used units should be more capable of doing damage to the enemy on average

2. Battles should feel 'realistic'...and so should not be totally deterministic

3. Players should have a high degree of control over battles and battle results (Battles should not be totally random)


#2 v. #3 is where I think the argument of balance is.

To achieve #3 you need
1. Predictable battle results
2. "Caps" on the best and worst random outcome from a battle (in Civ 1-4 in ANY battle it is theoretically possible for Either side to lose and the other to take no damage at all... this is VERY BAD for player control)


#2 is the Spearman v. Tank problem

Part 1: Nearly dead Spearman should never beat Healthy Tank... but that is ALWAYS possible in a totally random system, unless specific rules are put in

Part 2: A stronger unit should NOT be invulnerable to weaker units.. especially multiple weaker units


If the basic combat mechanics are designed right, you should never worry about these problems

Part 2 can be handled with damage (a good introduction in Civ 2.. it needs to be strengthened by making healing more costly)

Part 1 can ONLY be handled by eliminating the "chance to take 0 damage"






The thing to remember is that this is a game not a simulation... the economy+social aspects of civ are Drastically oversimplified to allow for player control and input. The Military model needs that to happen as well. There is a place for the RNG, but is should be limited... it can only go so far... only 2x the damage or 20x

I'd far rather a 50% of doing 2x damage than a 1% chance of doing 50x the damage because that 50x the damage is way to much for a random chance.... but that is what happens in Civ 1-4.

If you want to do 50x the damage you should have
1. gotten a better unit (tech/economy)
2. positioned the unit better (tactics)
Luck should NOT have that big effect in a combat Game

In a game like civ5 though, where hopefully it is not as dumbed down as civrev, having such an arbitrary discontinuity point will simply lead to people exploiting it to the fullest.

True which is why the system needs to be revised even MORE fundamentally.

The Problem is the Random probability of an Unlimited Massive failure.

In ANY battle it is possible to lose your unit (in civ 1-4)...ANY battle... the chance may by 1 in a trillion but it is there.

This is why my system would alway only have two possible results.... each with equal odds (in terms of the effect on a single unit)... people can understand 50% odds fairly well.

The "Maximum Bad result" would be limited NOT be loss of a healthy unit (although you might lose a damaged unit)

I would much rather help people to understand the nuances of elementary probability theory rather than just protect them from a nasty RNG result that lets them lose a 99% battle. Games don't have to be constantly pleasing the player, something I think Sid focuses too much on. Getting angry about a bad RNG result is just part of the experience

No it is an example of bad game design.
Games do have to avoid Unnecessarily displeasing the player.

The power that the RNG has to control the game in combat is too much, I wish people undertood elementary probability theory too, but You shouldn't design a game that relies on them doing so.

Imagine if your production each turn was odds, each turn there was an X% chance that the current project would be finished. Would this be a useful addition to city management...No, even though it would probably be more realistic than the current version.

Instead you can predict and control what you are producing

You should be able to do the same thing with combat... predict and control what you are killing/losing (the only significant 'uncontrolled' factor should be what the enemy units do)

 

[Ahriman]

This is represented by stacks. And I think the machine guns in civ4 represent the purely defensive ones in WWI that led to trench warfare.

This is a reasonable interpretation in a Civ4 setting.

But, no stacks in Civ5. An individual level is at the Regiment, Corps or Division scale (and there are no machine-gun regiments!), and a purely defensive unit is low value.

 

[PieceOfMind]

Krikkitone I fully understand the desire to remove the dependency of game mechanics on the roll of an RNG. I have said in the past (not in this thread though) that in general when simulating something, removing a call of an RNG is progress - a step in the right direction. Civ of course is not quite a simulation, but people like to think of it that way.

RNG element has been in civ games from the very start. Honestly I think the occurence of bad luck battles is by now a part of its character. So yeah, while RNG rolls determining production is absurd, RNG rolls in combat is not absurd.

You find high-level players absolutely abhor things like random events because they dislike having random elements influence their success in games. This is like in RTS games too where RNGs are rarely used.

It comes down to a matter of taste or opinion. I for one like the fact that combat in civ games has a random element to it. Every battle carries a risk, even if only a small one. If I could predict with 100% certainty what the battle outcome would be based on the two units involved, I wouldn't need to take risks. This could make the game more boring, IMO.


@Ahriman, concerning your reply to me...

Not sure what the problem is. Do note I talked about "overall defense modifier". Except for combat promotions, all modifiers are always added or subtracted from the defender's overall modifier. Once that modifier is determined, that's when you start adjusting the defender's strength according to the rule I mentioned. The attacker can only be adjusted by his combat promotions.

 

[Ahriman]

Not sure what the problem is.

My only point is; a % strength bonus to the attacker (eg from combat promotions) is not the same as an equal size strength penalty to the defender.

Not all % modifications are created equal.

 

[PieceOfMind]

Well, if you ignore attack combat promos, they are indeed all created equally. This is something I was slow to learn.

It is exactly for this reason that combat promos are good for attacker units but bad for defender type units. It's unfortunate that there is this oddity, but it's restricted to attacker combat promos.