Balance Discussion - Hack and Pierce

[DesertRose]

Been reading through old balance discussion threads (1, 2, 3, 4) but because those threads are up to 8 years old I'll have the audacity to make a new one instead of responding to one of those.

Spoiler

 

On 17/04/2022 at 7:50 PM, real_tabasco_sauce said:

Spears are fine to have both hack and pierce damage, as this speaks to their versatility. However, some more asymmetry between these values might be worth considering.

 

Spearman dealing about 50/50 Hack/Pierce makes them boring and very hard to balance as the individual Hack and Pierce Resistance of the target becomes irrelevant. It doesn't matter if the target has 5 Hack Res and 1 Pierce Res, 4 and 2, 3 and 3, 2 and 4, or 1 and 5, Spearman will deal roughly the same damage against all of them. That makes them hell to balance, as when they deal sufficient damage to one target to be considered good against them they very quickly become good against everything, making them OP, and vice versa.

This thread will be about (Citizen Soldier) units that deal Hack and / or Pierce damage, and how they can be given distinguished roles with just Hack Damage, Pierce Damage, Hack Resistance, Pierce Resistance and a Damage Multiplier against Cavalry.

 

The values presented are not set in stone but to illustrate the core principles.

 

So lets start with the core triangle:

Infantry Spearman:

• 75% Hack, 25% Pierce
• Bonus Damage against Cavalry

Infantry Archer:

• 100% Pierce
• Long Range

Cavalry Swordsman:

• 100% Hack
• High DPS (due to high RoF)
• -1 Hack Resistance (from Cavalry Melee)
• +1 Pierce Resistance (from Cavalry Swordsman)
• Faster than Cavalry Spearman

These three units form the most distinct Rock Paper Scissors triangle. Cavalry Swordsmen's high speed, high damage and high Pierce Resistance makes them decimate Archers. Infantry Spearmen decimate Cavalry Swordmen thanks to mostly their bonus damage, and Infantry Archers weed out most Spearmen before they can reach them.

The next three units still uphold the triangle but in a (sometimes far) less pronounced way:

Infantry Swordsman:

• 100% Hack
• High DPS (due to high RoF)
• +1 Hack Resistance (from Infantry Swordsman)
• Faster than Infantry_Spearman

Infantry Javelineer:

• 25% Hack, 75% Pierce
• Low Range
• High DPS (due to high damage)
• Faster than Infantry Archer

Cavalry Spearman:

• 75% Hack, 25% Pierce
• Small Bonus Damage against Cavalry (1.2-1.5)
• -1 Hack Resistance (from Cavalry Melee)

Compared to Cavalry Swordsmen: Cavalry Spearmen still beat ranged infantry, but lower DPS, speed and Pierce Resistance makes them less effective. They are slightly better against cavalry despite lower DPS thanks to their Bonus Damage. They perform about equally good against Infantry Swordmen thanks to part of their damage bypassing the +1 Hack Resistance of them.

Compared to Infantry Archers: Infantry Javelineers are better against Cavalry due to them partially dealing Hack damage. Their lower range does not make that much of a difference against cavalry as cavalry can close the distance quickly. They are weaker against infantry due to having to be far closer, and having partially Hack damage does not offer them any advantage; in fact, against Infantry Swordsmen it becomes a disadvantage.

Compared to Infantry Spearmen: Infantry Swordmen lose to melee Cavalry due to their better stats, but cavalry's weakness to Hack damage and their own Hack Resistance means they barely win cost-for-cost. They win against Spearmen due to their own higher DPS and Hack Resistance. They are better against ranged units due their higher DPS and speed.

Infantry Pikeman:

• 25% Hack, 75% Pierce
• Long Range, High Damage, but low RoF (DPS slightly below Infantry Spearman)
• Bonus Damage against Cavalry
• Slower than Infantry Spearman

Compared to Infantry Spearmen: 1-vs-1 Infantry Pikemen are the worst melee infantry unit. They become much stronger in numbers as their long range and alpha strike damage allows them to kill some of the approaching melee units before they can fight back, snowballing the fight in their favor.

Not sure about Cavalry Archers and Cavalry Javelineers, so I'll put them in a spoiler:

Spoiler

Cavalry Archer:

• 100% Pierce
• Long Range, less than Infantry Archer
• -1 Pierce Resistance (from Cavalry Ranged)
• About the same speed as Cavalry Spearman

Cavalry Javelineer:

• 25% Hack, 75% Pierce
• Low Range, less than Infantry Javelineer
• -1 Pierce Resistance (from Cavalry Ranged)
• Faster than Cavalry Archer

They are even more effective against melee infantry does to being able to kite them endlessly. However melee cavalry and ranged infantry can zone them out.

edit:

Here is the mod to try out these ideas:

 

Edited March 14 by DesertRose

[Deicide4u]

57 minutes ago, DesertRose said:

Infantry Javelineer:

• 25% Hack, 75% Pierce

Why, though?

I think giving melee units pierce damage makes no sense. It's also much easier to balance when you know that melee = hack and ranged = pierce. 

But, people chose to give "trash" (w/o metal cost) melee units high pierce damage, to artificially nerf them against buildings, thus promoting the capture mechanic.

[Emacz]

Classical Warfare AEA give Javs a very small hack dmg

[DesertRose]

6 hours ago, Deicide4u said:

Why, though?

In my suggestion melee cavalry has less Hack Resistance than Pierce Resistance.

Thus Javelineers dealing 25% Hack is a "hidden" 2.8% / 5.1% bonus damage against cavalry.

[wowgetoffyourcellphone]

53 minutes ago, DesertRose said:

In my suggestion melee cavalry has less Hack Resistance than Pierce Resistance.

Thus Javelineers dealing 25% Hack is a "hidden" 2.8% / 5.1% bonus damage against cavalry.

Add an attack bonus instead (say... "Javelineers 1.05x attack vs. Cavalry, Peltasts 1.15x attack vs. Cavalry") This would prevent any unintended knock on effects of giving your units multiple attack types. DE makes extensive use of attack bonuses for this purpose:

 

(Tooltip)

Spearman

Strong Against: Cavalry (2.0x)

Weak Against: Ranged Units, Swordsmen

 

(Tooltip)

Javelineer

Strong Against: Ranged Cavalry (1.10x), Elephants (1.5x)

Weak Against: Melee Cavalry

 

(Tooltip)

Civilian Citizen

Strong Against: Resources, Chickens (100x)

Weak Against: Everything Else

 

Just kidding about the last one.

 

[Perzival12]

On 8/3/2026 at 6:19 AM, Deicide4u said:

I think giving melee units pierce damage makes no sense. .

Agreed

How it should be:

Units with swords, axes, or other edged weapons (arrow edges don’t count) deal Hack damage.

Units with spears, arrows, pikes, and other tipped weapons (maybe some swords can count, but probably not given no thrusting animations) deal Pierce damage.

Units with maces, staffs, or blunt weapons of any type deal Crush damage.

Resistance balance should be redesigned, with Pierce and Crush resistance being about equal, while Hack remains slightly lower than them.

[Stan`]

15 hours ago, Perzival12 said:

Units with spears, arrows, pikes, and other tipped weapons (maybe some swords can count, but probably not given no thrusting animations) deal Pierce damage.

Isn't that directly contradicting the message you quoted?

[Perzival12]

2 hours ago, Stan` said:

Isn't that directly contradicting the message you quoted?

I guess your right, I read that message wrong.

[wowgetoffyourcellphone]

I'd rather Hack = Melee, Pierce = Ranged. Crush = Siege and/or Anti-building. KISS principle applies and prevents unintended consequences. 

[Adeimantos]

Like desert rose and perzival I prefer using hack, pierce, and crush damages to make some units have natural advantages against others and keep bonuses limited.

The way I see it, from a unit's point of view pierce is an attack you need a shield to block, hack is an attack you could block with a sword, and crush is something you can only avoid by dodging.

[Deicide4u]

11 minutes ago, Adeimantos said:

The way I see it, from a unit's point of view pierce is an attack you need a shield to block, hack is an attack you could block with a sword, and crush is something you can only avoid by dodging.

You can block mace and axe attacks using a shield, but you can't really block the axe attack using your sword. By this logic, mace is a pierce weapon and axe is both a pierce and crush weapon.

See how complicated this can get?

Compare that to the following hypothetical situation:

All melee units have a standard "Melee" attack type. All ranged units have standard "Ranged/Pierce" attack type. All siege weapons have standard "Siege/Crush" attack type. Units can have more than one attack type, which allows elephants and other specialized units to deal some Siege damage. Unit resistances are now scaled based on how they fare against certain attack types. So, a Hoplite can have high Pierce resistance and average Melee resistance. Swordsman will have higher Melee resistance, but he is weaker against Ranged attacks. Unless he is a Roman with a very large Scutum (thus, Romans naturally have better swordsmen).

Immortals, being without big shields, can have high Melee resistance, but lower Ranged resistance, etc.

[Adeimantos]

2 minutes ago, Deicide4u said:

You can block mace and axe attacks using a shield, but you can't really block the axe attack using your sword. By this logic, mace is a pierce weapon and axe is both a pierce and crush weapon.

See how complicated this can get?

Compare that to the following hypothetical situation:

All melee units have a standard "Melee" attack type. All ranged units have standard "Ranged/Pierce" attack type. All siege weapons have standard "Siege/Crush" attack type. Units can have more than one attack type, which allows elephants and other specialized units to deal some Siege damage. Unit resistances are now scaled based on how they fare against certain attack types. So, a Hoplite can have high Pierce resistance and average Melee resistance. Swordsman will have higher Melee resistance, but he is weaker against Ranged attacks. Unless he is a Roman with a very large Scutum (thus, Romans naturally have better swordsmen).

Immortals, being without big shields, can have high Melee resistance, but lower Ranged resistance, etc.

If you block a mace with your shield it'd still hurt, that was what I meant about crush damage. Anyway how you describe it is pretty close to the current game and I'm fine with that. I'd like more variety between units, what you said about swordman and spearman fits that. We should give the swordman another hack armor and the spearman another Pierce armor.

[ThePoshBarbarian]

Going to put my own implementation opinion out there, always kinda found it weird how 0ad does it but I've gotten used to it somewhat. Though if we were to take a page out of say a simplified "realism" argument, my implementation for damage types would be:

• hack = blade edge slashing/hacking (if you'll forgive the lack of better words to describe it) aka mostly good against non-armoured foes. (there's just one unit with falxes iirc, for macedonians, they can be an exception by just giving them a bonus against armour)
• pierce = spears/javelins/bolts/arrows piercing targets, doesn't struggle against armour, just a good all around option
• crush = a solid mass hitting another, aka big rock from onager/siege ballista crushing someone, or a getting a mace/club hit to the face. Great at dealing with really heavily armoured targets. Slings could be included in this but that's a can of worms.

This is just a little thought experiment I've had, honestly I prefer 0ad's current implementation of the damage mechanics. Thinking about a bit more deeply, the system I've written up has some problems when it come to classifying certain units, eg falxes for when thracian/dacians/getae get implemented, or how slingers work (sling bullets are technically speaking a crushing type of damage, but they are also known to pierce through targets given enough force).

To balance armour around this system would be difficult me thinks, I personally would use an armour types system with:

• non armoured = basically basic clothes, no armour, little to no resistance to any damage types
• organic = aka leather/linen/whatever material that is not made from bronze/iron, good resistance to slashing, decent to slightly below average resistance to crushing, little to none against piercing
• metalic = iron or bronze, great resistance to slashing, average to poor restistance against piercing, low resistance to crush.

Already I'm having problems describing this armour system because of what little I know about armour materials and the physics behind how they mitiagate certain types of damage makes this way too unimmersive/game-ified to me. Again, this was just a thought theory of an implementation, I still think that 0ad's current damage mechanics work very well for it (and imho, a way better representation than AoE2DE does).

[Deicide4u]

@ThePoshBarbarian

I really like your classification of damage and types, nothing wrong with it. But, it fits more in a tactical RPG squad-based strategy game than classic RTS.

Edited March 16 by Deicide4u

[Perzival12]

@ThePoshBarbarian, your classification makes sense, except for resistance needs to be reversed. In all actuality, leather armor is designed to block arrows and other pointed weapons, an axe blade or a good blow with a sword would cut through it like butter.

As for sling implementation, slingerss could just deal small amounts of pierce and crush damage.

Edited March 16 by Perzival12

[LienRag]

Not sure about armour (arrows can be designed specifically to work against chainmail, but afaik not against plate) but shields are good against pierce - which is another way to make units differents.

[Thalatta]

I have no idea how formulas work in this game (would like to know though), but along the “simplified realism argument”, I think bonuses should be avoided, they seem to me a sign that the parameters used to reproduce reality are either not enough or are being inefficiently used. I rather have a couple more parameters than who knows how many random bonuses here and there. This keeping in mind that it would be nice to eventually show the relevant info not as a bunch of text like it’s now, but as a given set of icons and values that is as simple and standard as possible.

It seems to me that with hack, pierce, crush, block (with a shield), parry (with a weapon) and dodge values, both for attack and defense, most things should be achieved (I don’t see any advantage to organic/metallic armor types, it’s just setting hack/pierce/crush defence values). Also, causing low damage should do very little, as a measure of actual penetration of defenses, to enhance differences, and penalising massing up the wrong kind of counter. Going through some of what has been said, and just as a first approximation:

Spearmen: high pierce attack, hard to block, low dodging.

Archers: pierce attack, easy to block, basically impossible to parry, hard to dodge.

Cavalry: low blocking, high dodging.

Maces: high crush attack, hard to block.

Axemen: high hack attack, hard to parry.

Swordsmen: high hack attack, hard to dodge, high parrying, high blocking if they have a shield.

Javelineers: high pierce attack, low crush defense, high dodging.

Elephants: very high crush attack, high hack and crush defense, can’t be dodged, can’t dodge.

Rams: extremely high crush attack, very high pierce defense, high crush defense, extremely easily dodged, can’t dodge.

Buildings: extremely high pierce and hack defense, very high crush defense, can’t dodge.

Slingers: high crush attack, low defenses.

Slingers should be good against troops that don’t have crush defense, cheap, sling faster and farther than bows (bow damage should decrease more with distance), but have no armor, be unable to form close formations when slinging, should have a straight line of sight (slinging over friendly troops was too risky, iirc), and, most importantly, take a lot of time to train. Regarding their slow demise (they would be around up to the Middle Ages though), certain late agricultural technologies should drastically increase how much their training takes, since they came from a more agrarian background, and besides some early “biconical lead projectiles” tech to increase their pierce damage, they should not be improved much by other techs (around the middle to late timeframe of the game sling technology was historically at its peak, while others kept improving).

Going back to unit’s values: archers’ arrows should be somewhat easy to block with shields, they would lose against cavalry, who should lose against spearmen, who should lose against swordsmen (here I wasn’t sure if they should win or lose against archers, I guess it should depend on how much their shield can block). I like the hard to block/parry/dodge dynamic of the mace/axe/sword, going along with the reality that crush damage transfers more through shield and armor, axes are hard to parry, and swords have faster attacks. I have not emulated all possibilities in my head though, all remains to be tested. Javeliners are good against elephants (from a distance), which along rams are good against buildings. It’s a matter of fine tuning to get what one wants, as Enrico Fermi told Freeman Dyson: “with four parameters I can fit an elephant, and with five I can make him wiggle his trunk". No bonuses should be needed!

Edited 2 hours ago by Thalatta
Changed Spearman from "very hard to dodge" to "hard to block", but these details need to be fine tuned better.

[guerringuerrin]

One thing 0 A.D. lacks is a diversity of archer types. I think the current ones resemble longbowmen more, but there are no shortbow units. I like the idea of a short-range ranged unit with higher damage than a longbow. Perhaps, in melee, it could fight with a short sword or dagger. I’m not sure how complex its implementation would be.
 

2 hours ago, Thalatta said:

should have a straight line of sight (slinging over friendly troops was too risky, iirc), and, most importantly, take a lot of time to train

I think it would make sense not to allow tight formations for slingers if we want to add a touch of realism, but limiting slingers to direct line of sight would make them quite useless. Beyond 0 A.D.’s realistic orientation, let’s not forget this is an RTS, not a combat simulator.