How much can you mechanically differentiate classes from each other?

[Val the Moofia Boss]

Mostly thinking about this in relation to games that have lots of classes, as a common idea I have heard is that eventually you run out of mechanics to meaningfully differentiate classes from each other and new classes begin retreading/infringing upon old ground covered by preexisting classes, and at a certain point you have enough classes that effectively do the same thing and the only real difference is the aesthetics you want.

The different knobs off of the top of my head:

Healer roles:
- Typical healer
- Regen healer
- Barrier/shield healer (anticipates heavy damage coming up in a few seconds and places a short duration shield on the party member to absorb first before his HP bar takes damage)
- Blood mage, siphon HP from enemies and redistribute it to allies. Redistribute or equalize entire party's HP pool (WoW Restoration Shaman's spirit link totem).

Tank roles
- Straight up regular damage mitigation/reduction, might have to press a button at the right time to hold up a shield to reduce incoming damage.
- Blood tank (HP healed as a portion of damage dealt, thrives when facing many enemies at once)
- DoT/stagger tank (ie Brewmaster Monk from WoW, a large portion of any incoming hit is converted into a DoT on the tank, allowing them to survive hits that would immediately kill other tanks)
- Shadow clone tank (Ninja from FF11, creates clones that soak incoming hits before the ninja runs out of clones and starts being hit directly).
- Parry tank, enters into a stance at the right time to reflect a huge amount of damage back at the enemy, the more powerful the incoming strike was the more damage dealt.

DPS differentiators:
- Melee/ranged
- Cast/charged up times or not (melee classes could have cast/charged up attacks, some ranged could attack while moving)
- Burst or DoT
- Elemental damage type dealt ie ice/fire/light/etc
- If ranged, does it send forth projectiles that can be blocked, do projectiles travel in a straight line or can they arc over obstructions like mortars.
- Abilities might deal more damage if dealt from a certain direction, ie the sides of the enemy or the rear. How much the rotation requires you to move back and forth between these spots might increase the difficulty of the playstyle, and might also simulate the idea that the class is a fast paced or agile attacker.

Support roles:
- Buffer
- Debuffer/Crowd Controller
- A supporter "class" disguised as a character that acts as an accessory riding around on another class or player, like the second head of a two headed race that is piloted by one player (Cho'Gall from HotS), or a little cute mascot race that sits on the head of another player (Abathur from HotS, Yuumi from League of Legends).

Pet class/Engineers
- If it is a pet class, whether it has one big pet/summon/robot or many smaller pets (ie Necromancer with lots of zombies).
- If the class constructs/summons/places something on the ground like a turret/shield generator/totem/landmine/etc.

Miscellaneous
- How much the class relies on consumable items rather than just innately replenishing resources like MP, ie needing arrows to shoot a a bow, or both balls and powder for a gun, or multiple different types potions for a chemist to mix and throw, or requiring expensive parts for their robot/mech to use.
- Classes that can manipulate the terrain like freezing water to cross rivers or raising earth walls to block incoming projectiles or trap targets from escaping.
- Whether the class can enter different stances or shapeshift forms.
- Whether the class is slow (ie maybe cast times/charged up attacks, or simply a lot of passive autoattacking in between abilities coming off of CD) or fast paced (ie button smashing and memorizing different combos like in a fighting game).
- How much mobility, ie can they sprint or do a flying kick across the battlefield or have to slowly run on foot
- Armor type ie how squishy they are. Heavier armors like plate could cause the wearer to become vulnerable to electricity or fluctuations in temperature. Lighter robes could be vulnerable to fire.
- If they can fight while mounted or not. Might overlap with pet class if the mount itself can fight (instead of, or in addition to the player class?) too, like a tiger mount clawing at the enemy (instead of or while the player also swings their sword?)

Are there any others I am missing?

[Tadeusz]

DPS classes may have another bar instead of MP that starts with 0 and fills during combat. These points are then used for skills or maybe some self-buffs. It's usually used for berserkers or monks.

[J1M]

There seems to be a mismatch between the title (asks an interesting question) and the OP which is a list of what popular games do today. Here are my thoughts for the former.

Foundations of design:

• In order to have advanced classes and interesting interaction twists you need a baseline for comparison.
• Choices with significant tradeoffs are more interesting than cosmetic ones.
• The best designs wed mechanics to established archetypes and themes.

Common pitfalls:

• Trying to keep basic classes balanced with advanced classes at high levels of play is not necessary. It's okay to retire these by giving them a level cap or similar.
• Not explicitly sharing the intended tradeoffs of a class with players makes it very easy for lazy designer to start ignoring those tradeoffs, especially as staff turns over ("paladins need a ranged attack!", "why don't I have a self-heal, everyone else does", etc)
• Looking only at output for differentiation. It is possible to create two classes that do the same job but feel very different if one of them expects a button press every half second and the other relies on things like stances or reactions that happen on average every 5 seconds.
• Every class ends up with a builder/spender resource system.
• Too many variations of spells (fast fireball, slow fireball, instant fireball, fast frostbolt, slow frostbolt, instant frostbolt), at least 2 of these should be cut so there is a meaningful difference between fire and frost.
• Removing drawbacks because players complain about them (for example, doing max damage requires standing still, but that is difficult to do given everything else going on).

Examples of small differentiators that can have a huge impact:

• A damage class that deals more damage or generates free resources when it takes damage, such as standing in fire.
• Positioning requirements involving more than one or two characters.
• Raid leader class that is not intended to provide throughput, but rather to preform in-game actions that lead to things like markers and raid buffs. (Commander role in Natural Selection is an example)
• Different damage types that actually matter.
• Classes that excel at different parts of the game (PvP, PvE, etc), especially if players can switch between them.

[TKVNC]

The problem with differentiation is that it locks classes into niches that make them either powerful, or ****, with no inbetween. Far better to just give a range of options within a catch-all class.

Fighter, mage, rogue

Add the ability to multiclass and you solve every issue.

[J1M]

The problem with differentiation is that it locks classes into niches that make them either powerful, or ****, with no inbetween. Far better to just give a range of options within a catch-all class.

Fighter, mage, rogue

Add the ability to multiclass and you solve every issue.

Ideally you need a team with varied skills to succeed and you are happy when your friend is doing great at their role.

[TKVNC]

The problem with differentiation is that it locks classes into niches that make them either powerful, or ****, with no inbetween. Far better to just give a range of options within a catch-all class.

Fighter, mage, rogue

Add the ability to multiclass and you solve every issue.

Ideally you need a team with varied skills to succeed and you are happy when your friend is doing great at their role.

Friendpilled

[WhiteShark]

Looking only at output for differentiation. It is possible to create two classes that do the same job but feel very different if one of them expects a button press every half second and the other relies on things like stances or reactions that happen on average every 5 seconds.

I disagree on this point. Pressing buttons more or less frequently to do the same thing isn't a meaningful difference; it may create an illusion of difference sufficient to fool some people for a while, but eventually someone will run the numbers and realize that they are the same, or that X is strictly superior to Y. Ideally, there would also be a difference in efficacy according to the type of encounter. For example, the stance/reaction damage dealer could be designed to do better in protracted combat against a single enemy, while the button masher made better against swarms of weaker enemies.

[J1M]

Looking only at output for differentiation. It is possible to create two classes that do the same job but feel very different if one of them expects a button press every half second and the other relies on things like stances or reactions that happen on average every 5 seconds.

I disagree on this point. Pressing buttons more or less frequently to do the same thing isn't a meaningful difference; it may create an illusion of difference sufficient to fool some people for a while, but eventually someone will run the numbers and realize that they are the same, or that X is strictly superior to Y. Ideally, there would also be a difference in efficacy according to the type of encounter. For example, the stance/reaction damage dealer could be designed to do better in protracted combat against a single enemy, while the button masher made better against swarms of weaker enemies.

When I say the "same thing" I am referring to something like "melee DPS". Practical translation of my example: a rogue and a paladin using seals in older WoW. One often wanted to spam input buttons on a shorter global cooldown and the other would have GCDs that were empty, which could be used to aid allies with blessings from time to time.

Not everyone wants to opt into carpal tunnel and I think in something like an MMO there's absolutely room for different classes having different APM ceilings.

[WhiteShark]

When I say the "same thing" I am referring to something like "melee DPS". Practical translation of my example: a rogue and a paladin using seals in older WoW. One often wanted to spam input buttons on a shorter global cooldown and the other would have GCDs that were empty, which could be used to aid allies with blessings from time to time.

Not everyone wants to opt into carpal tunnel and I think in something like an MMO there's absolutely room for different classes having different APM ceilings.

That's fine, I just don't think it's a sufficient differentiator by itself. I'd be pretty peeved if my class' job was done just as well by another class for less effort. I don't know WoW well enough to analyze your example, but if I assume that the rogue did more damage overall while the paladin had more support functions, that's a sufficient difference, but it's one of more than just required APM.

[Val the Moofia Boss]

Foundations of design:

• Choices with significant tradeoffs are more interesting than cosmetic ones.

I do think this is an often overlooked part of design. Things are often just as much defined - if not more - by what they are not than what they are. Naked berserkers that take considerably more damage due to their lack of armor, or hunters who have to carry expensive consumables around with them like ammo and food for their pets are a little more distinctive than classes with no visible drawbacks like that.

Common pitfalls:

• Trying to keep basic classes balanced with advanced classes at high levels of play is not necessary. It's okay to retire these by giving them a level cap or similar.

I think one fear might be that a player might not find any of the advanced classes that their current class can graduate into to be appealing, and would thus prefer to remain as their current class.

[Tweed]

Depends on how much work you're willing to put in to make them different from each other. ADOM's psionic class, whatever it was called worked by reaching out to enemy minds and I think some attacks could even pass through solid object which could be a blessing or a curse. The big problem was if you attacked an undead or chaotic monster, it did bad thing to your brain. So it wasn't just magic by any other name.

[J1M]

Foundations of design:

• Choices with significant tradeoffs are more interesting than cosmetic ones.

I do think this is an often overlooked part of design. Things are often just as much defined - if not more - by what they are not than what they are. Naked berserkers that take considerably more damage due to their lack of armor, or hunters who have to carry expensive consumables around with them like ammo and food for their pets are a little more distinctive than classes with no visible drawbacks like that.

Common pitfalls:

• Trying to keep basic classes balanced with advanced classes at high levels of play is not necessary. It's okay to retire these by giving them a level cap or similar.

I think one fear might be that a player might not find any of the advanced classes that their current class can graduate into to be appealing, and would thus prefer to remain as their current class.

I think this is a ludicrous fear harbored by people who have access to debug commands. "Professionals" that are no longer players. Imagine this person, a lead designer, sitting in a meeting arguing that we need to cut some side quests so that players have the option of continuing to progress as a scoundrel instead of as a Jedi Knight after they meet the Jedi council in KOTOR.

[NotAI]

They should be able to always refuse the main quest, only to be dragged back into it.

Suppose they play as a scoundrel. A rogue. "Picaro"

When they get to the Awesome Main Quest Room with the Captain of Awesome Guard of Awesome Land who, sitting across a Magnificent Hardwood Desk, tells him he is The Awesome One who will save The Awesome Land from the ****** Demon Lord, the protagonist should always have the real option to refuse the quest, and urinate on the desk. If not on the desk, then in the corner. But in full sight of the quest giver. Then leave.

The game should then have means to draw them back into solving the main quest anyway, while following another objective of some apparent side quest.

What is actually needed it to add hidden routes to side quests that bring you back to the main quest.