Sunday, June 10, 2012

Thoughts on the Trinity

It has become fashionable among the MMO literati to denigrate Trinity-based gameplay for PvE, which are games with a group composition of tank, healer, and damage dealers. I disagree with this view. I consider the Trinity-based system to be strongest group system for MMOs presented thus far, especially for the fantasy genre.  Here are some thoughts on the Trinity system.

A Digression on Magic

A few years ago, according to Mark Rosewater, the Magic: the Gathering design team conducted an experiment. They stripped Magic down to its fundamentals, playing "vanilla" games with basic land, basic creatures without abilities, and basic spells. The simplest game of Magic you can think of.

They found that--far from being boring--these vanilla games were surprising fun and intricate. They didn't need the crazy complex spells to make the game interesting. The basic skeleton of Magic was more than fun enough to sustain gameplay.

They also discovered that a lot of the fun came from the interaction between attacking and blocking creatures. This did not happen as often in the Magic of that time period because cheap removal spells had made it a lot easier for players to clear a path, or evasion abilities to avoid being blocked.

Because of this experiment, Magic R&D cut back significantly on complexity in future expansions, especially on non-rare cards, made removal scarcer and creature combat more important. They moved Magic back towards its vanilla nature.

This approach has made Magic enormously successful, with current sets setting new sales records, and outselling previous sets. An outstanding achievement for a game that is approaching 20 years of age.

The lesson here is that, for a game system, the very basic game system around which everything else is built should be fun in and of itself. You should not rely on complexity to add fun.

Relevance to MMOs

The above anecdote crystallized some of my thoughts on the Trinity. In my view, the very basic Trinity gameplay is a tank tanking a monster, the damage dealers killing the monster, and the healer keeping the tank up. This very basic, very vanilla, gameplay is fun. You don't need to have all the crazy, wacky abilities. Those extra abilities add spice and interest. But I think the Trinity skeleton is strong enough to sustain itself.

You don't need things like coordinating cooldowns or excessive dancing to make Trinity games fun.

This is in sharp contrast to all the non-Trinity gameplay I've played. The vanilla skeleton of those games is usually just a zerg, with the monster switching attacks at random. I don't think the zerg is fun. It contains none of the teamwork of the Trinity, none of the sense of the group being stronger than the individual components.

The Third Role

That's not to say that the current Trinity is perfect. But the real issue with the Trinity is that one role, damage dealing, is far more popular than the other two roles. The problem is not with Trinity gameplay, the problem is constructing the Trinity in the first place.

Here's an idea: what if the third role was not damage dealing? What if it was something else, like debuffs, or interrupts, or crowd control?

This would immediately allow the game company to equalize damage across all three classes. That would mean that healers or tanks could do as much damage as the other role. This would make soloing much easier, and would make tank and healer classes more attractive. It would also mean that you could do something like requiring a group of five to need one tank, one healer, one debuffer, and two others. That would give more flexibility for group composition.

Threat

A lot of Trinity games, especially WoW, have made threat a non-issue when tanking. And yet, when we look at the vanilla Trinity gameplay, I think threat is actually important. It's a large part of the interaction between damage dealing and tanking.  I think basic threat is a large part of what makes the Trinity gameplay tick.

That means that the modern move away from threat is working against the natural skeleton of the Trinity. That it might be better to re-emphasis threat, to cut away the elements that make threat excessively pointless.

To re-emphasize threat means a couple changes would have to be made. The tank losing threat cannot always be an auto-loss. The tank needs to regain threat quickly, but the group should be able to survive the mob switching targets temporarily.

In my opinion, one really sees how this works when tanking in TERA. There, the trade-off between threat and reducing damage is explicit. I think it emphasizes something that WoW tanking has lost.

Conclusions

Those are my thoughts on Trinity gameplay. The basic skeleton is fun and engaging. In my opinion, vanilla Trinity is far more fun than any other proposed vanilla system, especially the zerg.

However, I am not sure that the third role in the Trinity should be damage dealing. It might be better to equalize damage across all three roles, and have the third role take care of a different function.

Finally, I think the move away from threat is a mistake. Threat is an important part of vanilla Trinity gameplay. However, I think that the complexity of modern fights has masked that factor. Trinity MMOs might be better served by reducing complexity, but re-emphasizing threat.

29 comments:

Polynices said...

I suspect people need reminding that back in the days of Everquest when the term was coined the "holy trinity" was tank, healer, and *crowd control*. DPS was just along for the ride.

The notion of a "holy trinity" was that you had to have those 3 roles if you wanted to go into a dungeon at all, DPS was just something you added after you had taken care of the important roles.

When WoW came out and largely removed crowd control, the trinity lost its third leg and people put DPS on the list to get the count back to 3. But in the original sense of the phrase it's just a "holy duo" nowadays.

Azuriel said...

I think Threat is actually a component of the "needless complexity" you are referring to here. In the vanilla skeleton gameplay, the boss is always hitting the tank - the holy trinity does not function otherwise. Introducing "boss sometimes switches targets because someone else passed an invisible line" is no different than any other complex mechanic.

I am 100% behind you on the basic argument for the trinity though. As a player, I want to pick a role and specialize.

Kring said...

In M:TG you can be short on land and not cast anything. In a trinity based game you'll be short on tanks. Both is about the same amount of fun.

The games don't move away from the trinity because the trinity is not fun. They move away from it because paying customers are not able to play the game.

> Here's an idea: what if the third role was not
> damage dealing? What if it was something else,
> like debuffs, or interrupts, or crowd control?

Damage is the only task that wins the fight. And it's the only task that can shorten the fight, therefore reduce the time in which you can make mistakes.

Your idea would be broken by the players who would replace as many player as possible with someone only focusing on damage and therefore making every fight trivial. (Remember the Ragnaros fight video back at level 60 with 35 mages, one warlock (CoE) a tank and some healers? Took them a little bit more then 35 seconds to kill him, neglecting all combat mechanics.)

To counter that you might have to make damage irrelevant and make each fight take a fixed amount of time. Then again, people love to do damage.

Redbeard said...

Threat is what really keeps DPS on their toes, and getting away from that makes the DPS' role kind of unexciting. If you know that the tank will always (or almost always) have threat, you can just spam away without any understanding of pacing. And, to be honest, pug stories are always more interesting when there's a Rogue or Boomkin or Destro Lock out there yanking the threat away from the tank.

Anonymous said...

I would go a step further back.
Group RPGs had a front line of heavy plate wearers that were supposed to take the hits. But they always dished out damage aswell.
The leather armour midline was for CC and specials, but also damage.
The backline in robes were mages and clerics, for damage and healing mostly, but not exclusive.
The problem in MMOs is, that those roles got so reduced to one ability only aka stand still and take all the hits or stand still and cast (better two) heal spells every cooldown.
Specializations is a very nice thing, if you are better in doing it, not if it becomes the sole purpose of you being there.

Helistar said...

As I already replied some time ago on Azuriel's blog: the trinity is the inevitable consequence of health bars. As long as "life" is represented by a number, the only possible "specialized" roles are tank/heal/dps. With DPS being the most important, since combat will end when the enemy's hp bar is at zero. My limited experience with systems not using the trinity is that all-out DPS is always the "win" option. This is also what happens in trinity games when survival becomes less of an issue: you bring less healers/tanks and beef up the DPS.
BTW trying to get rid of the DPS by using debuffs/cc will inevitably fail: as I wrote: it's DPS which ends the fight, so all you'll have is a shift towards the classes which in addition to the debuff/CC, can dish out the more damage.

Threat: I'm happy that WoW is dumping it. As of today, too much of tank's gameplay was on "DPSing from the front". A TPS rotation is too much like a DPS rotation, I like the idea to move tanking towards a survival gameplay, with cooldown rotation, etc. I don't know if it will work, but right now tanking was too much like DPS + 2 buttons to press every couple of minutes for the -50% damage taken.

BTW I tried to design a combat system which does not use HP bars, but the thing spiraled out of control into an absolute complex mess (the idea was combo/countercombo based). It's hard to have a simple and visible thing like the "hp bar" without falling into the same trinity trap.

Milady said...

Very well put, and I happen to completely agree with you, Rohan.

It is remarkable that when the holy trinity is removed, the system that replaces it is more akin to zerging than to an organized group. This was one of my main concerns regarding GW2. If there is no specialization (in the trinity roles, or comparable ones), there will be no structure in the combat, just an all in in which people take turns to get hit (and die, and get ressed). Which is coincidentally exactly what happens in D3 Inferno. Since all classes are damage dealers, and tanks (Barbarian, Monk) are not viable until they have an absurd amount of gear, there is no structure in the group dynamics.

Juzaba said...

I disagree with the people posting that groups will just create the damage role anyway. Or, rather, while that may be true, it would be up to the devs of such a game to balance damage across roles and classes such that even if you used a Tank Class as your DPS, there was no actual loss of DPS. At least, that's what I understand Rohan's idea to be.

There are other options for "third role" in addition to CC/utility. For example, D&D 4e includes "Striker" as a role in their 4-way of roles. Though Strikers most represent the DPS role, they specialize in burst damage. Presumably, a third role could also be filled by classes that do the same baseline DPS as anyone else in the game but bring burst damage as their tool, presumably for use during burn phases and the like.

Such a plan involves even more careful balancing than the CC idea, since you'd have to prevent Strikers from being able to turn on the burst in any situation. Perhaps abilities like "for the next 20 seconds, your abilities do X% more damage against Vulnerable targets" where there were only certain mechanics that created Vulnerability. Something like that, while feeling arbitrary, at least gives devs better control over burst damage than a base +X% damage buff.

Point is: I agree. The holy trinity is not borked as a theory.

RJ said...

Something you either didn't notice, or didn't highlight, in your post is that threat gain is dis-emphasized not because it's "unfun" to lose threat or anything, but because Tank DPS is scaled to be specifically lower so as not to compete. In a system where the "Trinity" did comparable damage, the tank would only need a marginal threat booster in order to maintain, and thus it would be a lot tighter.

In other words, if the engine is only letting a tank do 100 DPS when the DPS is doing 1000, then the threat gain mechanics NEED to be utterly ridiculous, or else the system doesn't work at all.

Helistar said...

I disagree with the people posting that groups will just create the damage role anyway. Or, rather, while that may be true, it would be up to the devs of such a game to balance damage across roles and classes such that even if you used a Tank Class as your DPS, there was no actual loss of DPS. At least, that's what I understand Rohan's idea to be.

"Welcome to the world of tanks" :P

Unless there's some kind of tradeoff, in your game the best would be to have all tanks. Same DPS + extra survival! No healers needed :)

As for sustained/burst DPS, in WoW it's *already* a big deal (see the insane arcane mage stacking for HM spine which was used by Kin Raiders). Unless you allow DPS classes to multi-spec into different types, this opens more problems than it solves.

Not to mention the risk that given enough burst damage, you may end up with enough burst damage that it ends the fight before anything else happens.....

Bearness said...

Really great post. I completely agree with all ideas. I think the holy trinity is the best model for any large group (4+ people). Also, it creates dependency (tank needs healers, people need tank to protect them from monsters, etc), which I think enhances gaming experience and makes grouping feel meaningful. Of course, it sucks when you have someone who is not good at their role, but that is the case for any model.

I believe you're correct that 3rd role shouldn't be DPS because in holy trinity, you need to depend on each other for something unique you can't provide yourself. Everyone can DPS, it's just a matter of how much they can DPS. It's not something you "need" from another member. Vanilla WOW had it right that you wanted to sap/sheep a monster or two, and then take on rest of the group. But, it was missing the normalized DPS for all party members factor. Everyone should be able to do roughly same DPS while still performing their role.

I think best way to handle normalizing DPS would be like this. First, everyone's DPS abilities do roughly the same damage for obvious reasons. Second, DPS abilities for all classes should not require resources (mana, etc). Their only limitation should be cooldown. So, using your class specialty (for example, healing or pulling threat, or using CC) doesn't stop you from DPSing because you've used mana for healing instead of fireball. It will briefly interrupt DPS rotation, but that's it. Meanwhile, class abilities are based on something like a mana bar so that you have limited uses and must manage them during the fight. I think this will make the game play far more interesting and engaging.

Anonymous said...

There are really only a few ways to eliminate the trinity. One is to make all fights trivial so that no one has to specialize--they can if they want but it isn't required. The other is to force everyone to specialize both in dps and "damage taking" since heal and tank are really two sides of the same coin. I suppose the third way is to create difficulty by way of "dance" mechanics, where fights are puzzles that have to be solved with moving around and using items or something.

For me, specialization has become boring. I like single player games now because it is simply more interesting to have to use a broad range of skills to both kill and survive rather than rely on others to do the surviving while you do the killing.

Protectorate said...

I think your thoughts on the trinity and getting back to basics are great. I much preferred I stanching in BC to now, but I didn't like the putting a group together and waiting forever.
Speaking from a tanking role - I would love to get back to threat being an issue, it makes the game dynamic and more fun to me. However, unless game companies find a way to pair up good players with other good players in similar gear, bringing threat back with the current versions of quick pve matchmaking is a huge mistake. The reason threat became a non issue in WoW is solely because of the looking for group mechanic - a fresh 80 or 85 could keep threat when they were put in the same group as an ICC or DS dps.

Bearness said...

@Protectorate, I agree threat is a fun mechanic and WOW should have kept it. Fresh tanks were having trouble because because threat = DPS. So, if everyone did roughly same DPS, and if DPS was only a portion of the threat equation (ex. big heals or CC/interrupt abilities pull more threat than big DPS), that would've been a better way to handle the threat issue IMO. But, then WOW would have to redesigned from ground up.

Anonymous said...

Your idea of the holy trinity going back to the basics and equalizing damage among all classes is a great idea and would work wonders for a PVE only game. The problem is WoW must balance for PVP as well and thats impossible if all classes do the same damage. Tanks would smash face while taking little to no damage and healers would heal to full then kill you. Arena and PVP balance is what holds WoW PVE back from being alot more interesting and more enjoyable to play IMO.

Rohan said...

@Helistar, I actually think that World of Healers is more likely with equalized dps, because healing depends on boss damage output. Boss does more damage, more healers switch to healing. Boss does less damage, some healers switch to dps.

@Milady, the other alternative to people dying in sequence appears to be kiting. Whoever the monster is attacking kites the monster until it switches. I don't think random kiting is fun either, and judging by the reaction to Faction Champs in Trial of the Crusader, most raiders agree with me.

@Anon, I'm not really sure that PvP would be that badly off. If the third role had CC and interrupts, I think that would make things more equitable. It might even set up a rock-paper-scissors sequence:

Tank beats Interruptor because tank takes less damage.
Interruptor beats Healer because Healer's attacks are disrupted.
Healer beats Tank because the tank is eventually worn down.

Wulfstan said...

The GW2 beta is a great example of what you miss when you lose the holy trinity.

Every dynamic boss event lead to a zerging with no strategy/interplay between toons. Everyone swarmed the boss, spamming their abilities off CD, with the boss randomly turning and one-shotting people (mostly melee).

There might be a complex interplay of debuffs and cross-class combos somewhere in the gameplay, but it was no visible or trained. It would also be more complex to learn across strangers. As least with holy trinity you know what is expected on your role.

(following post on HT thoughts)

Wulfstan said...

I think there are some risks to making everyone have equivalent DPS. This assumes that everyone wants to DPS, and I don't think this is the case.

I believe the HT roles do specialise around distinctive gameplay...

For tanks it is mob control, and responsibility for self and others. There is a distinctive "leader" role. Raid tanking is about a partnership with another tank, rather than a competitive relationship.

DPS is about personal skill tracked in a very measurable way. It is very competitive between raid members. (Interesting dynamic in SWTOR without damage meters...)

Healing is reactive and responsive (even when done proactively), and again is a team relationship in raids.

The problems:

These roles don't appeal equally to people, resulting in inbalances.

Sometimes one role is harder than the others, forcing people into styles they don't like and further imbalancing numbers.

Threat is an interesting one: as a raid tank I took great pride in my threat output, and saw it as a measure of my ability. However, in 5-man PuGs I felt I was fighting my party, rather than fighting the mobs. For this reason, plus scaling problems, it had to get fixed.

dahut said...

In my opinion, the trinity tends to become fairly boring. If you take it down to its essentials it's a tank and spank. Tank absorbs, healer heals, and dps damages. It's only when extra abilities, fires, bombs, aggros swaps get added does it become fun in my mind.

Thinking back to my time in WoW the most fun fights I had were ones that veered away from the trinity.

The Faction champs in Trials of the Crusader.

Razorgore using the kiting method.

Mage tanking Korsh Firehand.

I must the Vael fight in BWL was pure trinity distilled and was quite fun.

If I were to boil down my preference it would be a more pvp experience than a trinity tank and spank.

Anonymous said...

With DPS being the most important, since combat will end when the enemy's hp bar is at zero.

Doesn't that make the healer the most 'useless'? Even a tank does damage while body blocking the enemies. Only the healer plays a game of whack a mole to keep everyone alive.
In a good trinity game, shouldn't a tank be able to take quite some damage without needing to get healed?
With a self heal, would a holy duo system work, with 3 people blocking enemies and being able to eat damage and 3 other dishing out more damage from behind? Occasional potion/self heal to stay alive for everyone?

Kring said...

> It is remarkable that when the holy trinity is removed,
> the system that replaces it is more akin to zerging than
> to an organized group. This was one of my main concerns
> regarding GW2.

I don't think that has anything to do with the holy trinity but only with the difficulty. You shouldn't compare WoWs hardmode raiding with GW2s quest "boss" mobs. You don't really notice the trinity in WoW's LFD runs anymore (besides the long wait time, unless you're a tank). Or when was the last time you experienced something like focus damage in an LFD?

An organized group or a zerg depends mainly on the required difficulty and not on if a trinity is used or not. GW2s system has a lot of potential for difficult fights. Like require ranged to cast conditions on the mob to reduce its damage and prevent it from killing melee. Of course, you don't see that on a level 3 world quest boss. But I also rarely see a mage sheeping in LFD heroics.

Fn0 said...

"The above anecdote crystallized some of my thoughts on the Trinity. In my view, the very basic Trinity gameplay is a tank tanking a monster, the damage dealers killing the monster, and the healer keeping the tank up."

This is not how it works in WoW, nor MoP. You have raid and tank healers. The healers often play whack-a-mole (IMO quite boring). If you play a disc priest you can make it less boring (and more raid heal, less tank heal) by going atonement. Then you do damage while healing. This is a hybrid with a clear primary role and a secondary role. In MoP, more than ever before, the tank can play better by mitigating more. The tank is actually sortof a (tank/self) healer much akin to the blood DK is now (again, fun IMO). Why are they fun? You're busy whole time, you're playing whole time, you need to optimize whole time and in your secondary role you also help your team. The difference between such and e.g. GW2 is that in GW2 everyone's primary role is damage dealing whereas in a game like WoW everyone has a primary role (healing, tank, DPS) due to LFD/LFR mechanisms (in normal and hardmode this is likely to change).

I am not sure about the holy trinity, but I sure do find it unrealistic the boss is always trying to kill the guy with the shield first whereas that guy does little damage but is somehow "threatening". In a game like WoW, bosses (and trash) are very dumb in their target picking.

"This is in sharp contrast to all the non-Trinity gameplay I've played. The vanilla skeleton of those games is usually just a zerg, with the monster switching attacks at random. I don't think the zerg is fun. It contains none of the teamwork of the Trinity, none of the sense of the group being stronger than the individual components."

This is not what my experience is in GW2 beta. You make the mistake thinking "holy trinity" is all what defines roles, but that is not true if classes have unique abilities and builds. But also: if you don't play a "holy trinity" game but can (with due respect) "tank" or "heal" with some abilities such as kiting (slows, teleport, knockback) or CDs (absorb, heals), wouldn't that work as replacement? Games which don't have "holy trinity" still have roles. And in GW2 from what I've seen there is various abilities like slow, teleport, stealth, knockback, selfheal, teamheal, stuns with professions able to respec by switching e.g. weapon or "stance". Some comment said this is about ressing the fellow who just tanked in GW2. Wrong! Only if he failed to perform! I read someone comment about Diablo 3. Remember in Diablo 3 you cannot respec on the fly; in fact your spec cannot be perfect having all the utility, and once you reach 60 your valor buff discourages respec. GW2 doesn't have this issue since at the very least you can respec by swapping your weapon, in combat. Is it enough? I don't know... yet.

"Here's an idea: what if the third role was not damage dealing? What if it was something else, like debuffs, or interrupts, or crowd control?"

Exactly.

So we have alternatives to think about:

1) Dynamic (based on what?) tanking, healing looking at games which implement this. Alert players when they are about to tank?
2) Everyone damage dealer with additional functions meaning everyone has to do damage output + their unique role.
3) Enrage timers (soft or hard). They exist so you cannot stack healers + tanks.


Anonymous wrote: "Doesn't that make the healer the most 'useless'?"

No, see my point #3.

Green Armadillo said...

EQ2 had this going for a while, where Dirges (a class that buff melee - including tanks) were so powerful that we were effectively guaranteed a group slot (6-player groups) even though we played functionally like any other stabby rogue/scout. The result was often 3 DPS LF3M tank, heals, Dirge.

As long as there is a non-zero number of DPS slots and DPS is fundamentally more desirable, adding more roles just means even more places for group-finding to bottleneck.

Rohan said...

@Green Armadillo, but there would be no "DPS" classes under an equalized DPS scenario. Other means one of [tank, healer, utility].

So a group could be tank, tank, healer, healer, utility, with one tank and one healer mainly dealing damage instead of formally tanking or dpsing.

Ahtchu said...

Loved this post, and agree with its basic premise. I'm glad, also, that you brought up the example about how simple != boring, as the inverse is often true.
I wrote a series on the necessity of the trinity system and how it manifests properly across the combat minigame. Might interest you.

Syl said...

The biggest issue of the trinity is that by the same virtue that it creates cooperation, it does also prevent it - due to its inflexibility.
so maybe it would be useful to make observations about different types of cooperation; and ask players what they prefer because you cannot have it both ways in the same game.

all that means is that both an MMO like WoW and an MMO like GW2 have their own target audience.
is that such a bad thing?

Craftygod said...

I was glad to see you refer to TERA tanking, as I think the combat system as a whole in TERA moves toward eliminating a lot of the issues with the "trinity" in WoW and other games. In TERA, a really good tank makes it possible for the healer to do some dps, and no matter how much a DPS wants to just stand and do damage, he can't. Everyone has to move, dodge, and be responsible for himself during a boss fight or he will die. Yes, the tank takes most of the hits and mitigates that damage through blocking, but the other players will still take lethal damage if they aren't paying attention.

However, the problem with the trinity in my mind isn't game mechanics, but socialization. I think the biggest draw back is the fact that groups are dependent on these roles to be successful,and that really comes down to finding a tank . . . so, why don't people roll more tanks? Are they not fun? On the contrary, I think tanking is a lot of fun, but there is the added problem of a tank having to lead the group through the dungeon, and we all know that tanks are quickly blamed if someone dies because of a threat drop, etc. In a way, tanks are held to a higher level of responsibility than any other player in the group (definitely more so than DPS), and this is why people don't really want to play tanks.

I'm not sure what the solution to this problem is, but as long as one type of character is held to a higher level of social responsibility, fewer people will play them, and we will continue to see an imbalance in putting parties together which is the biggest complaint against the trinity.

souldrinker said...

The question of threat is highly subjective one.
For a long time, making enough threat was the main tank chore at WoW. The stories told about great tanks were about how tank got disconnected without his character kicked out from the game and the raid killed the boss 4 minutes later without pulling threat.
However, under this approach the tanking doesn't feel different from DPSing. You're still maching buttons bit prioritise high threat abilities over high damage ones.

For me the real attraction of tanking was the control I had over the mobs. The fights where tank does really shine, where tank makes or breaks it, are almost always about proper positioning, or movement, or using utility abilities (rather than shield slamming on CD).

Imakulata said...

@Rohan:
The idea of making everyone a DPS+something else hybrid is an interesting one. If you consider the usual roles (DPS, tanking, healing, support, CC/debuffing), DPS is the only offensive one (CC/debuffing is ultimately a defensive one because its goal is to reduce incoming damage from the mobs). So I think you definitely have a point.

The problem with threat is that generating threat by itself does not harm the monster or help the group. Unlike other roles and other tanks' tasks, threat works neither against the enemies nor for the teammates but against the teammates. Additionally, WoW's threat does not work with the enemies being players. While allowing the tank to choose between a full attack and turtle mode for the group is not a bad idea, I would like to see it to be implemented in a different way. Maybe allowing the tank to decrease their enemy's defense at the cost of their own?

I think there is another problem with your solution is the three DPS+X hybrids, 2 pure DPSers. Based on WoW's trinity, it is possible people would aim for the pure DPS spots because of shared responsibility. What does the model offer to encourage people to try the hybrid roles with high personal responsibility (due to being the only one who is "supposed" to do the task)?

@Kring, I believe it is possible to set a low bound for defense, i. e. have the mob dish so much damage that X players need to play defensive roles no matter how much damage the offense deals.

@Milady, I haven't played any explorable dungeons in GW2 (despite being in 2 BWEs, my highest char is only 17) so I can't tell how it looks there but the huge events are not supposed to be too challenging. (@Kring says the same in his post, I think.)

The events also suffer from lack of parallel targeting which I consider to be a huge problem for bosses fighting huge groups. The boss's attack power has a soft lower bound based on how much damage can the group prevent and heal up, and a hard upper bound based on how much damage is needed to kill a single player no matter what. Since most bosses lack parallel targeting that would enable them to deal damage to more targets as a basic mechanic, there is a group size where the lower bound for the particular difficulty gets higher than the upper bound.