Well here it is, my commentary and/or attempt to balance the combat system in ESO. And like many long works that few will bother to read in its entirety, it has a few disclaimers:

• This essay is a WORK IN PROGRESS! It is incomplete, does not necessarily flow, and probably still requires proof-reading. I intended to update it as I get the time. In the meantime I didn't want this to collect dust in my hard drive.
• Many issues and topics are intricately interwoven, much like any combat system, and may require jumping back and forth to fully grasp the implications.
• When I make assertions I will try not to support them with anecdotes or what have you to try to improve my credibility. Not because I don’t have examples to call upon, not that I have no experiences with these, but because I feel that it would be somewhat of a given and rather pointlessly drag on and distract from an already major wall of text.
• This is a major wall of text. A TL;DR for each section has not been provided at this time. Read at your own discretion.
• Unless otherwise stated, assertions are made from an endgame perspective.
• The goal here is not the piece but the whole. One part may seem off-balance but could be countered by an equal imbalance somewhere else. I invite you to assess my assertions yourself and to point out any apparent contradictions. This was not all written in one sitting; they may very well be there.
• Please, oh, please, oh PLEASE do NOT quote the entirety of my post. For all our sakes.
• This is entirely combat related but I posted it in General Discussion. FORGIVE ME, FORUM MODERATORS!
• I am going to reserve the first few posts. I have reasons.
• Please be gentle with me. I put a lot of effort into this.

Now, in absence of a proper introduction, let's jump right in.

What makes ESO different from other MMOs? Arguably the biggest thing is the targeting system. This makes for a unique dynamic. As a DD or a tank, you need to pick your target out in a crowd, not by pressing tab to cycle through NPCs, but by keeping the center of the screen trained on them. There are a few implications of this, the largest being, against a swarm of monsters, it can be very difficult to pick out your target if it isn’t in the first row. This is more realistic than using tab targeting to attack a monster some distance away through other, closer obstacles (more monsters) in your way. There are many potential problems with not being able to target a dangerous monster that is “hiding” in a swarm of monsters that the tank is holding all at once, so the developers deviated from the standard practice of having the tank control and hold aggro against ALL the adds: they omitted an aggro system for tanks with the exception of a single-target taunt. This made it so that adds would, by design, split up and attack various group members all at once, decreasing the mob density thus making individual monsters easier to target—and the popular if not niche pet class archetype more problematic as pets obscure the enemy player. Healers experienced a similar paradigm shift as they could no longer single out a player for heals through a vast crowd of enemy and ally alike. This necessitated the standard of all heals being “smart,” targeting the most in-danger ally, or otherwise AoE. This arguably removed a bit of depth from the three roles as an unfortunate side-effect.

For whatever reason, the overarching category of support was not included in ESO, though that probably is tied to the arbitrarily small group size of four. Support is typically a difficult and complex role that can entail CC, buffing, and debuffing all while contributing as an off-healer/tank/DD. This would be a far more demanding role than the other three, but having a group consist of 1 player in each role doesn’t make for an entertaining group structure. I have 2 guesses for why the standard group size is only four: first, Elder Scrolls games are traditionally single-player so it may be a consideration to long-time fans, and second, it facilitates the any-class-can-do-any-role mentality. A consequence of a small group size, however, is players have a lower bar to achieve to solo content scaled to groups than if the group size were, say, six.

ESO has another problem in the general size and composition of its groups. The standard group size is four, suitable for 4 classes with only 3 general roles. These groups consist of 1 tank, 1 healer, and 2 DD, per design of the content. This means that, for four-man groups, the population is suited toward and ideally made up of 50% DDs, 25% healers, and 25% tanks. In order for this to be useful/effective/whatever, this should remain constant throughout all levels of content. Trials, unfortunately, break from this. From experience, trial groups of 12 consist of 1.5 tanks (an off-tank may be present) 2 healers (sometimes 3, but often considered overkill) and 8.5 (off tank had to come from somewhere) DDs. So, roughly 12% tank, 16-25% healer (let’s call it 20%) and probably 68% DD (rounding, and stuff. Nefarious pseudo-math). This creates a problem when people who are built to fit the population of trials come to use the group finder. There appears to be far too many DDs relative to healers and tanks, especially tanks. This means tanks get groups almost immediately, healers get paired to groups in short order, but DDs can wait near an hour or more to be paired. Ideally, this would self-correct due to the wait times, but then folks will find that their tank spec can’t get in a trial because people are only running with one, simply put that supply is twice as large as demand. These tanks are encouraged to DD then, to fill the hole.

The point here is trials should be designed to not only encourage the same or very similar group structure as typical four-man dungeons, but to discourage a significant deviation from that. Ideally, the standard should be 3 tanks, 3 healers, and 6 DDs, though slight changes in that should be permitted (for example, 2 tanks, 3 healers, and 7 DDs).

Unfortunately, however, this will not solve all of the problems with grouping and roles. We players put too much emphasis in the role of DD. DDs, outside of a designated debuffer, are supposed to DD and only DD. Kill, kill, kill is the ultimate goal for them, and two things contribute to this mindset. First, most mechanics in dungeons are related to how much dps your group has. Higher dps allowed you, in many dungeons, to cut the fight/boss mechanic short or even avoid it entirely. Second, heals are so powerful that healers have time on their hands; it’s not a feat on the healer’s part if no one dies, it’s kudos to the DDs for not standing in stupid and power creep is responsible for this. I am in a few math classes right now and it is giving me the urge to say that power creep is change in TTK over time, or d(TTK)/dt, but before we can get into that let’s look at TTK.

From a very broad standpoint, damage and healing are all based on a magic number known as “time to kill”, or TTK, which can be represented as a ratio of (health of defender)/(dps of attackers). Content is designed so that players have a specific range that their TTK falls into that must be enforced. This enforcement can be done by way of increasing the health of the defender, be it a dungeon boss in the case of PvE or a player in the case of PvP, or by decreasing the dps of the player. If we could know the intended TTK, we could quantitatively model “balance” in the game, an otherwise slippery qualitative term used by players when something doesn’t feel quite right.

Now let’s invent a term called “time to heal”, or TTH, and define it to be (health of target)/(hps of healer). This is the time it takes to fully recover all the health of a friendly target. This isn’t just some arbitrary number, it is proportional to TTK by some coefficient c, such that TTK/TTH=c. I’ll try to keep this basic; Suppose we are in PvP where everyone has a similar TTK and TTH, and further suppose c > 1 : what we have here is one healer out-healing the damage that one or more players can do, on average. It’s pretty clear that there are dangers to “balance” when c is in this range. Next let’s suppose that c = 1: then healing capabilities are just as strong damage capabilities such that one healer counteracts one DD, on average. And finally, if 0 < c < 1, then, on average, DD capabilities out-pace healing capabilities. Notice how I stated “on average” here. What that means is, given an arbitrarily large health pool, this is the trend we observe. It is not quite as simple as that. Dps and hps are average values, but are made up of many small DoTs/HoTs with large burst damage/heals thrown in. With relatively small health pools, burst can (typically to the advantage to the attacker) overcome a lower average TTK or TTH. We actually see that the game design facilities this phenomenon with executes, skills that receive around a 4x damage modifier to low health targets, and certain passives that also boost healing on low-health targets but to a far lesser degree, typically a 1.3x modifier.

So what do we do with these theoretical values? Well, now that everything in ESO is scaled to the level cap, things are a little different than before but the principle is the same. We can take the average group at average gear at any given level and measure their average TTK and compare that number to all the other levels. If scaling is done right, the TTK will remain constant throughout all levels—and if it is it reduces the meaning to having character levels for progression, interestingly enough. If the overall trend is that TTK falls as you go up in level and get better gear, as most games do and to a certain degree should, then we can actually see power creep. But of course, a small amount of this kind of power creep is necessary. The power creep that most people refer to as a problem is when either TTK decreases too much as you level and get better gear (at on-level content, which is now everything in ESO) or when TTK decreases significantly over the course of many updates. Naturally, TTH can be evaluated similarly.

What is the power creep we see in ESO and how do we know it exists? Players being able to solo group content is a pretty clear indicator. Group content, by definition, needs a group to complete. It is a contradiction to solo this content. In the current state of the game it is not uncommon to be able to solo group content, and some trial bosses can even be soloed. For these endgame players, nothing is scaled properly. There is a huge difference with freshly made characters, however. These characters cannot solo group content or trial bosses. In fact, a group of these players can take an unholy amount of time to complete these dungeons, or even fail to complete them on normal difficulty. There is a massive power gap in between the lowest character and the highest endgame character, and assuming the content is scaled appropriately for the newest of players, endgame characters are far too powerful and need scaling down.

For endgame characters, TTK and TTH have gotten far too low. Enforcement of TTK did come in One Tamriel when bosses got increased health, but TTH, which cannot realistically get any lower, was left unaffected by that change. TTH is so low that a single Breath of Life from a Templar, a single button pressed at a moment's notice, can instantly take two typical group members, half of the group, to full health. Tanks may need that button hit twice. Because healing is this powerful—far too powerful—there is nearly no danger from dungeons; if it’s not a 1-shot, it’s not a threat. Tanks aren’t needed so long as the other roles can avoid 1-shots, something that is the sole provision of bosses. When group content is designed appropriately, players’ TTH is higher than dungeon monsters’ TTK on the squishier roles but TTH is lower than that same TTK on the tank, meaning the healer can sustain the tank but not the DDs(or himself) when under heavy, consistent fire. When that delicate balance is struck, tanks are useful, necessary and valued. Admittedly, this balance is more difficult to attain in our combat system where the tank is not expected to hold everything and all group members are under attack in any given encounter. There are two obvious solutions to this: given the current TTH, which probably measures 1 second, everyone but the tank must be killed by dungeon monsters in less than one second, while the tank could survive the same onslaught for more than one second. Else, increase the TTH. That is, weaken healing substantially, such that current damage dealt by dungeon monsters has the described effect as above. This method is preferable. Until an adjustment like this is made, in all likelihood tanks will remain devalued as no buffer is needed between the typical player and group content monsters due to the incredible healing capacity of healers (or even DDs with a heal skill slotted).

PvP is the bridge between the TTK and TTH of players. Because of PvP, players’ dps must be a fraction of their health pool (at least in PvP). In PvE, players have been known to exceed 40k dps. Various modifiers, most in the form of the Battle Spirit effect in Cyrodiil and duels, are used to drop that down below 10k, a range where average damage loses its significance and burst becomes a major component. TTK and TTH offer less insight into the player vs player balance because of this, but more on that later. Suffice to say, for now, if TTH is increased—meaning healing power is decreased—to balance dungeons, players’ TTK will also have to be increased—meaning damage capabilities are decreased—to balance things out. A simple change to PvE healing affects PvP balance, which then affects PvE damage as PvP is returned to equilibrium. This can also apply in the reverse direction. This connection between PvE and PvP is justification enough to say that PvE is so hard to balance because of PvP. There is equal justification to say that PvP is so hard to balance because of PvE.

Thus, because healing merits a decrease, damage merits a decrease as well. However, once again using PvP as the bridge between PvE TTK and TTH, damage merits a decrease, and thus so does healing. [End of Current Edition]

Most Recent Edit: Corrected the less than 1 and greater than 1 values for c, adjusted wording in 10th paragraph.