Tank IDENTITY

[Lyth]

...

In HW, PLD/DRK was one tank slot and WAR was the other, much like WHM/AST was one healer slot and SCH was the other. It wasn't a deliberate design decision, but it happened as a consequence of the fact that ARR had so few tanks and healers. If you contrast this with DPS, you never really saw the same problem because the job pool was fairly diverse from the outset, such each job had a relatively narrow but clearly defined theme.

By the way, Steel Cyclone didn't have any self-healing on Stormblood's launch. You can see that if you pull up an older version of the job actions page circa 2017. That was only added later following complaints over Bloodbath's removal from WAR.

ARR's mistake was to try to set up PLD as a 'defensive tank' and WAR as an 'offensive tank'. Tanks need to be able to mitigate, do damage, obtain enmity, position and move, and you could argue self-sustain fits in there as an optional fifth category. You can have a small degree of variability, but large discrepancies lead to power imbalances. The way around this is that you just find alternative ways of achieving the end result. We saw this to an extent with Shield Oath's %DR vs. Defiance's %HP boost + %healing bonus, or even in Plunge vs. Tempered Will as anti-knockback tools. I don't think that there's any category of tank actions that 'has to be' done one particular way.

And that's where self-sustain comes in. I can think of many ways of doing self-sustain on a tank. PLD's classic spellcasting approach with Clemency has its issues because of the opportunity costs in cast time and resource costs, which makes it a dps loss. But that wouldn't be an issue if you had a way of generating a swiftcast proc that lets you cast healing and defensive spells instantly and for free. You wouldn't need a lifesteal effect built into Holy Spirit/Circle. Need some spot healing? Just weave in a Swiftcast Clemency. Done.

GNB has Aurora, which isn't all that interesting on its own. But what if you tapped into that 'soldier' design a bit more and gave them a Stimpack effect to boost their rate of regeneration along with other temporary bonuses? I've mentioned before that a lifesteal or the more lawful good approach of 'heal yourself on hitting the enemy (wink wink)' job like WAR could expand that same concept into mitigation by turning overhealing into temporary HP. And there are other unexplored options as well. Could you heal yourself on taking damage? Rewind your HP back to a previous timepoint? Could you convert unused shields into healing? Could you airdrop some health pickups for your party to grab? Could you build a heal-over-time effect into a stance or aura? Could you place down a support puppet that acts as a personal healer?

You can basically do this with any of the tanking 'categories' that I mentioned earlier. But it does require buy in from the community. If your tank with stimpack and regen mechanics turns around and starts demanding lifesteal as well as cure spells as well as HP rewind effects, then you end up with a bunch of identical tanks that all have way too much self-healing and defensive bloat.

[fulminating]

And that's where self-sustain comes in. I can think of many ways of doing self-sustain on a tank. PLD's classic spellcasting approach with Clemency has its issues because of the opportunity costs in cast time and resource costs, which makes it a dps loss. But that wouldn't be an issue if you had a way of generating a swiftcast proc that lets you cast healing and defensive spells instantly and for free. You wouldn't need a lifesteal effect built into Holy Spirit/Circle. Need some spot healing? Just weave in a Swiftcast Clemency. Done.
.

Clemency is a spell, no matter how hard you swiftcast it you'll still not be doing weaponskill/spell damage that gcd. Changing clemency to an ability probably wouldn't help much either because you'd now have a 1000/500p heal on demand which generates the opposite issue.

[Shurrikhan]

In HW, PLD/DRK was one tank slot and WAR was the other, much like WHM/AST was one healer slot and SCH was the other. It wasn't a deliberate design decision, but it happened as a consequence of the fact that ARR had so few tanks and healers.

I still feel like this was less a problem of what (or how little) jobs existed in each role prior as just... honestly taking the "if it ain't broke, don't fix it mentality" too far, ignoring opportunities for fleshing out their respective new job's (AST and DRK's) themes.

We had an incoming time-space support mage... and the only things we could think of for its core healing bonus was "Give it the choice between WHM mode or SCH mode"??? That theme could easily have offered either, in practice, among increased total output or increased eHP without having to simply copy HoTs and barriers verbatim.

ARR's mistake was to try to set up PLD as a 'defensive tank' and WAR as an 'offensive tank'.

Given that PLD had greater OT damage than WAR did and WAR had defensive-only mechanics that PLD had no equivalent to, I can't quite agree that the distinction was quite so clear-cut as that, but otherwise agreed.

PLD's classic spellcasting approach with Clemency has its issues because of the opportunity costs in cast time and resource costs, which makes it a dps loss.

To be fair, though, it was never sold as "self-sustain", only as an emergency support tool. The idea before was that Paladin's unreliable extra mitigation over time from blocking would take up that longer-term sustain* function.

But, yes, at least in the current state of things, I certainly wouldn't mind procced "Rapid Clemency" hotswapped button whereby Clemency becomes an ability for its next cast. I'd think that more fitting, at the least, than Knight's Benediction or Requiescat's healing-per-attack.

* To be clear, I'm referring to sustain as anything that increases your longevity, which can be via mitigation (flat or percentile) or healing (likewise flat or percentile, though we don't technically have the latter in XIV).

GNB has Aurora, which isn't all that interesting on its own. But what if you tapped into that 'soldier' design a bit more and gave them a Stimpack effect to boost their rate of regeneration along with other temporary bonuses?

...Hell yeah. Would love to see that theme played more with. I don't think we're going to get an outright Soldier at this point, so I don't think it'd be too problematic for its thematic elements to be split each across GNB and DRK.

You never want to design an action that always punishes you when using it, which was the problem with many tank actions pre-Shadowbringers. It ends up being dead hotbar space when you're playing optimally.

Ehhh. Any Raise, for example, is also that. As is, eventually... nearly every GCD heal. Since Clemency could feasibly save yourself from a death (far more dps-costly than a mere lost GCD and combo progress) or one's raid from a wipe, it certainly "always" punishing, so long as one doesn't narrow their sight to just the one player's offensive output.

Personally, I don't mind a job that is designed to have "saves the raid pitch-hitter buttons" having situational tools, so long as they come at little to no cost to that job's budget. Until ShB's oversights, for instance, I don't think Clemency was every considered part of PLD's sustain (combined effective mitigation and effective healing) budget, and so that was fine. It was a bonus tool that gave the job some unique affordances and further flavor.

[Shurrikhan]

I think what sometimes gets lost in the conversation about mechanics and what themes they may imply is their effect on gameplay through their affordances and differences from other means of reaching a roughly equal larger-scale effect (longterm sustain, on-demand eHP, etc.).

For instance, layered flat and percentile mitigation have multiplicative synergy, which can make you feel a bit more bundled up, dependent on coordination/ability-sync, strictly timed, or hard-but-brittle. (To me, that feels rather... Paladin.)

Or, take the difference between temporary HP and healing. The first is more timing-dependent, but the second is susceptible to overcapping. Extend them over time and the first makes the actual HP bar seem to move very little, even if one's max HP may hit amusingly extreme heights, while the second is likely far more in flux. Though when perfectly balanced, both will likely die around the same time, the first makes much more obvious that its life is finite --slowly spiraling downward-- while the second can at least give impression to the contrary --so long as one doesn't notice that, after each wave of healing, their HP has still bounced back up to a little less than before as their enemies chip them down. That difference in impression can make the second feel a little freer, a little more independent, but arguably the first can also feel more urgent and intense (since any of that temp HP that isn't spent before its duration elapses would go to waste). (For that reason, both are really damn good fits for Warrior, imo, so long oversights are avoided and the two can be truly balanced.)

Per-hit effects over a duration, instead of to a capped number of hits? It can certainly be more frustrating, especially if there's nothing extra to differentiate it from the capped version apart from obligatory SkS tiers, but... assuming it has that decent context? It rewards speed; it makes you feel more all-in; it makes timing the effect's CD, especially if bankable, feel more deliberate. (To me, that feels very GNB / blitzkrieg bodyguard.)