Dark Knight Megathread

[Shurrikhan]

It may not be the same thing as he might be thinking, but healing based on damage received has the potential to be so insanely broken (Read: Macrocosmos in P3s) or literally useless compared to current Oblation (Read: any single mechanic that sets your HP to low values, bypassing damage, like Charybdis -> double attack in O4s). P3s has already shown healing based on damage taken has incredible cheese potential. And with how cheesy TBN can be with ignoring damage downs, the last thing DRK needs is another shred of cheese.

Again, unless the mitigation would prevent a death, there's no difference in sustain produced between mitigation and healing. Though, when there is such a danger, mitigation is the better of the two. As such, even at best, Woe-eater is still inferior to a 4s 20% mitigation effect. But, we treat even 4s of 35% mitigation on a 25s CD (with attached healing and a further 4s of mitigation) as far from broken. This is mere 4s of 20%, but without even a max eHP increase, on a 40s CD.

At no point, though, is a 20% sustain benefit for/from 4 seconds (on a 40s CD) going to be "useless" compared to 10% for 10 seconds (on a 60s CD). The latter could, if the damage were somehow perfectly constant, provide a quarter more, but it'd also take a third longer to charge.

If DRK is to get any self sustain to match Aurora/Equil, I'd rather it be like them and useful in 100% of situations, not tied to external factors which cause its effect to wildly vary in use. Not everything needs to have the wheel reinvented.

I feel like things are getting compared oddly here. The answer to Aurora/Equil would in this case be heals from Bloodspiller, Quietus, and, say, Salt and Darkness -- bankable, dependable, purely stat-based. Oblation here simply acts like any other 4s window of increased sustain as provided via any other lv82 trait. It's very specifically not reinventing the wheel. It just also isn't giving excessive max eHP increases.

For people who doesn't understand what he means.

P3S does a mechanic called Death Toll + Life Agonies. Requires the Sunbirds to strike you to remove debuff (By dying like Neo-Ex Death), the debuff leaves a hefty Healing down debuff calculating around 50%, and Macrocosmos is % based than potency which means it ignores the debuff fully restoring everyone's HP in an instant. Whm/Sge/Sch have to AOE heal spam just to get everyone up. Life Agonies is a ultra raid wide AOE.

I'm aware, but this seems to fall more under the question of "Should we allow cheese? And to what degree?" SE could easily have Macrocosmos, too, check for healing debuffs before applying its effect, just as they could with any damage->healing effect. I don't think we ought to remove whole slews of possible abilities on the assumption that there might one day be a fight in which they present a unique advantage and SE will never adjust those skills' interactions even if they make a particular job a lock-in for the given fight.
____________________________

But if we're talking wish lists, I'd rather see each tank receive a central theme for defense, and then build their cooldowns around that. It really isn't that hard to keep them numerically equal, so long as we don't do anything stupid like target scaling flat healing, as a random example of course.

Same. Though, is it awful that I really want to build PLD around Cover as a broader, more integral mechanic? Not exclusively, but at least in part....
____________________________

Thinking on it, they should probably remove TBN from the MP cost and remove its free Edge/Flood. Instead, make it so TBN has an HoT on it, and shattering it burst heals you for the remaining potency. This changes the move into something you don't care whether it pops or not, you're getting healed by it the same regardless.

Why would we want to remove even more gameplay considerations / skill ceiling elements from DRK, though? Even the other on-demands wouldn't be nearly so lenient in optimizing them. And that's before accounting for any overheal issues or wasted capacity in using it externally.