Big pulls with Dark Knight.

[Apdemic]

I don't see how Paladin has the worst defense outside of Clemency, but I'm glad to finally be getting responses that address the actual topic instead of just pointlessly flexing.

Take Clemency out of the equation and, at least compared to DRK; Paladin still has better passive mitigation, the best Invulnerability in the game (compared to DRK having the worst) and Sheltron you can basically spam with no downside that always lasts it's full duration (while TBN is a potential DPS loss where your shield might get popped in 2 seconds). DRK basically only has TBN, which is obviously great, some very negligible passive self healing and AD which is only really useful on big pulls. I'd say it balances out pretty well. Clemency is basically a bonus.

I'm not as familiar with the other 2 tanking classes, but from what I've gathered GNB is considered to have the worst defense of all the tanks.

Paladin has the worst defense without Clemency. Adressing the things you said:

passive mitigation: correct me if I'm wrong on the numbers but I'm pretty sure Paladins shield gives a 20% chance to block for 20% damage reduction. That's a passive mitigation of 4%. DRK has passive self heal to compensate, so does GNB, never played Warrior. The migigation is propably still better than that passive self heal but the difference is neglectable.

Sheltron vs TBN: TBN is so much more powerful, it's not even comparable. Sheltron gives you 16% damage reduction (20% from 100% block rate minus the 4% you already had from blocks) while TBN gives a 40k shield on roughly the same cooldown. Come on, it's not even close.

best Invul: fair enough, Paladin wins on that one but then again, it's on a 7 minute cooldown.

DRK only has TBN: doesn't DRK have access to the exact same defensive cooldowns as Paladin? Both have Rampart, Shadow Wall is the same as Sentinel, TBN is better than Sheltron and that's pretty much it already for Paladin (again, not counting Clemency). DRK also gets AD on top of that.

AD is only useful on big pulls: isn't that pretty much the entire point of your discussion though? You said DRK is crap on big pulls compared to Paladin because it doesn't have Clemency. DRK has AD instead and it's amazing on big pulls.

GNB is considered the worst in defense: nah Paladin is worse, although not by much (again, not counting Clemency). The passive self heal should be pretty close to Paladins 4% passive mitigation. GNB has pretty much the same defensive cooldowns (Sheltron being a bit stronger than HoS due to its lower Cooldown) and then gets two additional cooldowns on top of that for free, although they are both not that strong (Camouflage and Aurora, which practically functions the same as another mitigation tool). Paladin can throw in Divine Veil, GNB throws in Heart of Light. Hollowed Ground is the only upside here and even then, Superbolide isn't that far behind it, in my opinion the strongest invul after HG.

[Goji1639]

Well, DRKs self heal is only on it's single target combo, not sure if GNB is the same. Whereas passive mitigation is working at all times, and the more damage you receive the more powerful mitigation is. I'd say it's far superior to the self heal, but someone would have to actually math out a variety of scenarios to confirm that.

TBN is more powerful overall, but in big pulls I feel like it gets diminished considerably. It's good when paired with certain CDs, but otherwise it gets eaten in seconds. Sheltron is active the entire 6 seconds no matter what, and like I said before, the more damage coming in the more powerful mitigation is. TBNs shield, however, doesn't scale with the amount of damage you're taking.

Also, AD is better with big pulls. You implied Paladin has the worst defense overall, though, so I kind of shifted perspective to overall. AD is only really a good defensive ability on substantial pulls.

I feel like when being played optimally all of the tanks are pretty well balanced between defensive and offensive capability. Paladins just have an incredibly powerful utility that allows them to sacrifice damage in order to become by far the most resilient tank in the game.

[ReiMakoto]

Well, DRKs self heal is only on it's single target combo, not sure if GNB is the same. Whereas passive mitigation is working at all times, and the more damage you receive the more powerful mitigation is. I'd say it's far superior to the self heal, but someone would have to actually math out a variety of scenarios to confirm that.

TBN is more powerful overall, but in big pulls I feel like it gets diminished considerably. It's good when paired with certain CDs, but otherwise it gets eaten in seconds. Sheltron is active the entire 6 seconds no matter what, and like I said before, the more damage coming in the more powerful mitigation is. TBNs shield, however, doesn't scale with the amount of damage you're taking.

Also, AD is better with big pulls. You implied Paladin has the worst defense overall, though, so I kind of shifted perspective to overall. AD is only really a good defensive ability on substantial pulls.

I feel like when being played optimally all of the tanks are pretty well balanced between defensive and offensive capability. Paladins just have an incredibly powerful utility that allows them to sacrifice damage in order to become by far the most resilient tank in the game.

PLD has no fluff cooldown (thrill, camo, to an extent dark mind but drks tbn strength sort of excuses a need for a fluff) and only has passive block+shelltron with no sustain outside of clem. Passive block and shelltron don't compare to to the self sustain and additional cooldown power of the other tanks. In order for shelltron to beat out blackest night you would literally need to take more than your entire healthbar in the 6 seconds that shelltron is active, and shelltron actively makes and cooldowns you did use worse as a result of diminishing buff returns, whereas blackest night and cds complement each other. Even compared to GNB pld has it rough, GNB gets access to an extra fluff cd in camoflage which is excellent on large pulls, and they get auroa, in dungeons their mitigation is roughly equal if anything, but in single target pld has it the worst.

[Goji1639]

PLD has no fluff cooldown and only has passive block+shelltron. Passive block and shelltron don't compare to to the self sustain and additional cooldown power of the other tanks. In order for shelltron to beat out blackest night you would literally need to take more than your entire healthbar in the 6 seconds that shelltron is active, and shelltron actively makes and cooldowns you did use worse as a result of diminishing buff returns, whereas blackest night and cds complement each other. Even compared to GNB pld has it rough, GNB gets access to an extra fluff cd in camoflage which is excellent on large pulls, and they get auroa, in dungeons their mitigation is roughly equal if anything, but in single target pld has it the worst.

This is the kind of information I'm interested in. At what point does diminishing returns kick in for damage mitigation? How exactly does Sheltron stack with your other CDs... is it additive or multiplicative? If mitigation really does taper off that significantly then maybe I have the wrong idea.

[Kabooa]

This is the kind of information I'm interested in. At what point does diminishing returns kick in for damage mitigation? How exactly does Sheltron stack with your other CDs... is it additive or multiplicative? If mitigation really does taper off that significantly then maybe I have the wrong idea.

There are no diminishing returns.

Multiplicative mitigation is as effective for every layer.

People are misconstruing what diminishing returns actually means. The amount you -need- to use at any give time does not reduce the effectiveness of further cooldowns. Using more than you need is ill advised, but you're still going to be a damn super tank if you pop it all at once.

You are always going to be 20% more durable when popping Rampart no matter what else is popped at the time. Multiplicative mitigation keeps everything equally effective unlike additive mitigation which makes every additional source exponentially more effective.

As a tank, your baseline (100%) is your HP x 1.25 (Tank Mastery) x 1.42-1.72 (Based on gear). This is your baseline, your 100%.

20% mitigation is 25% EHP. Several things have quirks to consider but this is the basics.
30% is 144%
40% is 172% etc.

You can get these numbers in the following manner: 100 / Mit, where 100 is your baseline and Mit is your current amount of damage you are taking. With 20% running, you are taking 80% damage. With 30%, you are taking 70% damage. With both 20% and 30% running, you are taking (.8 x .7 = .56) or 56% damage.

100 / .8 = 125
100 / .7 = 142%
100 / .56 = 178%

Now that our basic rundown is done, we'll get into more specifics.

Parries reduce incoming damage by 15%
Blocks reduce incoming damage by 20% This was nerfed as paladins used to scale upwards of 28-30%, which for obvious reasons couldn't happen with the lovely change to Shelltron in Shb.

Parries and Blocks cannot happen if the enemy critically strikes you.

Parries and Blocks cannot occur at the same time (in the case of Paladin).

Parries and Blocks cannot happen if you are casting a spell (paladin).

Aurora is garbage, and only excels at making giant pulls effortless through the global threat it generates.

[shao32]

I mean, by this logic you may as well take away all healers GCD heals and just let the group wipe if a healer can't keep it alive by weaving it's OCD heals into it's optimal DPS rotation. Why let healers constantly protect bad groups from their failures?

I think having the option to counterbalance healing/tanking and DPS output based on your groups capabilities makes both roles more engaging, and feels more dynamic.

That's not even a fair comparison, GCD heals are still need it on the game and outside of the lvl 80 dungeons, normal raids and other casual stuff where a skillfull healer can stand by using his oGCD only, the GCD are necesary in extreme, savage and ultimates + leveling dungeon being harder due the ilvl limitations making it being imposible to keep your party with only oGCD if you are not properly geared and fully know your job.
Despite of the amount of complains that healing is braindead easy right now shields and regens are still necesary.

Clemency on the other hand is not taken in consideration for anything, no savage, no ultimate, just nothing, it's not need on optimal gameplay for anything and the amount of situations where you can use it to save someone can be counted with one hand and i consider it encourage bad habits on several PLD casuals, thats why is consider the last resort skill and we always recomend never use it outside of those extreme rare situations where you can really make a diference by using it bcs thats how limited and useless is the utility of the skill and how is mining proper progresion of new PLDs who depend of it to much when they shouln't.

so no, PLD can keep clemency like now or they can recive an update where they can use the skill in more situations and not costing they dps output but i don't consider give the other tanks a skill to carry bad players in those rare bad runs and being useless if you are average to good tank a good thing when you can archive the same with being better and having a fully usefull kit that works on every skill level, apart is part of PLD identity and we have to much homogenization already.

[ReiMakoto]

There are no diminishing returns.

Multiplicative mitigation is as effective for every layer.

People are misconstruing what diminishing returns actually means. The amount you -need- to use at any give time does not reduce the effectiveness of further cooldowns. Using more than you need is ill advised, but you're still going to be a damn super tank if you pop it all at once.

You are always going to be 20% more durable when popping Rampart no matter what else is popped at the time. Multiplicative mitigation keeps everything equally effective unlike additive mitigation which makes every additional source exponentially more effective.

As a tank, your baseline (100%) is your HP x 1.25 (Tank Mastery) x 1.42-1.72 (Based on gear). This is your baseline, your 100%.

20% mitigation is 25% EHP. Several things have quirks to consider but this is the basics.
30% is 144%
40% is 172% etc.

You can get these numbers in the following manner: 100 / Mit, where 100 is your baseline and Mit is your current amount of damage you are taking. With 20% running, you are taking 80% damage. With 30%, you are taking 70% damage. With both 20% and 30% running, you are taking (.8 x .7 = .56) or 56% damage.

100 / .8 = 125
100 / .7 = 142%
100 / .56 = 178%

Now that our basic rundown is done, we'll get into more specifics.

Parries reduce incoming damage by 15%
Blocks reduce incoming damage by 20% This was nerfed as paladins used to scale upwards of 28-30%, which for obvious reasons couldn't happen with the lovely change to Shelltron in Shb.

Parries and Blocks cannot happen if the enemy critically strikes you.

Parries and Blocks cannot occur at the same time (in the case of Paladin).

Parries and Blocks cannot happen if you are casting a spell (paladin).

Aurora is garbage, and only excels at making giant pulls effortless through the global threat it generates.

It literally is the definition of diminishing returns.

As a vastly simplified example

Scenario A): You pop rampart and sentinel at the same time you will get: ( 5000 × 15 × 0.8 × 0.7 ) + ( 5000 × 5 × 0.8 ) + ( 5000 × 20 ) =‬ 162000 damage recieved
Scenario B): You pop rampart and sentinel staggered you will get: ( 5000 × 20 × 0.8 ) + ( 5000 × 15 × 0.7 ) + ( 5000 × 5 ) =‬ 157500 damage recieved

in otherwords you have recieved diminishing returns by popping more mitigation, which does not apply to a drks blackest night because it is a barrier not % mitigation. The gap between the numbers only grows the more you add, hence diminishing returns, and why its better to spread you cds out as appropriate.

Also Aurora heals the same (if not a slight bit more due to being a regen but only slightly) as wars equilibrium, free healing is free healing and on a 1 min cd it isn't anything to shake a stick at, as its basically 30k+ free healing every 60s provided your healer isn't overzealous.

[Kabooa]

It literally is the definition of diminishing returns.

As a vastly simplified example

Scenario A): You pop rampart and sentinel at the same time you will get: ( 5000 × 15 × 0.8 × 0.7 ) + ( 5000 × 5 × 0.8 ) + ( 5000 × 20 ) =‬ 162000 damage recieved
Scenario B): You pop rampart and sentinel staggered you will get: ( 5000 × 20 × 0.8 ) + ( 5000 × 15 × 0.7 ) + ( 5000 × 5 ) =‬ 157500 damage recieved

in otherwords you have recieved diminishing returns by popping more mitigation, which does not apply to a drks blackest night because it is a barrier not % mitigation. The gap between the numbers only grows the more you add, hence diminishing returns, and why its better to spread you cds out as appropriate.

Also Aurora heals the same (if not a slight bit more due to being a regen but only slightly) as wars equilibrium, free healing is free healing and on a 1 min cd it isn't anything to shake a stick at, as its basically 30k+ free healing every 60s provided your healer isn't overzealous.

Lets say you're a tank with 170,000 HP. You're faced with an incoming hit of 280,000. You have no Invlun because they're terrible design and trivialize things. In order to survive this hit, you need an EHP of 280,0001. For sake of ease we'll say you will take no further damage afterwards so you won't die to an instant auto after.

You pop your 30%. 100 / .7 = 143% (rounded up). 243,100. Not enough.
You pop your 20%. 100 / .56 = 179% (rounded up). 304,000. Bingo.

How much of an increase from 243,100 is 304,000? 25%.

No decrease. It did not diminish. It was the same ratio of increase as if you used it alone.

You are confusing need with DR. In the tank buster scenario, not stacking them is the "DR" scenario going by your "how much damage did I take" metric. You cannot have diminishing returns both when stacking and when not stacking with this mindset, but it's fortunate because it doesn't deal with either.

It's about the cost against the gain. The cost is the use of a cooldown, the gain is EHP.

Except invulns. Invulns are dumb.

Edit: Aurora is still garbage.

[Tint]

I admit I don't understand EHP.

What I think to understand regarding the diminishing returns is the following:

If you get hit by 100k damage and you use your 30% cooldown the damage is reduced by 30k so you get hit by 70k.
If you use the 20% cooldown the damage is reduced by 20k so you get hit by 80k.

If you use both together the damage is reduced first by 30% down to 70k and when you now use the 20% cooldown it's further reduced to 56k. So the 20% cooldown reduces 20k damage if you use it alone, but only 14k damage if you use it together with the 30% cooldown.

Please ELI5 it for me xD

[Kabooa]

Please ELI5 it for me xD

EHP stands for Effective Health/Hit Points.

It is how much "Raw" damage has to go out to kill you.

Keep in mind you almost never see "Raw" damage beyond very early in the game, as by the time you are level capped, based on your job, you have a certain level of defense, and in the case of tanks, a trait. Your "HP" at the bottom, the green bar, is not your "EHP". Tank "HP" goes far beyond what equal amounts of DPS "HP" does.

So lets say you're a black mage and you have 100,000 HP. Your defense score at level 80 end game armor will put you at approximately 18-20% mitigation. This means you take 80% damage, or .8. You find your effective hp by the following.

HP / Damage Taken. 100,000 / .8 = 125,000 EHP.

However, since we never don't have our defense, this is considered our "Baseline". Our 100%. The damage we see when we get hit, for example, 7500, is already taking our baseline mitigation into account, and is not "Raw damage". For all practical purposes however, it's 100,000 health vs 7,500 damage.

Your understanding of the end result of stacking mitigation is correct.

However this is not "Diminishing return". Diminishing return deals directly with Cost and Benefit. The End result is not the "Benefit", because the end result has more factors.

The benefit of more mitigation is higher EHP values.

In order for mitigation to suffer from diminishing returns, each additional layer needs to be penalized in its benefit, but it is not.

With 20% mitigation, 100,000 HP becomes 125,000 EHP. A 25% increase.
With 30% mitigation, 100,000 HP becomes ~143,000 EHP. An ~43% increase.
With both layered together, 100,000 HP becomes (100,000) / (.8 * .7 = .56) = 178,571, or about a 78.6% increase.

This effectively means that Every layer of mitigation in a Multplicative System is exactly as effective as it would be alone compared to if they are used together.

The above example provided by Rei is flawed because it paints a very specific scenario, despite being vaguely described.

"The tank is taking 5k DPS over 40 seconds, all damage intake is uniform".

My counter example is no less specific.

"The tank is taking 280k Damage over one attack"

This demonstrates that "Diminishing return" is not about how much damage is prevented. That is separate, and we can go back and forth on the validity of each example, but in the end it has nothing to do with the concept.

It's about Cost (The ability) and Benefit (The EHP increase), and in multiplicative systems, the Cost and the Benefit do not fall out of sync.

This does not mean you cannot use abilities unwisely. You still gain the benefit - But you just didn't need it.