Big pulls with Dark Knight.

[Tint]

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

:O

I really understand it now, thanks. I had to read it twice, but this sentence really is the key.

I was wondering why 20% mitigation is a 25% increase of EHP. But 125k x 0.8 really is 100k. So while the mitigated damage of each cooldown really is reduced with every layer, it does not matter because the EHP value is increased by this lost mitigation? o.ô;

100k / 0.8 = 125k. An increase of 25k EHP
100k / 0.7 = 143k. An increase of 43k EHP

125k / 0.7 = 179k. An increase of 54k EHP :O

So while I lose 6k mitigation, I gain 11k EHP by stacking them?

Math is really complicated. And I am actually still not sure if I really understood it >.>

[Kabooa]

Math is really complicated. And I am actually still not sure if I really understood it >.>

The long and the short of it is math is cool guys, I swear ;_;

The short version is match your defense to what's hitting you.

The long version:

You have 100 HP. Let's say you have 50% mitigation. Run our formula.

100 / .5 = 200.

At 50% mitigation we have 200 EHP, or 100% more.
If we are hit for 100 raw damage, it is reduced to 50 damage.

Lets add a second layer of 50% mitigation. So our formula is

200 / .5 = 400. We have 400 EHP, or 100% more than before.
If we are hit for 100 raw damage, it is reduced first to 50, then to 25.

So the first time we reduced it by 50. Neat. But the second time it was only 25. What gives?

However, by taking 25 damage, we have to be hit four times to total our 100 HP. Our 100 HP against 100 Raw Damage is acting as if it was 400 HP against 100 raw damage. Our effective HP.

If we didn't have the second layer, then we would take 50 damage, and only need to be hit two times to total our 100 HP.

Now, what Rei was saying is that you get less benefit out of stacking mitigation. Rei is an experienced tank. He's talking less about the theory and more about the practical application. Listen to him, he good. (All of this is mostly the math and theory behind everything, and if you found it interesting - cool. Math is cool ;_; )

Because if you pop everything, you're a god. For 15 seconds. You'll have 123908757 EHP. The measly 100 raw damage hitting you is nothing. NOTHING.

Then everything runs out. And you're back to 100 HP, assuming you're full. Suddenly the 100 raw damage is much scarier.

But pop too little, and you pop.

Effective tanking is matching defense to the enemy offense.

For large packs of enemy, you're generally aiming for coverage. You never want your 100 HP to only be 100 HP.

For bosses, you're targeting specific actions. Your 100 HP is usually sufficient until a big hit's coming out, and at least two layers is recommended. Commonly the "Short" with one "Long", such as The Blackest Night and Rampart or Shadowskin.

Or use an invulnerability, because why not.

[ReiMakoto]

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.

We are not talking about a need situation though, we're talking about dungeon pulls in this thread, its common sense that you need to stack cooldowns to survive a large hit, however we are talking about prolonged damage within a dungeon context, so you will be getting diminishing returns on your stacking of cooldowns, as its about taking the least damage possible, which you achieve by spreading out your cooldowns. So in this case using your cooldowns spread did decrease the total amount of damage recieved, hence diminishing returns on stacking cooldowns. If you absolutely needed to stack the cooldowns to live it would not be a diminishing return, as you would be dead, but we are talking about dungeons here. Thats why my example wasnt flawed as I was talking within the context of the thread and the discussion being had, if we were talking about surviving tank busters would absolutely be talking about stacking cooldowns.

I think that broadly we are talking about different definitions here. Im taking diminishing returns to be that you put more in at one point and got less total out of it than using less spread out, this applies to both dungeons and some of the good raid design where tanks are taking more frequent high damage. Youre taking a purely mathematical view of diminishing returns, which while valid does not adhere to the reality of the game design due to to how spikey damage is (popping more than necessary for a tank bust essentially wastes mitigation that could be applied elsewhere ect).

Ps. Auroa isnt bad fite me

[ReiMakoto]

Also while I'm here and awake @OP heres some examples as to why TBN is superior to shelltron and to the other short cd tank skills again using a simplified comparison.

Like before 2 scenarios, a pld and a drk, both taking 10000dps for 35 seconds, who spread out a rampart and a sentinel/shadow wall, and use a blackest night/shelltron paired with each rampart/sentinel. They both have 160000 hp total

PLD: ((10000*0.8*0.8*6)+(10000*0.8*14))+((10000*0.7*0.8*6)+(10000*0.7*9))= 247000 damage taken
DRK: (10000*0.8*20))+(10000*0.7*15)-80000(2 tbns at 40000 used their full value )= 185000 damage taken

Even if we double the damage intake to 20000 DRK still takes 44000 damage less than pld. I don't have dungeon numbers to hand but I don't think the damage scales high enough for shelltron to outweigh tbn (it'd need to be 40k+ damage per second uniterrupted for shelltron to overtake tbn). Now, PLD does have its passive block rate to consider as well, but drk also has abyssal drain and dark mind (on pulls with casters). tldr even if your tbn is being shredded it will still be providing more mitigation overall than the other short cd tank mitigations which makes DRK very powerful in large pulls. Obviously this example is simplfied but I hope it demonstrates to you OP how crazy tanky DRK actually is because even in the second boss-> third boss pull of anemesis I would doubt that you would be taking 40k per second uniterrupted damage for 40+ seconds.

[Kabooa]

I think that broadly we are talking about different definitions here.

Ps. Auroa isnt bad fite me

You are talking about mismanagement of a toolkit. That isn't Diminishing returns. In order for it to be diminishing returns, every scenario must have multiple layers being weaker, not just the ones made for it. Cut off 10 seconds of your scenario and see where that ends up. Add another 30 and look. The result will vary by scenario, which is the entire reason I chose "Tankbuster". The most effective use of cooldowns changes, but for a system to suffer from diminishing returns, it must be built in to every scenario.

You need only look at where Crowd Control works to see where DR is uncompromising.

It does not care about encounter length, damage intake, or any other factor.

1 use = 100%
2 use = 50%
3 use = 25%

For the same cost, you get less benefit. That's diminishing return. You don't have to invent a scenario to prove it does or does not suffer from it.

But no one crowd controls, therefore it's always going to be at 100% effectiveness if you did use it.

Does that mean crowd control does not have diminishing returns?

PS. Aurora is bad.

PPS. Except white mages. They CC the !@#% out of everything.

PPPS. And I totally acknowledge this doesn't matter in the slightest. Which is why I persevere.

[ReiMakoto]

You are talking about mismanagement of a toolkit. That isn't Diminishing returns. In order for it to be diminishing returns, every scenario must have multiple layers being weaker, not just the ones made for it. Cut off 10 seconds of your scenario and see where that ends up. Add another 30 and look. The result will vary by scenario, which is the entire reason I chose "Tankbuster". The most effective use of cooldowns changes, but for a system to suffer from diminishing returns, it must be built in to every scenario.

You need only look at where Crowd Control works to see where DR is uncompromising.

It does not care about encounter length, damage intake, or any other factor.

1 use = 100%
2 use = 50%
3 use = 25%

For the same cost, you get less benefit. That's diminishing return. You don't have to invent a scenario to prove it does or does not suffer from it.

But no one crowd controls, therefore it's always going to be at 100% effectiveness if you did use it.

Does that mean crowd control does not have diminishing returns?

PS. Aurora is bad.

PPS. Except white mages. They CC the !@#% out of everything.

PPPS. And I totally acknowledge this doesn't matter in the slightest. Which is why I persevere.

Im not sure what you mean, cutting 10 seconds off means that one of the resources wasn't entirely used making it not an equal comparison. As I said we are talking about the context of mass pulls here I've already admitted that if we are taking the context of a tank buster then it changes, but in the terms of mass pulls and various encounters I will still argue that it is a diminishing return, for the same input (2 cooldowns) you get less output (or in this case more damage recieved) If that is not a diminished return I'm not quite sure what we are arguing anymore. Even in a scenario where they both end at the same time (35s) not pressing both at the same time is still less damage taken (13700 vs 132500). Again I think this is coming down to a difference in definition, I'm saying it is a diminishing in this scenario as you have recieved less by using them at the same time, for the same cost (pressing 2 cooldowns) you have recieved less benefit (more damage recieved). It does not need to be a diminishing return in all circumstances I've admitted this, but withing the context we are talking about in this thread it is.

PS. Tell me why free healing with no cost is bad

PPS. Yes this doesn't matter but I feel like we are agreeing with each other mostly and are having an argument over semantics which is annoying

[Shougun]

You are talking about mismanagement of a toolkit..

Getting ready to lather butter on my own foot and eat it, since I've not read everything here. But as I was reading this diminishing return part I am thinking ReiMakoto is talking about how % are not additive they are multiplicative and they are explicitly talking about them in context of the damage numbers rather than the eHP, so stacking them is a diminishing return by definition of diminishing returns assuming you accept DR% as investment into the system and damage % reduced the output rather than how much did eHP change. Additive would of course make it not true in terms of damage reduced but thats not our system, and can be super OP lol, like some games make it once you break cap you gain health instead of reduce/lose XD.

This is an important difference because you're talking total power of a short period of time, important for busters, and they're talking about going from pack of monsters to pack of monsters over a longer period of time.

Or in other words you have 100 damage and two 50% damage reduction cooldowns. If you input their value of 50% separately they would both give you 50. If you input them together you get 50 and then 50% of 50 which is 25. So you have diminishing returns based on investment in terms of damage reduction. Same DR% input but less damage reduced ouput (one reduced 50 the other reduced 25, i.e. if you ran them separately you would have more damage reduced assuming the damage incoming stays the same at least).

You're talking about hp, so 100 goes to 150 and then 150 goes to 225 that last jump being 75 rather than 50 which is an increase in value. That said I'm not entirely certain you can calculate eHP that simply, as I'm pretty sure you should be calculating eHP in relationship to % damage taken. Something like: HP / % damage taken = eHP (I only mention this because I normally never do MMO math and so I just wanted to show it the far easier way of direct %s lol).

So the context of if you need it within that bracket of defense buff time or if you need to parse out the reduction over a long period of time is really important. But there does exist a situation/lens where there is diminishing returns.

Also I think Aurora isn't bad, has no casting time/cost, and in that because it adds over time if you were near cap and taking damage the chance of wasting health to cap would be a lot less than clemency near cap which would just be "why in the world did you just cure for 100 potency and toss the rest out the window?". Certainly bad if you need it all at once though, cause it takes a whole 18 seconds lol. A thought I had for it, though I don't really play GNB as a main (@80 tho) so I'd rather leave that to people who enjoy the job more (I lurv me Paladin), is that Aurora creates a shield of ~1200 potency for self and ~100 for everyone nearby- shield lasts for ~15 seconds, while the shield is active each tick heals you for y. Can play with the numbers to make it better/worse for balance, and probably bake the change around an expansion so other tank balances adjust too, but I feel like that would hit the vibe of an Aurora and GNB a bit better, Auroras being wonderful while they are there but a bit fickle and GNB with high tech shields. I imagine the visual could be a bit more tweaked to fit the name then too with him pumping out a chromatic dust cloud of defensive energy that shields his allies (with a focus on shielding himself/target).

[Kabooa]

Im not sure what you mean, cutting 10 seconds off means that one of the resources wasn't entirely used making it not an equal comparison.

PS. Tell me why free healing with no cost is bad

PPS. Yes this doesn't matter but I feel like we are agreeing with each other mostly and are having an argument over semantics which is annoying

Just as extending an encounter makes selected mitigation matter less. The common theme is that diminishing returns is not the defining factor here. The end result of "how much damage was taken" is primarily due to the scenario. Every example I can give you and you can give me can be countered by another equally valid example, because at the core the examples do not deal with Diminishing Return. They deal with effective cooldown use.

You have 100% coverage in a 30 second window yet mitigate less damage total. Extend it to a 60 second window and the % difference between the two sharply drop.

Of course in the 60 second window you can fit in another cooldown, and so we'd be stacking three in scenario A, and then account for-

You get the idea.

Multiplicative mitigation, the system itself, does not have diminishing returns. In order for it to have diminishing returns, every additional layer has to give less than advertised.

PS: Equilibrium is just the cooler Aurora, Clemency is miles better as a healing tool even factoring in the loss of a holy circle, and Abyssal Drain scales with pack size. Aurora should be a 600 Heal/Shield, with the shield transferable with HoS. /micdrop

So the context of if you need it within that bracket of defense buff time or if you need to parse out the reduction over a long period of time is really important. But there does exist a situation/lens where there is diminishing returns.

Hi.

Welcome to the Thunder Dome.

It's a matter of where you want to try and fit it. I do not subscribe to the end-result train of thought, because then it either is or isn't, and both are true. It's Schrodinger's Armor. It both is and isn't affected by Diminishing Returns until we open the box.

But the "End result" is determined by a multitude of factors. Even the vague example he provided is a very specific encounter, and it will change by tweaking values and adding or removing factors. It "works' in his example because the numbers fall that way. But it wouldn't work if the numbers fall a different way. The fact that Aurora covers the difference between them is a testament to how minor the difference it actually made is (lol Aurora). Subscribing to end result is muddying everything because why should we only factor in what makes it fit in to the box?

The presence of a white mage changes everything. Your DPS changes everything. Your party members change everything. Taking a half step to synchronize auto attack timers change everything. The encounter length. Dodges. Random parries. If we're going to cut out so many things just to keep it basic, then lets just keep it basic, yo.

And there is nothing more basic than just looking at what the system itself does, and the system itself is multiplicative mitigation, and multiplicative mitigation is designed to do two things.

1) Remove the potential to become invincible.
2) Keep every source of mitigation as strong together as it would be alone.

[ReiMakoto]

Just as extending an encounter makes selected mitigation matter less. The common theme is that diminishing returns is not the defining factor here. The end result of "how much damage was taken" is primarily due to the scenario. Every example I can give you and you can give me can be countered by another equally valid example, because at the core the examples do not deal with Diminishing Return. They deal with effective cooldown use.

You have 100% coverage in a 30 second window yet mitigate less damage total. Extend it to a 60 second window and the % difference between the two sharply drop.

Of course in the 60 second window you can fit in another cooldown, and so we'd be stacking three in scenario A, and then account for-

You get the idea.

Multiplicative mitigation, the system itself, does not have diminishing returns. In order for it to have diminishing returns, every additional layer has to give less than advertised.

PS: Equilibrium is just the cooler Aurora, Clemency is miles better as a healing tool even factoring in the loss of a holy circle, and Abyssal Drain scales with pack size. Aurora should be a 600 Heal/Shield, with the shield transferable with HoS. /micdrop

Im gonna end this here because Shogun seems to understand what i was getting at with this and different scenarios having or not having a diminishing return, and i dont think im gonna convince you, im saying the misue of the cooldown has caused a diminishing return by the definition of diminishing return you used earlier. If you extend to 60 seconds, rampart and sentinel wont have cooled down so the spread out scenario has still mitigated the most damage. But I digress, my point here was to show OP how blackest night interacts differently which i did earlier.

On Aurora, saying it could be better doesnt make it bad, equilibrium cant be targetted so its usless in an off tank scenario (n flash not withstanding), and im yet to raid with a paladin whos healed more with clem than I have with free aurora. That aurora buff you said sounds nice, doesnt make the current version bad, literally free heal is really good and can be used to great effect

[Kabooa]

Im gonna end this here

As you like.