Mograine's Notebook - DK Tanking Guide

[Deathelve]

could you say which specific trinkets you have?

Vial of Dragons Blood (956 mastery)
Relic of Niuzao (1434 stamina)
Zen Alchemist stone (809 mastery)

[Reniat]

But that's just it, healers (mostly) react to your health, which in turn reacts to the damage you just took.
So you already HAVE an idea of healer reaction. And by adding more Haste, you reduce the chance that your healers need to carry a damage spike on their own (similarly to how hit & expertise work for most tanks) because you can 'break' more strings of damage with DS.

By simply narrowing it down to "healers react to the damage you take" you are just going back to looking at dmg taken as opposed to the ideal of tank/healer stability, which is what we should be after. There are a myriad of factors to consider regarding healing reaction. Your burst size, previous bursts, raid wide damage, any number of mechanics, predicted damage, as well as their current mana can all affect how they react to that particular dmg intake. Each healer will react slightly different to each situation as well, so you'd have to include some incredibly large neuroscience functions into your healer simulator.

Haste also gains value if you can actually take advantage of the increased DS/min, which isn't always the case. You have to wait for burst damage to recover, so if the burst damage isn't coming too quickly (which is generally the case as if it is you should have a cooldown (or multiple cooldowns) active.) then you aren't gaining as much from the added haste unless you stop timing and just go for raw shield frequency, which just trades control for random mitigation, which isn't ideal.

More Mastery, on the other hand, only helps create the feeling of full-to-null situations avoidance currently give us where there is a large damage burst, then nothing, then burst again,... This is more difficult to heal as a healer, because the tank's required healing is less predictable.

It is only unpredictable if you are randomly placing death strikes as opposed to intelligently controlling your damage intake. This is also where cooldowns come in to play, as time you can't death strike is time you should be covered by some other form of damage control.

By relying on frequency over control, you are putting yourself at risk of being more unpredictable than with less ds/min but better placed mitigation through mastery.

[Reniat]

Vial of Dragons Blood (956 mastery)
Relic of Niuzao (1434 stamina)
Zen Alchemist stone (809 mastery)

Use the Vial and the Relic of Niuzao. That combo will get you by most anything. If you really want to maximize the use of the trinkets, track your buff from the Vial (it's called Elusive) and make sure you wait until that buff expires before you activate the on-use Niuzao trinket so that you're not stacking CDs. (and don't forget to actually use your on-use trinket )

[Tengenstein]

So how long do you recommend sitting on your DSs for? lets quantify what you're reacting to in order to squeeze off a DS, are we talking Hit strings or Breath attacks? I'm sitting here thinking that if i just sort of stagger my DSs about 4 seconds apart, I'm never going to take more than 2 unmitigated melee hits in a row from any given boss unless they get a haste buff, And if i know a huge hit is coming in I can still sit on a DS until after it hits, and furthermore becuase i'm now meleeing faster i'm going to get proportionally more SoB stacks over any given time frame, so my DS are going to heal more on average which in could actually be more important than the shield.

[Reniat]

Until you receive "burst" damage, which could be a single large hit or several quick hits. Anything that could be considered large fast damage that takes you from a safe level of health down to a low level of health very quickly, as this is when your healers need your help the most. You could space out your Death Strike evenly based on your current haste, but then you won't be able to immediately react to your health dipping without screwing up that interval, which will mean a larger interval later, and if you continue to react immediately to your health (like you should be) you've stopped using an interval to decide your death strikes.

If you could achieve a haste level high enough to where you could maintain such a high shield uptime that you never took a completely unmitigated hit then that would be fantastic, however i think you are overestimating the amount of rune regen you can get from haste and underestimating the shield amount you lose. Just to see how the numbers worked out, I went and with a build for max haste, putting it at the very top of the list and got a rune regen of 7.64s, which is only .69 seconds saved versus no haste at all (8.33s rune regen). Before that build I had 158.22% mastery, and after I had only 128.72% mastery. Now this will depend on your gear, but in my case I only got a shield lost to seconds gained ratio of 42.75% shield lost per 1 second of rune regen gained.

[Tengenstein]

But my point is that almost every big hit is well broad cast whether its Elegon's Breath, or an Overwhelming Assault or a three-stack exhale, you know when its coming, you don't need to sit very long at all because you have DBM timers for it and make it predictable and expected, and by frequent use of death strike you prevent yourself ever taking a succession of melee hits. Even Gara'jal with his hasted 1.2s swing timer you're gonna struggle to take 3 hits in a row without an absorb so your health isn't going to drop low without eating a boss spell nuke like Elegon's Breath, or an Overwhelming Assault or a three-stack exhale.

You're also ignoring that you will get more SoB stacks so your self healing will be stronger, especially when you delay you DS to counter a predicatable spell nuke, I don't know alot of healers that run with an absorb tracker especially ones who can see how big an absorb is on someone from some one else, I do know alot of healers whose raid frames show health bars and a tanks health going back up after a spell nuke is gonna make them feel alot more comfortable, than the size of an absorb shield they may not even see. the other advantage of frequent DS use when not countering spell nukes would be more RE/RC/BT procs/charges which means even more DSs making the likely hood of a hit string much less likely, and with BT you could still maintain intervaled DSs and use BT charges to allow you to bank DSs to be timed after spell nukes. getting the best of both worlds.

I'm not really disagreeing that mastery is not a great stat, i'm just not entirely sure i would put haste as "well below" in terms of survivability.

[Airowird]

By simply narrowing it down to "healers react to the damage you take" you are just going back to looking at dmg taken as opposed to the ideal of tank/healer stability, which is what we should be after.

No, it's the other way around. It works kinda like a weight-spring system: The more health you are missing, the more healing the healers feel you are in need of. Because an increase in healing output usually comes at the cost of efficiency, you could compare this to e.g. the metal strain in the spring which after a lot of bursts of a large enough force will shear and tear (=healers OOMing), making the weight (=your health) drop like a stone. Mathematicly, this is similar to a variating noise signal with a DC coupling you need to eliminate, or charging a laptop battery that needs to supply power in bursts at the same time, or a temperature regulation system in a chemical mixing vat, or ... Basicly, the math behind it has been used in billions of applications and it is so natural (both in the meaning of human perspective as in real physical nature), that it is not illogical to expect a similar behaviour from your healer team.

At the same time the situation becomes more complex (both figurative as mathematical) because healers can't 'flow' healing, it comes in bursts just like damage. this is were advanced electronics theory comes in to play with sampling, PWMs, differential feedback,... But as I said before, the model is too large and convoluted to simply create a single function for it (I'ld probably take 2 weeks of full time work to come somewhat near that, not including the deriving needed to actually value single stats), so a simulator is probably the easier and more robust way of determining stat values.

That said, with enough experience we can already estimate the effects of changes as we are doing now and having raided in nearly all formats since MC on a myriad of tank and healer classes, I dare say that I know healer reaction well enough to say that Haste seems a better choice for an experienced DK timing his/her Death Strikes around high peaks of damage income.

There are a myriad of factors to consider regarding healing reaction. Your burst size, previous bursts, raid wide damage, any number of mechanics, predicted damage, as well as their current mana can all affect how they react to that particular dmg intake. Each healer will react slightly different to each situation as well, so you'd have to include some incredibly large neuroscience functions into your healer simulator.

This is why I said that a sim is more effective, because a single base function would probably be three pages long, with the stat values around 7 pages in length. Simply not something a normal person can understand and too much work compared to a sim to build (especially considering it isn't a physical theory, values and interactions are subject to change in the future)

Haste also gains value if you can actually take advantage of the increased DS/min, which isn't always the case. You have to wait for burst damage to recover, so if the burst damage isn't coming too quickly (which is generally the case as if it is you should have a cooldown (or multiple cooldowns) active.) then you aren't gaining as much from the added haste unless you stop timing and just go for raw shield frequency, which just trades control for random mitigation, which isn't ideal.

Assuming you DS no later than right before your second pair comes off cooldown, the overall increase is technicly guaranteed. Even in the worst case scenario, where your only get an 'extra' DS in during low damage strings (compared to not having that haste) is a gain in Blood Shield. Haste also helps you by reducing the amount of time you need to sit on runes before a Big Bad Breath, which means there is a larger chance you are at a higher health point right as that big hit lands, basicly increasing the buffer you have afterwards to return to max health, just as you are trying to do with Mastery.

In the end, healers prefer a shield BEFORE a damage burst and a heal right after (or in the middle). Haste helps you for both of those by allowing more frequent shielding (thus more likely some amount of shield is up right before a burst) and less wait time until you can heal after a burst, while Mastery only increases that invisible buffer called shielding after damage spikes.

It is only unpredictable if you are randomly placing death strikes as opposed to intelligently controlling your damage intake. This is also where cooldowns come in to play, as time you can't death strike is time you should be covered by some other form of damage control.

Isn't this an argument FOR Haste? By increasing rune regeneration, you decrease the odds that you can't death strike, so you need your cooldowns less often!

By relying on frequency over control, you are putting yourself at risk of being more unpredictable than with less ds/min but better placed mitigation through mastery.

The control remains exactly the same. What Haste vs Mastery is about is frequency vs magnitude . Just like in the avoidance vs any other damage reduction mechanic debate, frequency trumps magnitude, even if the overall average is slightly higher, because the amplitude of the incoming damage vs average is smaller with Haste. This is exactly the reason why Protection Warriors still keep using Shield Block even if Shield Barrier technicly reduces more damage. Because taking 30% off of 3 hits is more reliable and predictable for healers than not taking damage from the first one and then taking two hits straight to the face.

[Deathelve]

I am really interested in this discussion, on my DK I seem to take burst but since I did not wait to use my DS I am usually in a bad situation. It feels like i should DS on CD to build a shield, but that can leave me in the cold when I need it after a burst. I am not saying that I never look at DBM and think ahead, I do that.

On reflection it is usually when I am tanking several mobs in a raid. I tend to wonder if I am actually doing my rotation as best as possible. But I cannot come with anything different to do. DS w/death, unholy, frost runes, HS with blood. I mean cant really do more then that in my mind. I will admit, I am new to my DK in raids.

I can see the advantage of haste, I dont know it is the absolute lowest priority for stats. I would think it is about equal with dodge/parry if you are gearing hit/exp 7.5%. You are already devalueing dodge and parry.

[Reniat]

You're also ignoring that you will get more SoB stacks so your self healing will be stronger, especially when you delay you DS to counter a predicatable spell nuke

You should already be being patient with your runes, so unless you're waiting to the point where you're actually losing throughput (which wastes haste) you would increase both frequency and melee speed, negating eachother. in other words, you won't gain SoB stacks since the quicker melee speed will be offset with higher spending frequency.

That said, with enough experience we can already estimate the effects of changes as we are doing now and having raided in nearly all formats since MC on a myriad of tank and healer classes, I dare say that I know healer reaction well enough to say that Haste seems a better choice for an experienced DK timing his/her Death Strikes around high peaks of damage income.

And I dare say i've tanked for long enough at every level of the game (from 10N normal casual guilds to guilds at the top 100 world) to say that haste in it's current format WILL NOT WORK for end game tanking. Ideologically it has value yes, and if it had the numbers to back it up it might be more viable, but getting ANYWHERE near the haste you need to remotely make a noticeable difference to your healers will absolutely DESTROY your shield size Your frequency will be better, but your magnitude will be so low you will be taking many large hits without the ability to recover. You'd have to get your rune regen speed to around 4.5 seconds, since you'd basically have to be able to DS after every third hit since the very first hit you take after Death Striking will more than remove your shield, since it will be worth little more than the base shield amount, which in this tier's terms would probably be (depending on the bosses damage output) around 50-80k, compared to a 120-200k shield from the same damage with mastery. In other words, you would need unreal frequency to compensate since you are going to start taking unmitigated hits immediately after each death strike. Your shield would be gone before your GCD was up. literally.


You're right, it does come down to an argument for frequency over magnitude, however in it current format for DKs haste's frequency simply doesn't hold up at all to the magnitude you are losing. It's still a valued stat, and I would take haste over hit/exp for survivability, but im sorry but a haste build simply won't be optimal for any meaningful content outside of fights like Lei Shi, where mastery/avoidance doesn't even work.

[Tengenstein]

Except you wouldn't have to get rune speed to 4.5s since luckily you have 2 pairs of FU runes without any extra rune procs its not particularily hard to get a DS every 4.5 seconds. Unreal? No even with no haste you can easily achieve this goal. but comppile that becuase you are using DS more often you get more RP and more RSs so more extra runes. so even more death strikes. You might get bigger shields but haste give you more Heal, you quote some approximate numbers, which are inconsistant with your earlier values where a full mastery build has shields about 25% bigger.


If youre intrested in TDR Mastery sims about 40DPS better than Haste for me at 4k rating difference and an incoming dps of 170k

[Reniat]

i'm sorry but 4.5 seconds is not possible. You have 8.33 rune regen, but that doesn't mean with two pairs you can death strike every 8.33/2 seconds, since it takes 16.66 seconds for both pairs to cycle back to back. The average DS gap is about 7-9 seconds, depending on your rune regen talent as well as your luck with procs.

I know you're looking for more proof that mastery > haste, but i'm sorry I can't give you any more proof than I already have. If you still don't believe me, I invite you to go tank meaningful end game content with a full haste build and see what happens. (and no, H lei shi doesn't count)

EDIT: oh, and merry christmas

[Airowird]

"Go find the arguments that support my case" is kind of the fundy argument, isn't it?

[Tengenstein]

The average DS gap is about 7-9 seconds, depending on your rune regen talent as well as your luck with procs.
EDIT: oh, and merry christmas

with my 451 ilvl i'm simming sub 7s CD on DS, with 4k haste rating you can get it down to 6.15 seconds on average, which on some bosses would be almost enough to get it to every third hit.

[Reniat]

simming and reality are two different things. I challenge you to find a single parse where there is an average of 4.5 seconds between DS. I don't know of any DKs at any level that would use haste, so you might have to make your own parses.

[Tengenstein]

so can you go back to when you use DS, at what health threshold do you decide to use a DS?

[Deathelve]

Reniat when you have time check your PM's.

[swelt]

This discussion about haste seems to me to be missing the point. Why are you starting from a comparison of haste vs mastery? Surely the much more likely question is haste vs dodge/parry and haste vs hit/exp. For one, there is no gear that ever has haste and an avoidance stat, so you will never be comparing haste+something with mastery+something gear in a meaningful way. A decision you will have the opportunity to make is a slot with haste+mastery (a 'dps' piece) compared to a slot with mastery+avoidance or mastery+hit/exp. Similarly, when I'm reforging, I don't see myself ever reforging from mastery->haste, but I could certainly be convinced to reforge into haste on gear that already had mastery.

I would posit that haste vs hit/exp is mostly about quality of life, unless Blizzard change death strike mechanics to require the attack to land; that haste is flat out better for survival than hit/exp but tanks may choose to retain some (or even quite a lot) of hit/exp for control.

So the best question you could answer me would be at what point in the parry/dodge diminishing returns curve does haste become 'better' for survival.


On a related note: I have been tanking in 25 man, but due to the holidays had to scale back to 10 man. The difference in incoming boss damage was huge. Playing a DK in 25 vs 10 is a massive difference. In 25 man normal mode kills (on certain bosses, like wind lord phase 2, empress, sha) death strike timing in what the guide refers to as 'advanced mode' is pretty much mandatory. The regular boss melee attacks are clearly tuned to stress the healers, but it really gives you the opportunity to work hard on your death striking. By comparison, the same 10 man encounters are much less taxing and can allow some pretty slack timing. This isn't a new thing - it's often been one of the more subtle differentiators - but it's worth remembering that 'good death strike timing' can apply to straight boss melee swings, not just big telegraphed supermoves.

[Reniat]

the lack of actual dmg intake in 10m is why i'll never go down to 10m (until 25m dies out, heaven forbid). There are other things to consider as a 10m tank. Your damage means a lot more than it does in 25m, for example. Even though you take less damage, you also may not often have a dedicated tank healer, so being aware of what they are doing and communicating is as important as it is in 25m, despite the lower pressure. That said, I think the most fun thing to tank is the 25m version of a tank damage heavy encounter (2nd phase 25m heroic grand empress anyone? bad string of rng there = fun times)

As far as haste vs avoidance goes, if you're just looking at survivability it really doesn't hold up. Based on the numbers of euliat's sims, raw TDR isn't a strong enough argument for haste. And the ideological argument for it (which isn't nothing) doesn't hold up since because mastery would still be your #1 stat, you wouldn't be able to get even close to enough haste to make a noticeable difference.

however...

once you factor in damage, things can change a bit. This is a good timing, since we just talked about how in 10m your damage is a larger % of the whole than in 25m, even if your damage from vengeance is lower. While I still don't think haste is a valid contender for survivability compared to dodge/parry, it definetely has some fight as a dps stat. The dps per point between haste and crit is negligibly close (and actually favors haste by a incredibly small fraction of a %), and the dps per point combined with the slight survivability aspect make it a solid dps for DK tanks.

Just taking from my gear, if i swapped every point of avoidance for haste (basically turning every single mastery/avoidance piece for a mastery/haste piece of the same stat budget) I would get a patchwerk DPS boost of 12.9%, which isn't insignificant. Now, getting this much haste would also get in the way of hit/exp capping, which is better dps per point until caps, which at 7.5% of each is also about a 12% increase but only requires 2550+2550 (5100) stat points as opposed to the 12.9% from 8000 proposed haste level from trading EVERY point in avoidance to haste, which may or may not be possible depending on gear options. Just like hit/exp however, haste scales linearly. This means that a 10m (or 25m) tank could slowly push his/her way into more hit/exp/haste until the desired balance of dps and survivability was found.

[lennygel666]

People, really don't follow this guide... It is absolutely terrible.
Lichborne is a COMPLETE waste. Purgatory is the best talent in that tree by far... Prepare for a hit that will kill you, summon your ghoul(or tell your healers to LoH/spam you when you drop) and boom, you now have an extra defensivve cooldown.
http://www.wowhead.com/talent#kLL|
Those are the talents that you should use for 90% of the fights.

Another thing, Blizz seems to have changed a lot in DKs lately(without actually mentioning it anywhere). The set bonus now works 'as intended'(which is weird, seeing how no other +heal increases your blood shield, while the set bonus does), SoB stacks are not consumed when you miss DS and, most importantly, hitcapping is actually important now(afaik, SoB also increases blood shield which it didn't used to do, like any other +heal).

http://eu.battle.net/wow/en/characte...ncaco/advanced
I have not tested the hitcap in raid yet but I will on sunday.

And please OP, rewrite this guide... It's even worse than the EJ guide...

[Reniat]

You don't have to like my guide or Tyvi's. If you think you can write one better than you are welcome to it. Theorycrafting isn't about monopoly, and if you think you can do better than I encourage you to do so. This guide is written by a tank who currently tanks at a world 100 25m level, and has tried his best to include everything that goes into playing at that high level as well as in normal/5m tanking. There isn't anything that I do to successfully tank at a world 100 level that isn't in this guide in some way or another. I can safely say Tyvi (at EJ) takes their guide just as seriously and puts just as much into it as I do.

Also i'm not sure why you think I devalue purgatory. I've used it for the entirety of the tier (except for when I needed to use AMZ as a raid CD for H elegon) and recommend it here in the guide for any progression tanking where a tank can get bursted down quickly. If that's not a risk at all (such as non-progression normal content) then there really isn't any reason to get purgatory unless you have abnormally slow healers, and the self heals from Lichborne would see more use. If you are getting bursted hard enough in 10N non-progression content that you need purgatory than I question how you are using CDs.

TL;DR
if you don't like it, write a better one. Just coming out and saying "this guide is teh suxor" doesn't help me as the writer, or the Blood DK community as a whole.

Hit/exp:

SoB has increased shield size as well as raw heal since beta, and the fact that SoB stacks aren't consumed by missed death strikes is an argument for ignoring hit, since if you are hit soft capped you won't miss DS in the first place. Hit/exp reduce the chance of losing SoB stacks, and unless you're hard capping (which requires a significant loss of mastery in our current gear) you can't guarantee every single SoB stack so the "100% consistency" argument doesn't hold up. According to Euliat's sims, soft capping both is a 2% mitigation gain versus 0% of each, and that's not including the lost avoidance which makes hit/exp capping a net loss in overall mitigation. Not to mention that in our current gear it takes a loss of mastery in order to reach even the soft caps, which is our best stat for gaining stability in 25H raid tanking (and all raid tanking). just going by your armory as of this post you're losing 2072.2 mastery by getting as much accuracy as you are. In other words, hit/exp capping doesn't hold up as a stability/consistency argument or an overall mitigation argument. there is only a 15% chance per hit of even seeing ANY benefit from soft capping each versus 0% of each, and the mastery lost in the process of capping accuracy is going to be lost on every single shield.

Another thing, why are you using Fine gems in red slots instead of keen gems? If you got a bit of exp from gems you might not have to lose as much mastery from reforging. Surely you can at least agree that mastery > parry.