이 사이트는 자바스크립트를 사용합니다.
을(를) 해 주십시오.
Should I Switch to a Blood Death Knight in Patch 8.3? - An Analysis
2019/12/27 시간 16:18
에 의해 작성됨
Recently, a number of streamers have advocated Blood Death Knight as the go-to tank for the upcoming Ny’alotha raid, citing a number of reasons, ranging from the type of encounter to class design. This has led to an increase in questions regarding the veracity of these claims. In this article, we’ll go over a number of the arguments put forward, and provide a balanced and objective take on it.
One of these claims does require us to look back in time at the history of the spec. Hold on tight and grab some popcorn. If you’ve been playing Death Knight for a while, you may recognize some of the moments mentioned in this.
About the Authors
This article was written by Mandl (Mandl#0001), a useful minion of
and BDK theorycrafter, in collaboration with emallson from Peak of Serenity. Both have been active, progression raiders and are heavily involved in optimizing the various tanks in Battle for Azeroth.
Claim #1: Blood Death Knights “scale better at the end of an expansion”
A common trope currently is the statement that BDKs “scale exponentially just like every expansion”. In order to be able to ascertain this one, we’re going to need to look at the history of the class, when they are remembered to be strong, and why this happened, over the course of the past 5 expansions.
Cataclysm: Late Firelands
Who does not remember the glorious days of the end of Firelands and the beginning of Dragon Soul, a time when mastery actually did something, where our toolkit suddenly felt like a breath of fresh air compared to earlier in the expansion, and where we were considered top dogs for the overwhelming majority of bosses in the last tier of the expansion?
The culprit behind this was not stat scaling, nor was it the (insanely good) set bonuses. At the core of this sudden reversal was a set of patch notes, applied on the patch of Dragon Soul (4.3.0):
now heals the death knight whether or not the attack misses, or is dodged/parried. As a result of this change,
no longer refunds its rune cost if it fails to hit the target, as the death knight will still receive the healing effect.
Blood Presence now provides an armor bonus of 55%, up from 30%.
Blade Barrier has been redesigned. It now passively reduces damage taken.
Bone Shield now has 6 charges, up from 4.
Cataclysm was one of the expansions with arguably the most dangerous raid design in terms of tank damage, as it still was very much like Wrath in that bosses would auto-attack for a significant chunk of your health. This was exacerbated even further by the prevalence of block tanks; due to the way mastery was implemented for protection paladins and warriors, it was possible for them to reach something called CtC cap - where the sum of all their evasive stats (Dodge, parry, block) would be high enough to guarantee that every hit would be blocked. Back then, a block was a fixed 30% damage reduction; this effectively meant that block tanks always had a cooldown rolling compared to Death Knights in the form of blocks.
Due to this, it was
for Death Knights to be able to successfully land
, as this was a significant chunk of mitigated damage (through blood shield). However, back then, a dodged or parried
would not provide any healing, and, due to the magnitude of boss damage, you'd be very likely to be dead 1.5s later. As a result, the only viable option was to sink 15% worth of secondary stats into expertise in order to prevent this from happening. Death Knights were the only tank with this requirement until this patch-level change.
A major stamina bonus and quality of life improvements were also pretty impactful on the Dragon Soul experience (particularly the blade barrier change; back then, blade barrier was a damage reduction effect that would occur whenever your blood runes were on cooldown. Making it passive opened up some gameplay changes on the rune front). This, however, was not the only factor at play; encounter design also had a large impact in this, with the overwhelming majority of Dragon Soul having mechanics that were flat-out designed to not be blockable, thus massively favoring death knights and bears. Gone were also the fights requiring ludicrous tanking setups such as Halfus, Baleroc et al.
All this has been confirmed and publicly highlighted by Ghostcrawler in a
On top of this, an article from Sacred Duty
tells this story from a paladin PoV
Mists of Pandaria
Another expansion, another tier in which blood death knight strengths suddenly showed up. In the expansion where tanks were designed around taking damage to deal damage (through Vengeance), one commonly attributes the success of Blood at the end of the expansion, with famous death knights like Troxism leading the charge on bosses being solotanked, to “late expansion blood”.
The reality is altogether different, and is, once again, the direct result of a set of patches:
Riposte is a new passive ability learned by Blood Death Knights at level 76. When the Death Knight dodges or parries any attack, they gain 75% of their Parry and Dodge as additional bonus to Critical Strike for 20 seconds.
Dancing Rune Weapon] no longer costs Runic Power.
Sanguine Fortitude no longer reduces the cost of Icebound Fortitude.
Scent of Blood now also has a chance to activate when the Death Knight dodges or parries a melee attack.
The major change in this list is the change to Scent of Blood; its original effect causes your melee critical strikes to stack a buff increasing the healing done by
by 20%, and causing it to generate 10RP. It also generating from attacks avoided contributed massively to the RP economy.
To give you an idea of the scale of this change, two logs from MoP are still available on worldoflogs and can be used for comparison:
A 10N Council kill, from Throne of Thunde
A 25H Malkorok kill
Despite being a similar fight time, with Malkorok being on the low side in terms of auto damage (and therefore dodges/parries), Scent of Blood generated 30% more RP for the death knight.
This set of changes on its own upended the entire gearing flow of Blood. Prior to 5.4, Blood Death Knights were primarily stacking Haste after hitting hit and expertise caps; due to the addition of Riposte, and the fact that it itself got amplified by Scent of Blood, suddenly, avoidance stats (dodge/parry) became monstrously valuable.
These changes were a reaction to the trend of nerfing Blood during the entire expansion (or buffing every other tank), as it was (initially) the highest DPS tank in early MoP, such as:
5.0.5 hotfixes: +25% prot warrior damage, +50% prot paladin damage, +60% thrash damage (druid), +50% keg smash damage (monk), +100% swift reflexes damage (monk)
5.0.5 hotfixes: 7.5%
nerf, death siphon AP coefficient nerf (minor), 8.5% rune strike nerf, 8.5% heart strike nerf
5.1.0 hotfixes: Heart of Fear has a mind control,
now no longer takes into account PvP damage
5.2 Blood Parasite now summons Bloodworms with 15% of the Death Knight's health (was 18%) and Blood Burst now heals for 25% per stack of Blood Gorged (was 30% per stack).
And the list goes on. One can summarize MoP as “we nerf DKs for 3 tiers then massively buff them”, which, incidentally, is now remembered as a result of late expansion scaling.
Warlords of Draenor
Warlords is an altogether different beast in that, instead of being a straight-up set of patch notes, the “scaling” that everybody remembers was the result of the discovery of a new playstyle, referred to as
Chains of Sindragosa
To put it simply, abusing a glyph (causing chains of ice to grant double RP), an affinity towards gearing for Multistrike, and Breath of Sindragosa, led to some pretty scary numbers, both defensively and offensively.
Before this playstyle was fully utilized, Blood was mostly brought in as a DPS tank (for some reason, Blood Boil was tuned for ST damage while doing full damage to other targets. This was adjusted partly through Highmaul). This playstyle change and a couple of additional items (vial of convulsive shadows from Maidens in BRF being the most notable one - who doesn’t like an on-use multistrike trinket, which, whilst not being required, was a really noticeable addition to the kit) led to the dominance of Blood as we knew it back then.
mediocre in early Legion content, to the point that tanking certain bosses (Ursoc, Cenarius) was considered unwise, bordering on the insane. Cenarius, in particular, highlighted why a fight centered around reducing upfront damage spikes (the bigger the spear hit, the bigger the pool of corrupted ground) just did not work with Blood. It was so bad that we had to get Icebound Fortitude brought back from pruning halfway through Emerald Nightmare, and even with that, we were still on drake duty on that fight. To make matters worse, back then, Bone Shield had a 1s internal cooldown and periodic damage could remove a charge. Most of the DoTs in Emerald Nightmare had a 1s tick period.
Nighthold was no different, with certain fights requiring a drastic amount of skill compared to other tanks (Spellblade Aluriel being the prime example); on the other hand, pure magic damage fights like Star Augur saw us solo-tanking a raid boss on progress, reinforcing the “BDK is great at magic damage” trope that started in earlier expansions, back when
was the only active mitigation that scaled off magic damage. This stemmed from a combination of factors - AMS could be used to reset the fel phase debuff, and even after the tuning pass after Serenity killed it, the tank damage was very much a joke.
The advent of the new artifact traits and the addition of a specific legendary (Skullflower’s Haemostasis) led to a drastic increase in the power level of the spec; previously, our weakness was our (comparatively) low EHP and our damage intake. 7.1.5 saw us gain +20% (and up to 15% extra from relics) healing and damage from
, Vampiric Aura (25% leech on the group on 33% uptime), a copious amount of small buffs from every artifact trait gaining a rank and Souldrinker (a fresh take on the old Shadow of Death) increasing our EHP by a percentage of our overhealing. This, combined with a few extremely lenient encounters (Kil'Jaeden being the only exception, Argus being a close second) “fixed” blood in raids and made us mandatory in M+.
With all this, I hope the case is clear: BDKs have never directly scaled abnormally off a stat (or stat budget) on its own without a hotfix, and the only time this was the case was the Scent of Blood buff. Every other case remembered by the community as “late expansion scaling” was the direct result of balance patches.
Claim #2: “Blood Death Knights scale exponentially off stats”
To an extent, every spec in the game has non-linear scaling with certain stats and interactions between them. This can be verified with scaling plots straight off simulationcraft, or DTPS analysis for defensive benefits.
What blood does not have, however, is a stat that gets massively better the more you have of it. To highlight this, we’re going to walk through each stat and their interaction with the rest of the kit.
Crititical strike benefits are twofold; offensively, it grants an increasing chance to your chance to critically strike, and defensively, 100% of critical strike rating is turned into parry rating (through
Due to the fact that no heal in our base toolkit can crit, we’re left with the only defensive influence from crit being parries, and this itself has diminishing returns. The further up you go, the less effective parry % per parry rating you get, bringing it back down to sensible relative gains.
Unlike certain other specs in the game, we do not have abilities that generate a disproportionate amount of resources through crits. The closest we get to this is Blooddrinker, and this is just down to it being a 100% health transfer.
Mastery gives us extra AP and larger blood shields, linearly. Nothing surprising or unusual there; this interacts somewhat with versatility (to the point that vers slightly amplifies the value of mastery), and with DTPS, but due to the low coefficient per percent and the fact that Death Strikes heal for significantly less than prior expansions, we’re left with a stat that is mediocre at best in the majority of circumstances.
Haste provides additional GCDs and runes. Contrary to popular belief, the “extra runes” part is fully quantifiable, and assuming proper play and no haste debuff like the one on the first two packs of Siege, all the additional runes end up being spent on Heart Strike, netting a 50% RP gain over “natural” runes.
This, however, is still not exponential and does not have additional scaling terms beyond the 10% we get from all sources through bone shield. Defensively, it has no interaction with other stats than mastery (simply because more death strikes => more blood shield).
Versatility is a linear increase in healing done and linear reduction in DTPS; it amplifies the value of mastery and haste slightly as a result.
In conclusion, we can conclude that the following interactions are present:
Mastery, Haste and Versatility interact
Critical Strike has no real meaningful interaction with the rest of the kit beyond reducing DTPS
The real, difficult question is “how much do they interact”?
In order to evaluate this, a pretty sizeable number of iterative simulations were performed, and analyzed (together with the folks from Peak of Serenity). Emallson’s analysis and breakdown of the dataset is available as a gist:
The takeaway are the following charts:
All of these plots are a fixed scenario of 700k/1.5s physical DTPS (physical damage is the overwhelming majority of damage taken in just about any scenario except the following:
Rixxa Fluxflame, Mogul Razdunk (pure fire)
King Mechagon p1 (pure nature damage through Pulse Blast).
Prophet Skitra in the Ny’alotha raid
Based on this, a hierarchy appears. Crit is the most valuable secondary for EHRPS reduction according to the sim data. Versatility is worth less than Crit, but more than Mastery / Haste.
Crit's value is entirely in the increase in Parry rate; this stems from the
skewed nature of damage profiles in BfA raids
, and is something that has been extensively discussed by other class writers.
It is overwhelmingly clear that the following are true:
On fixed physical DTPS, the effect of haste and mastery on EHRPS are essentially equivalent
Haste does not “make you self-reliant”
Additionally, this analysis allows us to attempt to fit a model on the data, in order to figure out whether we can expect linear scaling, or anything else. Again, the full analysis is in the gist, but the most interesting chart is the following:
This clearly highlights the fact that, contrary to popular belief, haste has nothing “exponential” about it, at least not in Battle for Azeroth and most definitely not in the range of secondary budget we’re likely to experience.
As usual, situations differ; an encounter with overwhelming magic damage will massively devalue critical strike defensively. This, however, isn’t that common; from PTR testing, a large amount of bosses were dealing comparatively less magic damage than some of the bosses we’ve seen in the Eternal Palace, some even in absolute. As the encounters get tested further, we’ll be able to gauge this, but for now, things look very much like 8.2.
What tank should I reroll to in Patch 8.3?
If you are considering to reroll to anything, I would strongly advise monk still. There will be opportunities for mass grip to shine in Ny’alotha, but some of them can easily be covered by a VDH (vexiona), are not strictly necessary or can be covered by DPS DK single grips (hivemind). In particular, both frost and unholy are looking very appealing, which is something else to consider.
The magic damage argument also does not play in favor of BDKs. With the recent buffs to VDH passives (15% passive DR from all sources), their massively increased health pool and the ability to benefit from defensive talent and essence choices, something which BDKs just do not have (to the point that even our cheat death has been malfunctioning the entire expansion due to its interaction with Will of the Necropolis), the only thing we have over them is AMS’ potential to immune debuffs. This, just like every tier, should be considered on a fight-per-fight basis and not as a “blanket” (especially since protection paladins can do the same, along with significantly more group-wide utility).
At the end of the day, your success at anything is determined by your ability to play the spec optimally. Rerolling to a new class always brings a significant penalty in the form of ramp-up, both mechanically and in terms of gear, and BfA exacerbated this even further with essences. Blood Death Knights are not a panacea, and actually not a simple tank to master due to their indirect feedback loop (mistakes do not immediately crush you; instead, it all stacks up against you in the form of potential RP deficits and missed Hemostasis stacks), and are definitely not worth rerolling to because of “scaling” or “magic damage” - those two are age-old myths. It is not the first time, and it likely won’t be the last, that
the perception of a specialization has skewed the meta
If you do decide to play one, or continue playing one, you won’t be that badly off either; we're still perfectly viable, and tanking choices in 8.3 are looking to be somewhat more balanced compared to earlier tiers. If you do, I’ll see you on Acherus; if you don’t, I wish you all the best in your tanking endeavors!
Find out more about all the Tanks in our Class Guides!
Blood Death Knight
Vengeance Demon Hunter
와우헤드 프리미엄 이용하기
한달에 $1보다 적은 값
으로 광고를 보지 않고, 프리미엄 기능을 해금하고 사이트를 후원해 보세요!
댓글 0개 보기
댓글 0개 숨기기
댓글을 달려면 로그인하세요
로그인이 되어있지 않습니다. 댓글을 남기려면
을 해 주세요.