Upcoming Tank Squish

Forum Avatar
Game Designer
#1 - Sept. 3, 2014, 11:08 p.m.
Blizzard Post
Hey all. I've briefly mentioned this in tweets, but wanted to provide some more concrete info here for you. In an upcoming build, we're making some major changes to tanks.

But don't freak out, tanks, this is for the better. It'd be very easy to see datamining of these changes and think the sky is falling; don't be deceived! If you do the classic "scroll to my class, read datamined changes to my class only", you'll be sad and angry. And I'll be sad that you're sad and angry.

The core issue that we're aiming to solve is with the power level of tanks' defensive abilities. And I don't mean their gameplay. Their gameplay is great, especially Active Mitigation; that's not changing. I mean their total defensive effectiveness. Health, damage reduction (flat, percentage, random avoidance, etc), passive and active, cooldowns, etc.

Tanks have gotten a lot of power creep. A tanking specialized character should be better at tanking than a non-tanking specialized character, obviously. But by how much? 3x? 5x? 10x? It's gotten to be more like 50x or even 100x. Fun fact as a point of comparison: Most of the Mythic raid bosses (first raid tier) we've been testing deal tank damage roughly equivalent to what Heroic Lei Shen (second raid tier) did, after accounting for the squish and level difference. Some even more.

So, what are we doing about it? We're effectively doing a squish to tanks defensive effectiveness, and to dungeon and raid mob damage to tanks. This list is going to sound like a big pile of nerfs, but it's important to understand that it's not. It's happening to *all* tanks, and mob damage is coming down as well, to compensate.

Additionally, before anyone points it out, yes, some of these changes affect the defensive effectiveness of non-tanks. We didn't forget that; this was done consciously, because we think these changes should be made anyway, or we're considering other changes to compensate them if necessary.

Here's a summary of the actual changes (please post this around, and discuss this, not the datamining which will paint a very skewed and incomplete picture):

  • Creature damage has been retuned. In particular, the damage of creatures intended to be tanked in dungeons and raids has been drastically reduced to offset the below changes.
  • The amount of Armor on Plate, Mail, Shields has been reduced significantly.
  • Many tank abilities that increase maximum health have been reduced in effectiveness: Guarded by the Light, Stance of the Sturdy Ox, Bear Form, Empowered Bear Form, Blood Presence, Veteran of the Third War, Shadow of Death (Enhanced Death Coil), Ursa Major. In most cases, the magnitudes of these effects have been reduced.
  • Many tank abilities that passively reduce damage taken in some way have been reduced in effectiveness: Unwavering Sentinel, Improved Defensive Stance, Mastery: Critical Block, Defensive Stance, Sanctuary, Guarded by the Light, Stance of the Sturdy Ox, Bear Form, Blood Presence, Primal Fury. In most cases, the magnitudes of these effects have been reduced.
  • Many tank abilities that provide Active Mitigation have been reduced in effectiveness: Shield Block, Shield Barrier, Shield of the Righteous, Mastery: Divine Bulwark, Bastion of Glory, Guard, Stance of the Sturdy Ox, Shuffle, Savage Defense, Pulverize, Frenzied Regeneration, Tooth and Claw, Death Strike, Rune Tap. In most cases, the frequencies or magnitudes of these effects have been reduced.
  • Many tank abilities that provided long-cooldown temporary defensive buffs have been reduced in effectiveness: Shield Wall, Last Stand, Demoralizing Shout, Divine Protection, Guardian of Ancient Kings (Protection), Fortifying Brew, Barkskin, Bristling Fur, Survival Instincts, Might of Ursoc, Bone Shield, Dancing Rune Weapon, Icebound Fortitude, Vampiric Blood. In most cases, the durations of these effects have been reduced.
  • Resolve has been changed to no longer scale with Stamina, only incoming damage over the last 10 sec, and its scaling rate changed.
  • Resolve % = 100 * MAX(0, 8.5 * (1 - e^(-0.045*DamageMod)) - 1)
  • DamageMod refers to the % of a basic equal-level creature attacking you that you've had directed at you over the last 10 sec.
  • EDIT: See Post #40 on Page 2.


We understand that these changes will sound scary, but ask that you keep an open mind and try to comprehend the complete picture. Tuning may be pretty rough for a few builds as we iterate on these changes, so please be patient.

I'll try to answer questions as I have time, but please understand that we're all extremely busy trying to finish and polish things over here, so time is limited.

Thanks, everyone!

EDIT 1: Moved Bristling Fur to the Cooldowns category, had accidentally miscategorized it.
EDIT 2: Fixed another typo.
EDIT 3: Clarified that "Shadow of Death" is the buff from the Enhanced Death Coil perk.
Forum Avatar
Game Designer
#33 - Sept. 4, 2014, 12:16 a.m.
Blizzard Post
09/03/2014 04:21 PMPosted by Morrie
Is the plan for tanks to still do ~70% of DPS damage? Just wondering.

I don't know that we've given specific numbers, but those plans haven't changed.

09/03/2014 04:21 PMPosted by Celladore
First thing stood out to me is that some of the long-CD abilities being altered will have an affect on Melee DPS's raid survivability.

And while I will concede that in 1st teir, incidental AoE damage on the raid (melee specifically) wasn't too terrible in Mists, I'd like to know if this is something that's being kept in mind as you alter those cooldowns. Especially in the long run, as we proceed to the 2nd (and maybe 3rd) teir of raiding in WoD.


Yes, it is indeed being kept in mind.

09/03/2014 04:29 PMPosted by Síxxs
I'm confused as to why these changes would be introduced near the very end of mythic testing. Would it not have made more sense to make all of these major class changes, such as the gimping of Holy Paladins and now the squishing of tanks at the beginning of mythic testing so you have more than a week or twos worth of data to play with?


We'll still be doing considerable testing.
Forum Avatar
Game Designer
#36 - Sept. 4, 2014, 12:36 a.m.
Blizzard Post
09/03/2014 04:29 PMPosted by Quynt
I am somewhat concerned with one change. The reductions to both the power and duration of long cooldown tank CDs.


We just reduced the duration of long cooldown tank CDs, except for a couple where we changed the cooldown or magnitude instead. I don't believe anything had both the magnitude and duration changed.

09/03/2014 04:35 PMPosted by Jaeger
This sounds interesting. It sounds like non-tanks will have better survivability compared to tanks. In my opinion, non-tanks should be able to take a hit or two without being one-shot while tanks can take a few more hits before they risk dying. Healing a non-tank should be doable for short periods of time, but not sustainable over a large part of an encounter, but tanks shouldn't be indestructible.

That sounds like the goal, so I'm interested to see how it plays out.


Indeed.

09/03/2014 04:32 PMPosted by Fortegigas
Well if this is scaled with encounters, I guess it's not too bad, but how will this effect our effectiveness in PvP?


We had been planning to raise the amount of damage tanks take in PvP further, but are doing this instead. We're going to see how that works out, and may leave it at 25%, or even remove it if that looks reasonable based on testing.

I'm confused as to why these changes would be introduced near the very end of mythic testing.

Because the issues here were highlighted by Mythic testing. Numbers changes are significantly easier to make than mechanics changes.


Also quite true.
Forum Avatar
Game Designer
#39 - Sept. 4, 2014, 12:43 a.m.
Blizzard Post
09/03/2014 04:49 PMPosted by Kasuto
What is the exact number for "a basic equal-level creature" at level 100?


Data!

Level, Baseline DPS
90 847.125
91 1161
92 1425.06
93 1796.76
94 2179.35
95 2646
96 3114.45
97 3238.65
98 3364.2
99 3489.75
100 3615.3

...and this made me realize I forgot one very important part of the new Resolve when I typed up this post. Apologies, going to add it to the post now.
Forum Avatar
Game Designer
#40 - Sept. 4, 2014, 12:49 a.m.
Blizzard Post
Actually, can't edit it into the OP; post is getting too long. So, editing in a link to this post.

  • Resolve has been changed to no longer scale with Stamina, only incoming damage over the last 10 sec, and its scaling rate changed.
  • Resolve % = 100 * MAX(0, 8.5 * (1 - e^(-0.045*DamageMod)) - 1)
  • DamageMod refers to the % of a basic equal-level creature attacking you that you've had directed at you over the last 10 sec.
  • This Resolve amount is much much larger than previously. To compensate for that, baseline healing and absorption from tanks is being significantly reduced. The net result will be a curve that keeps abilities such as Word of Glory, Frenzied Regeneration, or Shield Barrier scaling roughly in parallel to other %-based abilities such as Shield of the Righteous, Savage Defense, or Shield Block; weak when fighting weak enemies, and strong when fighting strong enemies.
Forum Avatar
Game Designer
#41 - Sept. 4, 2014, 12:57 a.m.
Blizzard Post
09/03/2014 05:25 PMPosted by Choll
09/03/2014 05:10 PMPosted by Ashunera
...This is the part that breaks the game. You should never feel invincible until you severely over gear something. Mitigation, positioning, keeping a group moving, these are the roles of a tank.


With these changes, all that you said are the roles of a tank can be served by a DPS in defensive stance. If that puts out better DPS and we still survive, then why be a tank?

The need to be very careful reducing damage done by mobs just so they can reduce our defensive ability. There is a point where a tank is no longer required and can just be done by a plate wearing dps.


There is a point, but it is *extremely* far away, and we are nowhere even remotely close to that point. This isn't going to make non-tanks able to replace tanks.
Forum Avatar
Game Designer
#47 - Sept. 4, 2014, 1:19 a.m.
Blizzard Post
09/03/2014 06:15 PMPosted by Theck
So if I'm understanding this correctly, at level 100,

DamageMod = (damage taken in last 10 seconds) / ( 3615.3 * 10 ) ?

Is the time-decay mechanism of the previous version of Resolve, in which an attack was worth 2x Resolve at the time of impact and decayed away continuously to being worth 0 Resolve after 10 seconds, gone entirely?


The time-decay mechanism from the previous version is still there. So it's (weighted damage taken in last 10 sec) / (3615.3 * 10).
Forum Avatar
Game Designer
#59 - Sept. 4, 2014, 4:03 a.m.
Blizzard Post
09/03/2014 07:57 PMPosted by Littlejohnny
Also I have to echo the opinions above about AM. If AM is nerfed too much it will get to a point where it doesn't matter if you use it or not. AM should feel meaningful to use.


Fear not. We're not making anywhere close to that level of change.
Forum Avatar
Game Designer
#96 - Sept. 4, 2014, 6:13 p.m.
Blizzard Post
09/04/2014 10:09 AMPosted by Ariellè
Or, y'know, they don't like the gameplay of consistently having to 2-shot tanks in order to make their survival threatened.


This.
Forum Avatar
Game Designer
#198 - Oct. 2, 2014, 5:37 p.m.
Blizzard Post
Changes to mob damage are still ongoing, and it's something we're hand-tuning on a dungeon-by-dungeon and boss-by-boss basis.

One issue that has likely been significantly impacting the feel of tanking in dungeons and in multi-mob scenarios, is some legacy logic from 5.4 Vengeance that is no longer needed or applicable in the current tanking environment.

On live in Mists, Vengeance has diminishing returns with respect to damage taken from multiple targets, because such damage proved to be much less deadly in practice than the same amount from a single source. See this post for details: http://us.battle.net/wow/en/forum/topic/9573038382#12

However, with the general increase in tank health with respect to incoming damage, and the move away from "spiky" damage intake, tanks spend much more time partly injured. In Mists, what killed tanks was sudden bursts of damage. In Warlords, what kills tanks is sustained damage that exceeds incoming healing and self-healing. And when it comes to sustained damage, it is immaterial whether you're being attacked by one huge enemy or many smaller enemies.

Thus, we're removing any diminishing returns from Resolve when tanking multiple targets. A hotfix that makes this change should be in effect on beta within the next day. This should significantly improve survivability in dungeons and other multi-mob situations.
Forum Avatar
Game Designer
#202 - Oct. 2, 2014, 6:41 p.m.
Blizzard Post
10/02/2014 11:06 AMPosted by Zyzx
Thanks Watcher! Is there anything the community can do further to help with the tuning? (i.e. run older dungeons at level appropriate)

Being specific with your feedback, mainly. "My self-healing feels weak and healers can't keep me up, this feels terrible!" is valid feedback, but it's hard for us to address directly. On the other hand, "I was tanking Heroic Slag Mines on my premade DK and I couldn't survive the two-ogron pull before the last boss if I didn't have all my cooldowns up. They did way more damage than any boss in the dungeon." is very actionable.
Forum Avatar
Game Designer
#237 - Oct. 4, 2014, 12:28 a.m.
Blizzard Post
Tanks for all of the feedback.

Warning, the following is complex and techy; you don't need to know this in order to tank in Warlords.

As we've described, the design of Resolve is intended to maintain a balance between abilities that use percentage effects (such as Shield Block) and those that use flat amounts (such as Shield Barrier). In the below I'm just going to refer to "Barrier" and "Block" as examples, but this applies to all tanks' active mitigation and healing.

An important factor there is where the balance is at with no resolve. Obviously, if you're fighting a critter that's doing 10 damage to you, Barrier is going to win (and that's fine). So, we have to choose a starting point for Barrier with no Resolve. At that level of incoming damage, Block vs Barrier is balanced; beyond that, Resolve kicks in and starts bringing Barrier up.

When we implemented this version of Resolve, we chose that starting point as being a hard-hitting outdoor creature (or pack of creatures). We chose that so that you'd see Resolve taking effect as you learn to play in solo content. In order to achieve that, we had to bring the static abilities like Barrier down much lower, so that with no Resolve, they were balanced with Block against the hard-hitting outdoor creature (which is still pretty weakly hitting, in the grand scheme of things). We implemented that by making Barrier (and similar) 60% weaker at baseline, which is the right balance point against that hard-hitting outdoor creature.

However, that has some downsides as well. It feels negative, and it also means that for tanks with heals, out of combat healing yourself up after a fight is quite slow. So, we're going to revise that starting point up to be a hard-hitting dungeon creature, which makes the abilities balanced at their natural effectiveness; no 60% reduction needed. This means that Barrier will be stronger before that point, but Resolve won't kick in at all until you're being hit by a hard-hitting dungeon creature (or equivalent). To reiterate, this has no functional difference above that; self-healing will be unchanged against a raid boss, for example.

Again, thanks for the feedback on this, and for following along through our iteration process. It helps!

P.S. For theorycrafters, the 8.5 changed to a 3.4 in the Resolve formula.
Forum Avatar
Game Designer
#246 - Oct. 4, 2014, 1:05 a.m.
Blizzard Post
10/03/2014 05:49 PMPosted by Rilac
So basically you're saying that after this change that you just spoke of...

1). We will self-heal a lot more in solo/out of combat (so that Ret isn't more survivable than Prot and it ain't taking us 60+ seconds to heal up after combat)
2). We will self-heal about the same during raid content (so tanks aren't overpowered like in MoP)

Is that about right?


Yes.
Forum Avatar
Game Designer
#249 - Oct. 4, 2014, 1:15 a.m.
Blizzard Post
10/03/2014 06:03 PMPosted by Ahanss
Any thoughts on Resolve's scaling at high levels of damage taken? As of right now, we're getting pretty absurd levels of it in raid situations; even the mythic tank dummy with no stacks takes us about 80-85% of the way to the hard cap on our healing. I don't agree with the folks that say to make it linear (since there needs to be something to account for our better gear scaling this expansion), but even I think that Resolve's scaling is pitiful at high levels of damage taken.


Vengeance was a huge additive increase on a trivial base amount.

Resolve is a multiplier on a very significant base amount that grows with gear. As your gear improves, that multiplier needs to grow slower in order to produce the same effect. Your gear improving will make those smaller increases in Resolve more meaningful.