Monday, August 20, 2012

WoW MoP (Beta) Class Balance Analysis - deeply analyzed

You say that you are completely confident in your ability to balance the classes. What does that mean? I believe you when you say you can make a computer play a class on every spec on a patchwerk style fight and get the same numbers. Thats where my faith starts to drop off.
There are two general strategies to balance the game. One is to guarantee that everyone's damage is pretty close on a single target that doesn't move and has no real gimmicks (e.g. Patchwerk). The other is to guarantee that everyone's DPS will be equal on every encounter. The latter isn't a realistic goal, and honestly, we wouldn't want that anyway. We think it's cool when an encounter happens to play to your strengths, and it can also be fun when an encounter emphasizes your weaknesses and then it's up to you to try and overcome those challenges. As I've said before, it's not the straightaways that make racing fun -- it's the curves.

I will include my standard caveat that some players see top 100 parses and assume that is the damage they will do. Skill still matters enormously for being able to eek out those numbers. One of the consistent characteristics of the best guilds in the world is they are able to deliver near-Patchwerk numbers even on very complex fights. I watch them do it and still don't quite understand how.

Also, there are only a few fights where absolute DPS is the key to winning. Mastering mechanics usually has a much bigger impact on success. Players tend to downplay that in their quest for chart-topping DPS. Plenty of people raided (and even had fun doing so) back when DPS specs did double or triple the damage of each other. That's not an attempt to escape from balancing the game, but from where I stand, players overall place far too much emphasis on balance equaling fun.

The third point is also true, but that's why it surprises me so much that you've disregarded (actually you've specifically stated that you don't want to hear) feedback about certain classes and specs being very not fun.
We don't disregard it. Not at all. It's just very subjective.

I think it's easy for players posting on our forums to believe that this is our primary way of interacting with players. A very small, almost trivial, number of players post on our forums. It's very dangerous to assume that several very passionate forum posters are speaking for the community. We have a very storied history of making a change because some players argued for it very passionately, eloquently and logically, only to have a bunch of other players then get angry because they were perfectly happy with the previous design.

If there was some very safe, easy and exploit-free way of polling the entire player base, then it might be possible to know if players overall found a mechanic fun or not. Without that, we're left to just making judgment calls on whether we buy the argument or not. Sometimes we do and sometimes we don't.

The 10% melee/hunter attack speed buff from the dummies is stacking with Unleashed Rage(10% speed) from enhancement shaman.
Not intended I'm sure.

There were several bugs with the various sources of the 10% attack speed raid buff stacking incorrectly. They should be fixed soon.

Should Expertise convert to Spell Hit for pure casters? Or is that only supposed to work for melee classes that use weapon attacks that deal spell damage? (Pardon my ignorance here, I know nothing of melee stats.) Basically, I'm wondering if every Human Mage and Warlock is going to be looking for swords now.
Yes, Expertise should indeed convert to spell hit, even for pure casters. Casters might have a slight preference for certain types of weapons (as melee have for years). We make some mage blades, but not a lot.

In other words, tooltips currently do not reflect how much a spell is actually supposed to heal for once the benefits of PVP Power are factored in.
PvP Power does not change spell tooltips. This is partially because PvP Power only affects you in certain areas (specifically not in dungeons or raids).

13,20% Mastery - Increases the damage done by your demon servants by 13,20%. Increases the damage you deal in caster form by 13,20%. The damage you deal while using Metamorphosis increased by 39,60%
It said 39%, but it was lying. It only gave you 39% for 2 sec and then dropped off to 26% or something but didn't display the drop.

These are the kinds of bugs we find all the time before launch and require us to adjust numbers, which is why we implore you to try out the changes and make absolutely sure that we were just nerfing you out of spite before you curse GC and his Throne of Lies.

GC, is it the design intent that some classes can do 15% more single target damage than others? Currently the classes on the bottom of the DPS sims are that far off the top classes.

Can you clarify what is your target balancing spread between the best single target class and the worst single target class? It seems to me that 5%ish would be acceptable, but the current spread is close to 15% which is just outrageous.

Are the sims accurate enough to go on at this point? They don't yet match what we're seeing.

I don't want to change the topic of this thread, so I'd prefer if we didn't debate the issue here, but I don't believe sharing our target numbers is a good idea. My concern is that a chunk of players would want to endlessly debate whether the numbers we chose were fair for their class, and another chunk would want to argue about whether those numbers were achievable.

What I mean by the latter is that if we said "Our Balance druid target is 115,000 DPS," then anytime someone saw a number on Raidbots or wherever that was lower than that, they would demand buffs. They demand buffs today based on sims that may or may not mirror reality. Other players would argue that it takes excessive RNG to achieve the targets, or movement-less fights, or that we need to introduce more cleave encounters. None of those concerns are necessarily invalid, but they are a huge distraction. We'd rather players focus their communication efforts with us on whether their class is fun (though paradoxically, this is not the "fun" thread). Honestly, I think it's more fun in game design in general to see how much you can achieve rather than trying to match a target.

As an example of the former, look no further than the "hyrbid tax" debate. The debate is ongoing and often angry, despite the fact that I have refused to address the issue for years, on the grounds that sharing our position doesn't actually seem to bring closure to any discussion. Rather it pours fuel on the fire. I think coming out with actual target numbers would do the same. Sharing our philosophy for how we do quest design or how we define interesting rotations can provide some behind-the-curtain insight that players seem to enjoy. Sharing numbers just seems to make people mad.

Now, don't get me wrong. You can't be a game designer without a healthy tolerance if not outright love for endless arguing. We just think there are plenty of aspects of the WoW design upon which to give feedback, and we don't want the conversation to be dominated by "Are the developers' target DPS numbers on target dummies fair?" Maybe that's the wrong call. Maybe we're being too conservative. I'm generally an "ask me anything" sort of guy. But it's also one of those rabbits you can't stick back in the hole once it hops out.

If you disagree or want to explore the topic further, feel free to create another thread. I'll read it.

Death Knight (Forums / Skills)
DKS -- Frost Fever is incorrectly using spell crit chance instead of melee crit chance, which will be fixed. Blood Plague already uses melee crit chance.

Anti-Magic Zone absorbs 20% less spell damage per attack power for Unholy. This multiplier was added back when Unholy got 25% attack power as a spec bonus. That spec bonus was dropped to 15% in the last beta patch. Can the Unholy-specific Anti-magic Zone multiplier be reduced to compensate?
The Anti-Magic Zone modifier was recently reduced by the same proportion. The intent, as you point out, is for them to stay in sync.

Can you please confirm this is intended/acceptable results, or otherwise address it? Or if I screwed up badly in my math somewhere I'd like to hear that as well. Dual Wield Frost - Howling Blast Vs. Obliterate
Yep, that sounds about how we expect.

Base damage for DK diseases were buffed this patch, but the numbers seem off a little bit. The base damage for Frost Fever is now 1390, while the base damage for Blood Plague is 172. Was an extra 0 mistakenly added to Frost Fever (should be 139)? Or is Blood Plague missing a 0 (should be 1720)?
The former. Frost Fever's base damage was accidentally increased too much. It should indeed be 139 at level 90.

Druid (Forums / Skills)
Is Feline Swiftness intended to stack with speed enchants?

Is it intended that spells cast in cat or bear form via Predatory Swiftness or Nature's Swiftness interrupt the swing timer?

Monk (Forums / Skills)
Did something change with the Elusive Brew calculations this patch, specifically with 2h weapons?
The tooltip updated, and some people are reporting getting 3 stacks (consistently) with a 3.6 speed weapon (the PvP polearm), which should be impossible given the formula we were given, while others are still getting 2.

We removed Tiger Strikes from Brewmasters and adjusted the proc rates of Elusive Brew and Gift of the Ox to account for the reduced number of normal attacks that they'll be doing.

What's up with Chi Wave? I spent a while on the dummies and it's doing about 2.5x much per chi as BoK and 1.2x as much as RSK, but in actual raid testing it's very inconsistent. Sometimes it will hit bosses 4x and sometimes it will just hit once and then disappear. I remember reading a blue post saying that the second tier of monk talents weren't intended to be used rotationally - is this still the case?

Right now it's a required talent and provides a 15-20% increase in DPS on bosses on which it bounces multiple times and is worthless (except as a small extra heal, although expel harm + healing spheres may be better options) for bosses on which it bounces only once. How is this ability supposed to work, and can it please be adjusted to work consistently?

There are some nerfs coming to the level 30 row, that should resolve this. It’s not intended that the level 30 row is a pure DPS increase for Windwalkers, but should provide a very strong option for trading a little damage for a lot of healing.

Paladin (Forums / Skills)
Execution Sentence is consuming Glyph of Flash of Light, but is seeing no increase in the healing done. Is it supposed to consume the effect and receive the increase or is it not supposed to interact with the glyph at all?
Ideally it would benefit from the glyph and consume the buff, but ES is a complex spell, so it's possible we'll have to exclude it. Sounds like a bug regardless.

Shaman (Forums / Skills)
A.) Stormstrike was buffed to 300% weapon damage, but Stormblast still appears to be set at 200%
Stormblast should be 300% as well. We will fix.

Is there a chance you might reduce the mana cost of Chain Lightning? We are running oom very fast on 2 targets and I honestly don't think it should be that way seeing as it is our spellcleave.
As Ele, Enhance or Resto?

Warlock (Forums / Skills)
Warlocks -- the next beta build will make it look like we nerfed Demo enormously. These are actually just adjustments because mastery was only giving 66% of the benefit it was supposed to in Meta form, which is now fixed.

Paladin (Forums / Skills)
Did WoG and EF change and somehow those changes didn't make it to the Patch notes? My values are down for it on just about every AP scale I've tested by between 3 and 7% (easily noticeable on the EF HoT tick)
The tooltip for EF is incorrectly not including its base healing (but I believe is functioning correctly). Regardless, EF's initial heal should exactly match WoG's heal.

EF is behaving properly from its tooltip. However, the PvP power conversion factor has changed. In previous patches, it was ~265. This build it's ~530, nerfed by exactly a factor of 2.
While there was a tooltip bug difference between EF and WoG (which should be identical), I think what you are seeing with the PvP Power conversion factor is that we straight out reduced the amount of healing provided by PvP Power.

It is important that PvP Power affect healing because otherwise healers wouldn't value the stat or the gear. However, unlike damage, healing is not offset by resilience, so stacking PvP Power just makes healers better and better. We knew this was a risk, but now the consensus is that healing is just much too good in PvP. We reduced the amount of healing provided by 50%. That seems like around the right number, but we may still iterate on it.

Oh, so PVP Power conversion percentage is still ~265 to 1%, but its effectiveness on healing spells has been reduced by half. Although I don't do PVP, I do wonder if there's a chance that the tooltips for WoG and EF will reflect the expected value based on both SP and PVP Power?

I don't know if you've seen our spreadsheet, but even with EF there are tooltip differences between what's expected and what's actually happening. Like for example, at 17630 AP, the EF tooltip indicates we should be expecting heals of 1024 from the HoT portion of EF -- however, due to PVP Power (16.05% in the character sheet in this case, and in reality much less since you have disclosed that PVP Power effectiveness for heals has been reduced), we're actually getting 1105 ticks.

I think we’re still missing something that you’re trying to tell us. PvP Power globally affects all healing (except percent based heals), so should be completely abstracted from any discussion about WoG or EF or any particular heal. If this is not what you’re seeing, we can investigate that. If so, can we simplify the discussion by completely leaving PvP Power out of it? PvP Power should be completely irrelevant to WoG vs EF or really any talent in that row, or priority list, or anything.

Class Feedback Blue Posts

Originally Posted by Blizzard Entertainment
Hunter (Forums / Skills)
Deterrence and Kill Command
We wanted to prevent Kill Command during Deterrence, because that felt like cheating, but it's a bug if it's stopping heals.(Blue Tracker / Official Forums)

Paladin (Forums / Skills)
Beta Retribution Feedback
It should really only work with 3HP finishers.
We're changing Divine Purpose so its chance scales with Holy Power. The 25% is with 3. You should be able to benefit from using a 1 Holy Power ability, but we don't want to encourage fishing. (Blue Tracker / Official Forums)

Crusader Strike Change Feedback
Paladins were balanced around generating Holy Power through Holy Shock, Holy Radiance and Tower of Radiance. We like the "holy warrior" kit and thought using Crusader Strike situationally for Holy was interesting. We didn't want it to be the most effective way to play a healing paladin, which is how it was trending, especially in heroic 25-player raids.

I agree, we were tuned with the idea in mind that CS was part of our rotation.
No, I'm saying you weren't. We calculated paladin healing to be at X. With using CS as a major part of the rotation, the actual level was potentially X + Y. We want you to be at X.

As I mentioned above, I'm not even convinced it was a serious problem. Yes, it's mana efficient to generate Holy Power for almost nothing, but you could have also generated that Holy Power with an actual heal. WoG / LoD don't heal for enough for that to be your entire arsenal. A paladin who casts HS, DL, DL -> LoD will handily outheal one who casts HS, CS, CS -> LoD, unless mana is a crushing issue, which we don't believe to be the case. Taken to extremes, a healer who never casts a heal has awesome mana longevity, am I right?

However, this is one of those cases where we didn't even want there to be a question. We didn't want there to be any confusion about our intent. You should generate 99% of your Holy Power with Holy Shock, Holy Radiance and Tower of Radiance.

Now we are left with ability's that are not on par with other healers. After the nerf our mana efficiency is now worse imo then a lot of other healers.
We haven't seen any evidence of this, but if it comes to pass, we'll fix it (and not by buffing CS).

I think they must have had us balanced around CS because the alternative is frightening: that there was a meta playstyle that was used extensively, talked about extensively on these very forums and others, and Blizzard did not attempt to balance around it. Does that sound like good design to you?
It wasn't extensive. Not many paladins were doing it in our raid tests. However some of the paladins who were doing it were very talented healers in top-end guilds, and those guys can sometimes be trendsetters. This was a precautionary change, not a crisis. (Blue Tracker / Official Forums)

Priest (Forums / Skills)
Halo vs. Holy Word: Sanctuary
We agree with the argument that Sanctuary and Halo are apples and oranges. However, Sanctuary is intended to have a 30 sec cooldown, not 40 sec. That's a bug. (Blue Tracker / Official Forums)

Shadow Orbs Bug
This was a bug caused by the new orb UI. We fixed it recently so the orbs / ravens should persist for 60 sec now. (Blue Tracker / Official Forums)

Warrior (Forums / Skills)
Beta Warrior Feedback
Heroic strike now competes for rage with slam, so if you're slamming with rage, many times you'll not have enough rage to hit heroic strike when TFB is about to expire. Which means if you want to guarantee that you have enough rage for the big heroic strike, you have to not slam and leave gcds open, which is boring. And you have to gamble constantly between trying to stack TFB with more overpower, or using the rage on a weaker heroic strike before the buff wears off.
I want to dive into this a little bit again, because I still think there's some sort of miscomm between us and you guys.

1) If you have 60 or more rage, you should be able to Slam and Heroic Strike. No problems there.
2) If you have less than 60 rage, you might have to choose between Slam and Heroic Strike.
3) If Taste for Blood has plenty of duration left, then go ahead and Slam. You'll probably get enough rage to Heroic Strike soon too.
4) If Taste for Blood is about to fall off, you're at the greatest risk. In this case, it may be worth not Slamming because you need to have 30 rage. Even in this case though, you shouldn't be sitting on empty GCDs. This is a great time to Storm Bolt or Dragon Roar or Battle Shout. If every one of those is on cooldown, then you might have to wait a GCD or two, but that should be pretty rare. As you get better gear (and honestly that might be part of what we're seeing here) it should be even more rare because you'll have more rage to Slam away.

So what am I missing here? I want to understand the feedback.

GC, just a quick question. Is this change trying to suggest that Fury warriors should never equip a shield to go defensive?
Swapping to a shield should not turn off the Single-Minded Fury damage bonus for an SMF warrior. It's possible a Titan's Grip warrior who swapped to a one-hander and shield will not get the Single-Minded Fury bonus. However, TG warriors can also equip a two-handed weapon and shield.

GCs arguments and justifications almost make it sound like he has some kind of personal vendetta against the warrior class and wants us to suck and not be invited to raids... never gives us a real answer to anything other than "working as intended" or "we'll see" but nothing ever changes unless its another nerf or new penalty added to an ability... if we show the slightest hint of getting better "it's a bug... here we'll fix it"
Yeah, that's probably it.

Execute may not be half of my total dps but its a huge chunk that im missing out on 80% of the fight. Ghostcrawler, I dont understand your reasoning sometimes on these issues, you wanted player feedback, its given, but you counter saying its not an issue. Why bother even continue reading our feedback if you already have a replay to shoot it down?
Think about it from my point of view. If every time players say "Please fix this" we then always go and fix it, then we're really not making the decision. We're letting players make it. The way we like to design is through informed decision making. We like to gather information and then make the call. That doesn't mean we're always going to make the call you want, but at least we'll know how you feel about it.

In this case, I wanted to see if there was something we hadn't considered about why high executes are bad. I didn't see many arguments we hadn't considered, but there was one we found potentially compelling. Essentially, if a group is stuck on phase 2 of a boss they might feel compelled to swap out a warrior, since their contribution is really going to come in phase 4. Now if phase 4 is also tough, you made a mistake, but if phase 4 is easy, then you may have made the right call.

In my humble opinion it's the same deal as some classes being amazing on burst or multidotters. Granted, sub 20% happens in most boss fights unlike multidotting, but it could be viewed as a perk. If a fight lasts 4-5 minutes Shamans and other burst classes are ridiculously strong, on council fights multidotters shine and then warriors can have the flavor of killing you if you're close to death.
That was more or less our logic. Warriors might be great when executes matter and weaker when executes are irrelevant, but those don't happen with such regularity that it would be a major problem. Groups tend to like the dot classes even on single target fights. They just excel at group fights.

The same response came with shield wall and spell reflect, people dont like it and yet, you completely ignore the feedback and reply; "we are happy with warriors requiring a shield" well, thats fine and dandy that your happy, but wouldnt you rather have the players who are using the warrior be happy?
I've tried to explain our shield logic there several times. I'm sure you could find the responses. Now you may disagree with out logic, but that's different than our ignoring the feedback.

Of course, reading between the lines in GCs responses, it looks like they may not have fully taken CD stacking during Execute phase into account, so there's also the possibility that Warriors are competitive for 80% of the fight and just flat OP for the final 20
No, I just wanted to make sure there was a good explanation for the high crit rate.

There's nothing wrong with the idea of having your damage more backloaded. I think the concern right now is that it is too backloaded. It's just been kind of funny to hit the last 45 seconds of a beta raid encounter and have an ability go from zero to #2 in WoL while accounting for 15% of my damage, all in ~45 seconds.
Yeah, many of the responses seem to be "it feels weird" and not "it's fundamentally broken" (even when they say the latter, it looks like they often mean the former). Now feeling weird still counts, but it's not as "must fix" as a critical design flaw.

I seem to remember one of your first suggestions being a short cooldown slapped on. Intuitively, that seems the best — Execute can be powerful enough to kill in PvP without crowding damage contribution in PvE.
We don't like Execute having a cooldown. You still need to do other attacks already to make sure you have enough rage to spend. If we decide it's a problem, we're more likely just to shift some of the damage from Execute back to other attacks. (Key word is "some" - we don't want it to be weak.)

We're not changing the shield requirements of Spell Reflect and Shield Wall. You can read my previous comments on them, Die by the Sword and Mass Spell Reflect. Our stance hasn't changed.

We are going to try reducing Execute damage by about 20% and redistributing the damage to yellow attacks. I'm not sure that nerf will be significant enough to make some of you happy (what an ironic thing to say) but it might be enough of a compromise between Execute feeling awesome but not so awesome that the rest of your attacks feel weak.

You would only use HS and Slam over 60 rage if HS has a TfB stack (or stacks) already because a no stack HS is less than a slam in termsof damage. Otherwise you would have to weigh out the near term benefit against filling a future GCD which might otherwise be empty with a slam. The problem is, you can't really predict the availabillity of TfB procs in the future - will you have an empty GCD, or will you have multiple overpowers? If you hold off on the HS, perhaps those additional OP's will buff the HS you saved to a higher degree - is it more, or less DPS. There is simply too much on-the-fly calculation and gambling involved. Furthermore, you dont really have to use HS to burn off rage to prevent overcap - one slam pretty much negates 3 white swings worth of rage. Again with the unpredictabillity - how will enrages change that decision?
Why would you ever use the HS until Taste for Blood stops proc'ing? You Overpower until it stops, then you HS. If you have 60 rage (or think you will soon) then you go ahead and Slam because you'll still have enough rage to Heroic Strike as well. It's the "on-the-fly calculation and gambling" part that I'm not quite understanding. If you're choosing to HS when you don't know if the proc is done, then I guess that's gambling, but why do that? That's like bleeding off all your rage when you don't know is a CS is about to proc. If you're choosing to Slam when that means you won't be able to HS, then why do that (unless maybe the target has to die right now)?

It seems to me that the two decision points are: Should I Heroic Strike and should I Slam?

Should I Heroic Strike? Is Taste for Blood at 5 stacks? Then, absolutely HS. If not, then is Colossus Smash up? Then it's still worth it to blow the stacks with an HS. If not, then will the buff expire before I get another stack? If so, then Heroic Strike. (It's off the GCD, so you don't even have to delay another ability.)

Should I Slam? Will Slamming mean not enough rage for an upcoming Heroic Strike? Then, no.

What am I missing?

SMF warriors being able to eat is an unintended consequence of the bullet-proofing we put in to stop TG warriors from swapping to SMF for Execute. Should be an easy fix.

1. Slam is better than HS at 1 or 2 stacks, so you should never HS at that amount of stacks.
Okay, I think this is part of the problem. On our current build, Heroic Strike is at 110% weapon damage and Slam is at 220% weapon damage. That means a one stack of Heroic Strike ties with Slam and two stacks on always wins.

3. With 60 rage banked up, if you get lucky Overpower procs you'll have to HS before you reach 5 stacks or you'll waste rage due to cap. This is also RNG.
This part I am not understanding. If you have enough rage, then Slam. Otherwise, wait for the 5 stack. When are you at rage cap but unable to Slam?

4. Now say you have 60 rage, a 5 stack TfB and no filler avaiable. You have to outweigh the short term benefit of Slam + HS vs the possibility that those 30 rage you spent could turn into a 3 or higher stack in the next MS cycle.
If you're at 5 stacks, you should Heroic Strike because it can't go any higher. The only reason to delay would be if CS is about to come off cooldown. You shouldn't have to pool rage for multiple Heroic Strikes. If you use one, you should be ready for the next one by the time it stacks -- you're talking about a MS and 3 Overpowers. I'm not sure what gear you're talking about -- is this the ungemmed PvP gear at level 90?

Maybe another way to ask this is what percentage of the time are you Slamming in between buffed Heroic Strikes, and what percentage are you double Slamming?

Again, it's all probably due to low rage generation. As Arms you generate 36-45 rage during each MS depending on swing timers, which is less than you might be required to spend.
Yeah, I am starting to wonder if that's the issue. If you don't have enough rage to Slam regularly, then the rotation is going to feel empty. That will get better with more gear, but it sounds like Fury's rage feels good already.

I'm with GC on this one. The choice is merely between heroic strike and slam when you have rage. Why are you guys so steeled on 'must fill all globals!'? One of the big problems warriors had was too many globals as arms, you had no time for any utility buttons.
I understand that some players just like filling all globals, and the fact that warriors always become rage flooded in the last tier of content exacerbates how off it feels going from heroic raid gear to quest gear for the new expansion. We still want to keep some sense of progression that better gear means more rage which means more attacks, but maybe that's just too hard to deliver. On the other hand, warriors already have the highest APM of any class when you consider they can fill almost every global and have several off-the-global abilities, including a legitimate attack.

I do appreciate those of you attempting to clarify your concerns and not just venting. It is helpful.

I was curious if there was any chance of getting all of our attacks to trigger auto-attack? Not an actual swing, but putting us in a ready combat position. Kinda sucks to hit bloodthirst and fall right back out of the attack mode.
Can you explain this in more detail? If I target something and hit BT, then I start autoattacking.

The problem involves being out of range, out of rage, facing the wrong direction, the ability being on cool down for 4.5 more seconds, or disarmed etc.
We made a change so if you are in range (on any class) you will start autoattacking even if the special attack you used didn't go off (if for example you have no rage). It also works if you are facing the wrong direction.

It will not start autoattack if you are out of range. That might be something we can fix.

I just tested this on Slam and Mangle, and it works as I described. If you can test beta and verify that this is not happening, let us know. This is a change from live. (Blue Tracker / Official Forums)

No comments: