Ghostcrawler returns to the conversation, talks about hit and expertise on tanking gear

This is great news for the community -- we'll have more of an insight into the design decisions happening at Blizzard, understand better why folks like Ghostcrawler are making the choices they are, and no doubt get that mystical pony.
The main point about Ghostcrawler returning, as written by him, is:
When we started these blogs, the idea was to foster developer communication to the players without some of the inherent problems of posting in forums. Some players have pointed out recently, and we totally agree, that the blogs up until now have been from a very high vantage point. We looked for topics with universal interest that would feel important and newsworthy. That has worked overall, but we also feel like we've lost something from when I used to be down in the metaphorical trenches talking to players in the forums.
So we're going to try something a little different. We're going to unleash some blogs that are much more conversational and less proclamational (that's a word now). If we deliver on this, it will hopefully feel like you're eavesdropping on our design meetings. You won't always learn a lot about exciting new features coming to the game, but you will (ideally) learn something about the design process itself. (When we have big, exciting news to share, or 'State of the Game' style blogs, we'll still do those as well.)
So we're going to try something a little different. We're going to unleash some blogs that are much more conversational and less proclamational (that's a word now). If we deliver on this, it will hopefully feel like you're eavesdropping on our design meetings. You won't always learn a lot about exciting new features coming to the game, but you will (ideally) learn something about the design process itself. (When we have big, exciting news to share, or 'State of the Game' style blogs, we'll still do those as well.)
Ghostcrawler's full post, along with all the tanking discussion, after the break. And for the record, his post is titled "Coffee with the Devs." Just remember, he likes gin in his coffee (although what kind, we don't know ... yet).
So how is the view from way up here? It's great actually -- we're really happy with how Cataclysm is going so far, and we have big surprises on the horizon. On the other hand, there are details you can see at ground level that you can't make out from 10,000 feet.
When we started these blogs, the idea was to foster developer communication to the players without some of the inherent problems of posting in forums. Some players have pointed out recently, and we totally agree, that the blogs up until now have been from a very high vantage point. We looked for topics with universal interest that would feel important and newsworthy. That has worked overall, but we also feel like we've lost something from when I used to be down in the metaphorical trenches talking to players in the forums.
So we're going to try something a little different. We're going to unleash some blogs that are much more conversational and less proclamational (that's a word now). If we deliver on this, it will hopefully feel like you're eavesdropping on our design meetings. You won't always learn a lot about exciting new features coming to the game, but you will (ideally) learn something about the design process itself. (When we have big, exciting news to share, or 'State of the Game' style blogs, we'll still do those as well.)
But to pull off this more casual blog style, let's establish a few ground rules:
1) No promises. I'm going to be talking about a lot about things we might do or things we could do. You shouldn't interpret this brainstorming as patch notes. Our creative process is insanely iterative. We might pitch dozens of ideas before we find one we like. That can be really exhausting if you're not used to it. If you're more interested in final decisions and not idea churn, then this style of blog won't be for you.
2) Don't read too much between the lines. I'm going to point out a lot of design flaws in our game. "Oh no! Goatcaller admitted WoW was deeply flawed! It's shark-jumping time!" Look, Blizzard is very critical about our own designs. There is virtually nothing in World of Warcraft that could not be improved. That has always been the case and will continue to be the case. Just because I'm going to be sharing that more frankly with you doesn't mean that the game now has more cracks in its foundation than it ever did. There is an old saying (misattributed, from what I understand, to Otto Bismarck) that laws are like sausages; it is best not seeing them being made. My old friend and mentor Bruce Shelley used to apply the same maxim to game design.
3) No complaints about the topic. If we didn't have an interesting discussion about a topic recently, e.g. shaman mechanics, I'm not going to invent one. That doesn't mean that the class is perfect, or that we don't love shaman players, or that the shaman class has no direction, or that the class design is frozen in carbonite. I'm not going to keep hash marks next to every class and spec to make sure I've covered their "Very Important Issues" lately in a blog. World of Warcraft design being what it is, we'll probably eventually get around to talking about everyone on here, but it may take weeks or months or years. My team is responsible for areas of the game including classes, items, encounters, trade skills, achievements, combat, and UI, so my blogs will probably stick to those topics.
Okay, all that preamble is out of the way now. I'll probably refer back to it sometimes, if we have some players stomping all over the ground rules.
One topic we've been discussing lately is the role of Hit and Expertise on tank gear (or more precisely, plate tanking gear). The conventional wisdom is that Hit and Expertise are threat stats, and you may need to swap them out with some of your mitigation stats depending on the situation. Realistically, unless you severely overgear the content, we don't think that is actually true. Tanks almost always worry about survival first and foremost, which totally makes sense, and are willing to trade off threat stats for better mitigation in almost all situations. It's much harder to progress if the tank explodes than it is if the cat occasionally pulls aggro. (It's not quite that simple, but I'm going to gloss over details and exceptions since I spent so much text on the preamble up above).
Once upon a time, taunts could miss, and so Hit was marginally more interesting than it is today. Once upon a time, having a boss parry your attacks could speed up its swing timer, which turned Expertise into a (often weak) survival stat. Boss parries felt very random though, both in the sense that sometimes the tank would suddenly take much more damage than anticipated and there was no easy way to know which bosses had parry speed up. (Today, you can assume none of them do.) Until recently, interrupts could miss, but asking a tank to stack a bunch of Hit just for those few opportunities when they were probably going to hit anyway but disaster would occur if they did not felt crummy too.
The problem is that there aren't a lot of stats that are interesting to tanks. Stamina and Armor are great, but their stat budget is often in lockstep with item level. (It would be interesting to consider if we could make that not the case once again, but that's the topic for another blog.) We got rid of Defense as a stat that tanks needed to worry about. We have managed to make Mastery pretty good to excellent for tanks, so that's at least one stat they like to see. Dodge and (if you're a plate-wearer) Parry are good, and slightly interesting because of talents like Hold the Line. But beyond that, it starts to go downhill. Sure Haste and Crit can sometimes be fun, but really they often aren't worth the trade off. That leaves us with Hit and Expertise. We'd like to make them more interesting to tanks. But how?
One way is by turning them into defensive stats. They are defensive stats for Blood death knights, because the DK self-healing is tied into Death Strike, which can miss. It might be possible to do something similar for the other classes. Imagine if Shield Block had to actually hit the target. Presumably you raise your shield, but not high enough to intercept the incoming blow. Now hit becomes a mitigation stat for warriors as well. We might have to adjust the mitigation amount on Shield Block or give warriors a small Hit bonus so Hit capping wasn't totally unreasonable, but you get the basic idea. You could do the same with paladins (make Holy Shield more interesting?) and druids as well (Savage Defense could proc on a hit).
Is this a good idea? We're not sure yet. You won't see this change in the 4.1 patch for certain. There are trade-offs to making Hit and Expertise more valuable. Gearing as a tank might be more fun for experienced players, but it also might be more challenging for less experienced players. The number of struggling tanks in your Dungeon Finder groups might go up. Some less knowledgeable players (and to be fair, this stuff doesn't exactly explain itself on the character sheet) might stack Hit way too high at the expense of a more valuable mitigation stat, such as mastery.
It is the kind of thing we're talking about though, and if you want to make a contribution to the tanking forums but aren't quite sure on a topic, here is one potential possibility.
Greg "Ghostcrawler" Street is the lead systems designer of World of Warcraft. He still has Buru's Skull Fragment.
When we started these blogs, the idea was to foster developer communication to the players without some of the inherent problems of posting in forums. Some players have pointed out recently, and we totally agree, that the blogs up until now have been from a very high vantage point. We looked for topics with universal interest that would feel important and newsworthy. That has worked overall, but we also feel like we've lost something from when I used to be down in the metaphorical trenches talking to players in the forums.
So we're going to try something a little different. We're going to unleash some blogs that are much more conversational and less proclamational (that's a word now). If we deliver on this, it will hopefully feel like you're eavesdropping on our design meetings. You won't always learn a lot about exciting new features coming to the game, but you will (ideally) learn something about the design process itself. (When we have big, exciting news to share, or 'State of the Game' style blogs, we'll still do those as well.)
But to pull off this more casual blog style, let's establish a few ground rules:
1) No promises. I'm going to be talking about a lot about things we might do or things we could do. You shouldn't interpret this brainstorming as patch notes. Our creative process is insanely iterative. We might pitch dozens of ideas before we find one we like. That can be really exhausting if you're not used to it. If you're more interested in final decisions and not idea churn, then this style of blog won't be for you.
2) Don't read too much between the lines. I'm going to point out a lot of design flaws in our game. "Oh no! Goatcaller admitted WoW was deeply flawed! It's shark-jumping time!" Look, Blizzard is very critical about our own designs. There is virtually nothing in World of Warcraft that could not be improved. That has always been the case and will continue to be the case. Just because I'm going to be sharing that more frankly with you doesn't mean that the game now has more cracks in its foundation than it ever did. There is an old saying (misattributed, from what I understand, to Otto Bismarck) that laws are like sausages; it is best not seeing them being made. My old friend and mentor Bruce Shelley used to apply the same maxim to game design.
3) No complaints about the topic. If we didn't have an interesting discussion about a topic recently, e.g. shaman mechanics, I'm not going to invent one. That doesn't mean that the class is perfect, or that we don't love shaman players, or that the shaman class has no direction, or that the class design is frozen in carbonite. I'm not going to keep hash marks next to every class and spec to make sure I've covered their "Very Important Issues" lately in a blog. World of Warcraft design being what it is, we'll probably eventually get around to talking about everyone on here, but it may take weeks or months or years. My team is responsible for areas of the game including classes, items, encounters, trade skills, achievements, combat, and UI, so my blogs will probably stick to those topics.
Okay, all that preamble is out of the way now. I'll probably refer back to it sometimes, if we have some players stomping all over the ground rules.
One topic we've been discussing lately is the role of Hit and Expertise on tank gear (or more precisely, plate tanking gear). The conventional wisdom is that Hit and Expertise are threat stats, and you may need to swap them out with some of your mitigation stats depending on the situation. Realistically, unless you severely overgear the content, we don't think that is actually true. Tanks almost always worry about survival first and foremost, which totally makes sense, and are willing to trade off threat stats for better mitigation in almost all situations. It's much harder to progress if the tank explodes than it is if the cat occasionally pulls aggro. (It's not quite that simple, but I'm going to gloss over details and exceptions since I spent so much text on the preamble up above).
Once upon a time, taunts could miss, and so Hit was marginally more interesting than it is today. Once upon a time, having a boss parry your attacks could speed up its swing timer, which turned Expertise into a (often weak) survival stat. Boss parries felt very random though, both in the sense that sometimes the tank would suddenly take much more damage than anticipated and there was no easy way to know which bosses had parry speed up. (Today, you can assume none of them do.) Until recently, interrupts could miss, but asking a tank to stack a bunch of Hit just for those few opportunities when they were probably going to hit anyway but disaster would occur if they did not felt crummy too.
The problem is that there aren't a lot of stats that are interesting to tanks. Stamina and Armor are great, but their stat budget is often in lockstep with item level. (It would be interesting to consider if we could make that not the case once again, but that's the topic for another blog.) We got rid of Defense as a stat that tanks needed to worry about. We have managed to make Mastery pretty good to excellent for tanks, so that's at least one stat they like to see. Dodge and (if you're a plate-wearer) Parry are good, and slightly interesting because of talents like Hold the Line. But beyond that, it starts to go downhill. Sure Haste and Crit can sometimes be fun, but really they often aren't worth the trade off. That leaves us with Hit and Expertise. We'd like to make them more interesting to tanks. But how?
One way is by turning them into defensive stats. They are defensive stats for Blood death knights, because the DK self-healing is tied into Death Strike, which can miss. It might be possible to do something similar for the other classes. Imagine if Shield Block had to actually hit the target. Presumably you raise your shield, but not high enough to intercept the incoming blow. Now hit becomes a mitigation stat for warriors as well. We might have to adjust the mitigation amount on Shield Block or give warriors a small Hit bonus so Hit capping wasn't totally unreasonable, but you get the basic idea. You could do the same with paladins (make Holy Shield more interesting?) and druids as well (Savage Defense could proc on a hit).
Is this a good idea? We're not sure yet. You won't see this change in the 4.1 patch for certain. There are trade-offs to making Hit and Expertise more valuable. Gearing as a tank might be more fun for experienced players, but it also might be more challenging for less experienced players. The number of struggling tanks in your Dungeon Finder groups might go up. Some less knowledgeable players (and to be fair, this stuff doesn't exactly explain itself on the character sheet) might stack Hit way too high at the expense of a more valuable mitigation stat, such as mastery.
It is the kind of thing we're talking about though, and if you want to make a contribution to the tanking forums but aren't quite sure on a topic, here is one potential possibility.
Greg "Ghostcrawler" Street is the lead systems designer of World of Warcraft. He still has Buru's Skull Fragment.
Filed under: News items






Reader Comments (Page 1 of 5)
(cutaia) Mar 23rd 2011 1:41PM
They might need to include those rules at the beginning of each post. Aw, who am I kidding...the complaining is gonna happen regardless.
More importantly: I'm glad to see more usage of the word "Goatcaller."
Knob Mar 23rd 2011 2:11PM
Yeah, there's a startling lack of goats in Azeroth, and the more goatcallers we have the better it'll be.
Rortek Mar 23rd 2011 7:55PM
Goatcaller promised me a goat
Joakim Mar 24th 2011 7:14AM
The reason there are too few goats in Azeroth is that Dormus the Camelhoarder finally got tired of getting beaten up for hoarding camels - wich everyone obviously love - and changed his hobby to goats.
Right now there's a lot o angry dwarves in Dun Morogh, but hey: Wha can you do, huh?
Terrant Mar 23rd 2011 1:42PM
Adam: Best WoW Insider tag ever.
Alf Mar 23rd 2011 1:46PM
"That leaves us with Hit and Expertise. We'd like to make them more interesting to tanks. But how? One way is by turning them into defensive stats.It might be possible to do something similar for the other classes. Imagine if Shield Block had to actually hit the target. Presumably you raise your shield, but not high enough to intercept the incoming blow. Now hit becomes a mitigation stat for warriors as well. Is this a good idea?"
NO GHOSTCRAWLER, ITS NOT A GOOD IDEA
Darky Mar 23rd 2011 2:44PM
Welcome to the world of blood tanking.
VSUReaper Mar 23rd 2011 3:21PM
Actually, I think that it has merit... It just all depends on how it is executed.
I think expertise should not change that much, but there should be a penalty for having your attacks dodged or parried, just not what it was in Vanilla-WOTLK: parry hastened blows. It needs to be a small penalty, but not so small that it can be ignored; perhaps something to do with our next attack or resource.
I like the idea with hit, but with a small change. Dont make it so much a mitigation stat like GC described, but make it so that when a shield tank has to block, it checks 2 numbers: your hit rating and your block rating (via mastery). Hit determines if we can actually block the attack (just like hit determines if you actually hit) and then it checks the block chance to determine how much you block.
Kaphik Mar 23rd 2011 3:30PM
As a warrior tank I love this idea!
adamjgp Mar 23rd 2011 4:18PM
I'm concerned about this. I am a warrior tank, and I stack mitigation stats exclusively. If there is hit/expertise on an item it is always reforged.
That being said, if they make being hit/expertise capped a virtual requirement, that's going to eat into my avoidance/mitigation stats. Tanks are already incredibly gear dependent. A lot of onus is put on tanks to not be squishy, and currently a tank needs to have better gear than dps or heals in order to be effective. I know that vengeance helps with threat, but we still go SPLAT w/o enough avoidance/mitigation.
Just concerned that I won't have enough gear slots in order to be an effective tank. Sure the guys in full 359/372 won't have a problem, but those of us in 346 gear are going to get roflstomped. I don't want to come across as 'the sky is falling', because there aren't too many details about how this will be implemented, but i really don't like this proposed change.
Squatch Mar 23rd 2011 4:54PM
@Alf: Can you please elaborate? Why do you not think it's a good idea? Yelling in caps that it's not a good idea isn't necessarily the most constructive way to say you disagree lol.
@adamjgp: I have a tank that's in all iLevel 346 gear with a couple of iLevel 359 epics and my strategy has been to cap Hit and Expertise simply because it's a hard cap that I know I can reach. Yeah yeah, I know, days of capping something for tanks went away with Wrath, forgive me if old habits die hard. However, my OCD has been satisfied with this approach and so far, I've polled every healer I've run with (heroics, no raiding on my tank yet) to see if they think I'm squishy in general and so far all of them have been quite happy with healing me. So, just wanted to let you know that I think your concerns are unwarranted as far as running heroics are concerned. My overall strategy has been:
1) Stack hit and expertise to caps (I'm over 26 expertise at the moment because I find threat gen an issue against raid geared DPS and as a pally, it's an excellent threat stat).
2) Stack as much mastery as possible.
3) Balance Dodge/Parry so neither suffers more than the other with diminishing returns.
Gemming: All my gem-sockets have stamina in them while picking up socket bonuses (so red sockets have Sta/Exp, yellow have Sta/Mastery, and blue sockets are pure Sta).
Maybe it's not the best tank gearing/gemming strategy out there, but it's straight-forward enough for me to grok pretty well and I've found that my original concerns with this strategy yielding a particularly squishy tank were completely unfounded.
Hope this info helps!
pancakes Mar 23rd 2011 5:25PM
I'd say getting hit/exp capped for heroics is probably a good idea, since the bosses don't hit hard, and there's a lot of trash, so vengeance won't stack high enough to provide meaningful threat.
However, at the raiding level, threat stats lose value over time as the AP from vengeance gets high enough to offset misses &c. If they're less valuable than stats providing combat table coverage, why would you keep them on gear?
The reasons it's a bad idea IMO is that it will mean tanks will have another cap needed to maximise mitigation, it will devalue mastery until you're hit capped, as well as moving back into the era of ridiculous spikes in damage that they deliberately tried to get away from after wrath.
Pfinferno Mar 23rd 2011 10:38PM
@Squatch: I can say from personal experience, the difference between raid necessity and heroic necessity in Cataclysm is HEAVILY noticeable. I had healers praising my survivability in heroics, and then when we started raiding, I had to reforge and even revamp my spec because my stat weighting wasn't working. I tried to go for the hit cap for a while to pump some threat into bosses, but bottom line is the mitigation stats are far too necessary to reforge them into hit and/or expertise. This is more because of the new healing style than tanking in general. Healers need to be very smart with mana. Wasting mitigation stats on hit or expertise means more mana spent in raids keeping the tank alive. You can get the threat without necessarily being hit-capped. You can't survive a boss fight if you're never mitigating.
Skrotus Mar 24th 2011 2:45AM
Really I'd love an excuse to actually go for the hit/exp caps, missing with abilities just feels sloppy, but currently the tradeoff is too big for too little gain.
bella Mar 29th 2011 10:28AM
@Squatch - Do you use CC in your heroics? A 333 geared tank doesn't take much dmg when there is only 1-2 mobs hitting them at a time. (Lemme say this before I get flamed about CC - It's not a bad idea for CC. I've just encountered more DPS and healers bitching and moaning when there is CC involved. So personally, I've stopped using it 99% of the time. I take ALOT more dmg than before but my mitigation and health has balanced that out.)
I don't think that exp or hit are bad stats, I've always had the thought that as our gear increases the need for exp and hit will increase. Vengeance can only do so much for threat once the DPS is pulling double or triple what they are now. At that point we WILL need exp and hit to do our jobs correctly. Until then, meh. Why screw with somethin that ain't broken imo.
MattKrotzer Mar 23rd 2011 1:48PM
Man, I really hope they stick to the idea of tying block to hit rating as well. It would be great to go to a triple roll system.
Hit check to see if you raise your shield high enough.
Block check to see if you block it.
Critical block check to see if you block for twice as much.
That doesn't seem at all obnoxious...
As it is now, we already have a block % tied to mastery. Trying to force hit into the equation is just silly.
Cheeselandman Mar 23rd 2011 1:57PM
Agreed. Personally, I think tanking is pretty good- when you think about it, dps classes have to worry about hit, expertise, and often one other primary stat. Dps classes may like haste or crit or mastery, but there tends to be one that is more valuable than the other. This leads to essentially the same three stat system that tanks have.
My point is, tanks still have plenty of choice with gearing.
Chokaa Mar 23rd 2011 2:06PM
Matt - I must have read your post like 3 or 4 times before I realized the sarcasm. All those D&D roll check talks in the past. I was like 'yea, triple roll sare great for a tabletop but come on, dude. NOT IN M WARCRAFT!
MattKrotzer Mar 23rd 2011 2:18PM
Avoidance is complicated enough for most players to understand in a meaningful way. I really hope that they don't force hit and expertise down our throats again.
However, as the next tier looms, I'm skeptical that Vengeance alone will be enough to maintain threat over the DPS when they get further upgrades. Currently, tanks can and do forego any hit/expertise, without any substantial penalty.
Hal Mar 23rd 2011 2:52PM
I thought the entire point of Vengeance was that it was supposed to provide tanks with sufficient threat, particularly once DPS stats start ridiculously outmatching that of tanks.
It does seem they've created a problem with the tanks not really "needing" threat stats. Wouldn't it be easier to change the scaling of Vengeance so that hit/expertise are more valuable for threat?