Shifting Perspectives: Why effective health needs to die, part 2

We've discussed in the previous column why effective health is important; here, we're going to discuss why it's not as important as you might think if you had nothing beyond the collective opinion of the Tanking forum to go on. While this has something to do with the mob mentality of the forums themselves, it has more to do with how the concept of effective health isn't usually placed in context. Tank death on hard modes is quickly attributed to EH discrepancies, with rather less discussion on encounter mechanics, inappropriate gear, or that great but frequently unacknowledged bugaboo -- player error. It is for this that I say effective health needs to die.
What is effective health?
I neglected to put some hard numbers on this in the last article, but calculating base effective health is actually pretty simple. It's your health as modified by the damage you'll take after armor contribution (AC), or Health / (1 - AC%).
A 50,000 health tank with 25% armor contribution has 66,666.67 effective health (50,000 / 0.75).
A 50,000 health tank with 50% armor contribution has 100,000 effective health (50,000 / 0.5).
A 50,000 health tank with 75% armor contribution (the maximum functional AC) has 200,000 effective health (50,000 / .25).
One of the things you'll notice is how quickly each additional point of armor makes each point of health keep you alive longer. If you were facing bosses who hit for exactly the same amount on exactly the same swing timer across three tiers of raid content, and the only stat you were able to improve was armor, you'd still be significantly easier to heal at the end of it. Armor improvements make each hit fall for a smaller and smaller percentage of your overall health. In other words, a boss that started off hitting you for, say, 30% of your life in the first tier of raid gear would wind up hitting you for 20% at the third. This turns you from a tank who could be four-shot into a tank who could be five-shot, and the difference in survivability is one healers invariably notice.
So what's the problem?
The problem is that these bosses don't actually exist, that EH is affected by more things than just armor and health, and that what's killing tanks in modern raids has less to do with EH than it ever has. Effective health matters most versus bosses that dish out tremendous melee damage and are not likely to kill you by other means. While this does describe a subset of difficult Wrath bosses, it's a far more accurate descriptor of their Burning Crusade counterparts.
With a 15% chance per each melee hit to land a 150% damage attack (the crushing blow), the BC raid boss was to built to deal devastating physical hits, and tanks were built around the ability to avoid or absorb them. If a boss blew through a warrior's Shield Block charges or caught a paladin without Holy Shield up, you had to live through the 15% chance that the blow/s that followed would crush. Versus bosses with a fast swing timer (Morogrim Tidewalker perhaps being the most famous among them), the odds of multiple crushing blows occurring to a warrior tank despite his/her best efforts approached 100%. And druids, as we've observed, were nothing more than walking piles of effective health in an era that obliged them to eat crushing blows as a matter of course. The tank that needed EH the most was (unsurprisingly) the tank that had the most of it.
The "crushing blow" of Wrath
Blizzard was forced to move away from this model by the introduction of the death knight, which -- as a 2H-weapon tank -- could not be designed to avoid crushing blows or mitigate them without creating serious balance problems. In the place of the crushing blow, developers increasingly began to program boss attacks reliant on enormous burst damage that could only be reliably survived through the use of cooldowns. A max-rank Sartharion-3D Flame Breath, hard-mode Mimiron Plasma Blast, Surge of Darkness on hard-mode Vezax (where he is not typically kited), heroic Gormok-25's Impale, and heroic Anub'arak-25's Freezing Slash are abilities universally cited by raiders as being among the greatest contributing factors to tank death. A huge amount of effort and time goes into ensuring that your tank survives these abilities, because nothing else you do matters if the tank gets one-shot or dies before your healers have time to react.
Not surprisingly, these encounters are just as frequently cited as the fights where the class of tank makes/made the greatest difference. Blizzard did indeed nerf both the death knight and bear after observing that the former's cooldown-based tanking dovetailed a little too neatly into the array of encounters designed with tank cooldowns in mind, and the latter because druid health pools scaled too well.
However, none of the listed attacks furnish particularly compelling arguments in favor of the supremacy of EH:
- Flame Breath and Plasma Blast are both magic-based attacks that ignore armor entirely.
- Surge of Darkness on a hard-mode Vezax at then-current levels of gear meant that a tank with high EH but no cooldowns simply died with less +overkill.
- Impale is a stacking bleed effect that ignores armor.
- Freezing Slash is an interesting case. Tank survivability is undeniably affected by EH here, as the stun immediately drops tank avoidance to 0% for its duration. While the druid's EH advantage is usually cited as the reason for the bear's use as MT on this fight (and there is some truth to that), I think people are missing the forest for the trees. The druid's sole damage-reduction cooldown (Barkskin) is usable while incapacitated or stunned, which (to my knowledge) is the only tanking cooldown of all four classes usable under those circumstances. The druid owes its advantage on the fight not to EH superiority but to the fact that our sole damage-reduction CD has been appropriated from an ability initially designed for use in PvP. Otherwise, the Anub'arak encounter actively punishes a raid with high-HP tanks, with Leeching Swarm causing more healing to Anub'arak in phase 3 and ratcheting up the already-high raid DPS required to kill him.
If the stuff that's most likely to kill you doesn't have much to do with EH, why does everyone make such a big deal out of it?
Because EH as players typically use it is a simple answer to a complicated question -- "How did I really die?"
The underlying issue with the version of EH as discussed on the forums is that EH is presented as a static set of numbers showing that bears are "best" and warriors are "worst." The truth of the matter is that a tank's actual EH at any given second exists in an ill-understood miasma of talent choices, cooldowns, trinkets, healer procs, and boss debuffs, set against the backdrop of bosses who hit for variable amounts at variable speeds with equally variable special abilities:
Do you have a priest or a shaman on you constantly to keep generating Inspiration or Ancestral Fortitude procs?
10% less physical damage taken is a mandatory tank buff on hard-mode content, period.
Does your DPS warrior or warlock reapply Improved Demoralizing Shout or Improved Curse of Weakness to the boss religiously (if the tanks can't apply an improved attack power reduction themselves)? Are debuffs like Improved Thunder Clap or Infected Wounds up at all times?
While Improved Demoralizing Shout was once part of the standard Protection warrior build, most people eschew it nowadays in favor of a Deep Wounds build for threat. Bears are able to pick up Feral Aggression a little more easily (albeit, like the warrior, at a significant cost to threat production), but they are not exactly common in raids these days. Thus, equally mandatory boss debuffs often fall to DPS players, who -- let us be frank -- are not always religious about applying and reapplying them.
Does your hunter keep Scorpid Sting up? Did you moonkin unglyph Insect Swarm?
While an additional 3% avoidance is not a contribution to EH per se, it is a contribution to Time To Live. The higher the percentage of your health a boss hits for, the more that avoidance will play a role in whether you live or die.
Who did your paladin Beacon? Does the Discipline priest know what he's doing? Do you have a Discipline priest and/or Holy paladin? Are tanks constantly Earth Shielded by shamans and "hotted" by druids? Is there a Val'anyr in the raid, and is that player assigned to the tanks on encounters like Gormok?
That you need dedicated tank healers with a high level of throughput on single targets and the ability to prevent/shield against incoming damage goes without saying, or should. Limited healing resources in the raid need to be deployed intelligently on encounters like heroic Northrend Beasts.
What cooldowns do you have up? Do they mitigate physical or magical damage, or both? How frequently can they be used, and what kind of dent do they make on the incoming hits?
Not all tank cooldowns are created equal, and different classes have wildly variable means of impacting their EH through the use of one. I don't think it's a mistake that the tank with the highest average EH (the bear) is the one with the least EH from cooldowns. After writing a great deal on this, it became apparent that it merits a separate article, and I'm going to push that to next week (although it may not necessarily be a Shifting piece).
Rather than asking these questions, it is easier -- though less honest -- to say that a tank died because they couldn't stack sufficient EH. Now, it has to be said that more EH does help on encounters where the boss hits like a freight train on top of dishing out special abilities (Gormok would be the standout example here), but it's disingenuous to argue that you're incapable of tanking an encounter that has been successfully tanked by others playing your class elsewhere. It's a roundabout way of saying that your raid shouldn't be forced to play more intelligently to get past a hard-mode encounter, and most reasonable players are unlikely to see that as a sympathetic argument.
I'll be continuing this next week with a look at cooldowns' impact on EH.
Every week, Shifting Perspectives treks across Azeroth in pursuit of truth, beauty, and insight concerning the druid class. Sometimes it finds the latter, or something good enough for government work. Whether you're a Bear, Cat, Moonkin, Tree, or -- for some unaccountable reason -- stuck in caster form, we've got the skinny on druid changes in patch 3.3, a close look at the disappearance of the bear tank, and thoughts on why you should be playing the class (or why not).Filed under: Druid, Analysis / Opinion, Features, Raiding, Classes, (Druid) Shifting Perspectives






Reader Comments (Page 1 of 4)
paul Nov 27th 2009 4:25PM
Allison, is it wrong that I get this giddy about your articles? I think I need help :/
tim Nov 27th 2009 5:50PM
I feel the same way, definitely my favorite writer here, though the new warlock Leviathan is doing well in his own right.
One thing I wish though, is that Allison got into the underlying problem with mathematical models like EH. Though EH is fairly new, it draws its roots from the idea that random is bad (something very popular in the earlier days of DPS theorycrafting). While the avoidance of randomness (no pun intended) may make more sense for a tank, I think the overemphasis on such models ignores that fact that
1. WoW is a game with chance variation at its very CORE (variable damage, glancing blows, dodges, parries, blocks, crits, etc.).
2. Encounters are designed with this chance variation in mind. Blues have not only hinted at this but stated it outright.
So while I don't think that EH "needs to die," good tanks who understand it need to spread the word that EH is a theoretical model which, depending on the fight, may or may not paint the appropriate picture on a particular tank's performance.
Heilig Nov 27th 2009 4:29PM
Great article. The long and short of effective health is that there is no current content where tanks are dying because they are getting hit too hard by melee swings, which is really all effective health will measure. You don't die on Gormok from his melee. You die from a combination of his melee + the stupid high bleed damage. You die on Vezax because he's designed to kill you if you stand there with no cooldowns while he smacks you for 35K+ every second or so. You die on Sarth from a 70K Flame Breath. None of these is a melee boss. Those days are over, and so are the days of effective health.
I for one welcome our new cooldown overlords. It makes tanking much more interesting, although Chill of the Throne is going to put a little bit higher priority on EH over avoidance for physical survivability now.
William Nov 27th 2009 4:50PM
"You don't die on Gormok from his melee. You die from a combination of his melee + the stupid high bleed damage."
That sentence is self-contradicting, much like the article itself. If you died to a combination of melee and bleed, and since the bleed is unavoidable, then the melee is what finished you off.
CD usage is whole different story altogether. It's class/raid dependent and always important regardless of how much EH or avoidance you have.
Allison Robert Nov 27th 2009 5:26PM
To William:
Gormok does a crap ton of melee damage, but short of the bleed ability, no geared or experienced tank should die only to that. His melee isn't all that terribly different from the amount of health you'd expect to see lost on the Patchwerk offtanks before you outgear the encounter. It's not burst damage, it's just high damage, and healable damage at that. If a tank is dying to Gormok's melee before the bleed even comes into play, then one of two things is wrong:
a). The tank isn't really geared for the encounter, or:
b). You don't have a tanking problem on your hands, you've got a healing problem.
Now, after the bleed becomes an issue, the raid logs may well show a tank death to Gormok's melee, but it's still Impale indirectly killing the tank, in much the same way that tanks could die to pre-nerf Gurtogg's melee damage, but that's not what was really killing them.
William Nov 27th 2009 5:44PM
@Allison
You are right that clearly a tank shouldn't die "only" to the bleed ability. However, what a tank does die to is the massive spike of damage that occurs when unfavourable RNG lines up 3 of Gormok's abilities occur within 1-2 seconds of each other (Impale DoT + Impale Hit + Melee swing). That's a near instant spike of upwards of 50K HP on a TOC-25 geared tank. The bleed itself is not burst damage, but the combination of his 3 hits together very much is. Since the Impale DoT + Impale Hit are unavoidable, the only way to survive is to have enough HP left to make sure the melee swing doesn't finish off the tank. In essence, it is very much the melee hit that kills the tank, which makes stacking stamina (EH) and/or using CDs mandatory.
Allison Robert Nov 27th 2009 6:38PM
To William:
I think you misread that, or that it wasn't otherwise clear (rereading it, I'll just say I wasn't that clear). No tank should be dying solely to Gormok's melee damage, period. If they are, that's a sign that the tank isn't geared for the encounter or that there's something going wrong with the heal team or what have you. A geared and experienced tank on that encounter can die to a combination of Gormok's melee and Impale ticks as you describe, but I think you're undervaluing avoidance somewhat here as well. Decent avoidance isn't necessarily about avoiding that one hit that would otherwise have killed you -- it's about avoiding the chain of hits taken that forms the 50K health/2 second scenario you describe. To say that EH is the *only* means of surviving this is somewhat misleading, when a geared tank doing ToGC-25 is extraordinarily unlikely to be rocking less than 50% avoidance.
Where EH comes into play is a scenario Ciderhelm's previously described -- which is to say that over the course of a 3-minute or longer boss fight, the odds of hitting at least one bad avoidance streak are basically 100%. Ideally this doesn't happen in the midst of a nasty Impale tick, but sometimes it will. That's when EH is important, whether it's in the form of what you had walking into the encounter (armor/health, although -- as we've noted -- Impale ignores half of that) or the temporary form it assumes by way of a cooldown.
Heilig Nov 27th 2009 6:45PM
You're agreeing with us, William. Yes, it makes CD's mandatory. That's the entire point. The Impale and the bleed from the Impale are unavoidable and the DoT is not affected by armor. The only part of your equation that utilizes effective health is the melee hit and the bad luck of possibly having the Impale land near a melee swing, which is a VERY rare occurrence. When he has 4 stack of Rising Anger, you either have a tank that stacks stamina (not armor, not effective health, but raw stamina) or you blow cooldowns. Regardless of which you choose, the strategy for tank survival in that encounter is not to stack effective health.
Theck Nov 30th 2009 9:19AM
"When he has 4 stack of Rising Anger, you either have a tank that stacks stamina (not armor, not effective health, but raw stamina)"
You and Allison are both missing the point of Effective Health, and using the term incorrectly.
The true definition of Effective Health is "the amount of raw damage a creature has to deal to kill you." Most mathematical definitions (see Satrina's work for example) have made the assumption that this is physical damage, and formulate their derivations using Armor as the primary mitigation stat. This is because when those derivations were performed in Burning Crusade, physical damage was usually the type of damage that caused spike deaths.
However, nothing in the definition of EH says that it needs to be physical damage. I have derived a formula for EH that is valid for ANY type of damage over at maintankadin (which I would gladly link you to if I could, but MT is down at the moment). It is basically a weighted average of the inverse of your mitigation for each type of damage.
That means that for a fight with a considerable amount of incoming damage from magical or bleed sources (where armor does nothing), your Effective Health needs to be calculated differently. You still need enough Effective Health to survive the worst-case burst scenario (with or without cooldowns, as appropriate for the fight).
Saying that "Effective Health" has to die is somewhat silly, since Effective Health is not a game mechanic. It is a tool that measures survivability, nothing more. The fact that people don't understand the tool, or use it improperly as you have in this article, doesn't make it a bad tool.
As an analogy, a hammer doesn't become a bad tool because some people choose to try and use it to drive in a screw. A tape measure doesn't become a bad tool because someone tries to use it to measure the distance between two cities. One has to make sure they're using the tool appropriately and in the right scenarios (for example, EH is a fairly useless tool for Faction Champions).
-Theck
Theck Nov 30th 2009 10:33AM
Maintankadin is back up. Here's the link:
http://maintankadin.failsafedesign.com/forum/index.php?p=514810&rb_v=viewtopic#p514810
Allison Robert Dec 1st 2009 4:39AM
@Theck:
I don't have a problem with effective health, although WHY I don't have a problem with it was mostly addressed in a previous article. I wound up breaking up a very lengthy piece in order to avoid what journalists call the MEGO - My Eyes Glaze Over - effect.
As a TL:DR, I have a large problem with how EH is used as an excuse for everything that goes wrong during an encounter, but the concept itself is sound. All of the points you've made in your comment are points that I've made in either the article above, the previous article, or in the comment thread here (e.g EH is not solely impacted by the interaction of armor and health, the original version was calculated for a time where burst physical damage posed a routine threat, etc.), I won't elaborate on this too much apart from saying I agree with you. But I feel obligated to note that the statement "Effective health needs to die" was not intended as a blanket condemnation of the idea, but rather its frequent and just as frequently-erroneous use on the Tanking forums.
I'm finishing up this week's Shifting column as I write this (not on EH, a moonkin piece this time out), and when I get a moment I would be very interested to check out what you've written over at Maintankadin.
Theck Dec 1st 2009 9:28AM
Allison,
Whether "EH needs to die" was meant as a blanket condemnation or not, that's how it reads.
I have read both articles, and I understand your complaint. Unfortunately, while the message is reasonably clear from the article (that you're frustrated with the over-emphasis and misuse of EH on the blizzard forums), the title is doing you a great disservice by suggesting to your readership that you think the either the concept itself or the underlying game mechanics that make it a useful metric are at fault.
Perhaps a better title for the article would have been, "Why EH isn't everything," or something similar. In other words, something that more accurately reflects your point rather than misrepresenting it.
Furthermore, you make some fundamental errors in both articles that distract from the argument. As another commenter pointed out, you're not even using the proper form of the old EH definition, which was Health / (1 - Mitigation). You've jumped directly to the version that assumes "mitigation = armor." While that may be representative of the type of erroneous uses you see on the Tanking forums, it isn't the correct definition, and gives informed readers the impression that you know little more about what EH means than the posters you're chiding in the article.
This also leads to mistakes like this one:
"but death knights weren't overpowered in Tier 7 and Tier 8 because they were the highest-HP tank. They were overpowered because they always had a cooldown up to trivialize the high-damage boss attacks that occurred at predictable intervals during the fight."
This is *still* an issue of Effective Health. A mitigation cooldown is a temporary increase in effective health. By chaining mitigation cooldowns, a DK was able to have on-demand EH, which was very strong for fights which were patterned with large, fairly predictable damage spikes. The fact of the matter is that they were able to achieve the highest EH of any tank, on demand, exactly when it was needed.
In several places, you suggest that because attacks are magical or bleed sources, they don't reward EH stacking. Again, this is a mistake borne out of the incorrect definition of EH. Any time a tank encounters burst damage, it becomes a question of "Do I have enough EH to survive this burst?"
In another comment, you said you didn't like how this article came out. I think that the problem is that you're trying to fit too much into one article, and aren't able to do it all justice as a result. I think this article (meaning just "Part 2") would have been stronger if it had been split into two:
-one focusing on "This is the *wrong* definition of EH, this is why it's wrong, and here's the right one"
-another focusing on how cooldowns and raid buffs interact with EH, how other factors influence tank death as much or more than just having enough passive EH, and showing that this is why the "my class needs more EH" argument falls flat on its face.
"Part 1" did a decent job of giving the history of tanking balance and how EH was thought up in the first place. But "Part 2" just feels like it's trying to tackle too much content, and doing so insufficiently in the process.
Karilyn Nov 27th 2009 4:36PM
I still say this article is a message in search of a point.
"Effective health needs to die"
"Boss abilities like Flame Breath and Impale reward the ability to stack raw health"
I mean what's the point of these articles? Are you just trying to argue that Armor is the less important half of the Effective Health formula? I think we all know that. Stamina is far more important to survivability than any other tanking stat.
It's like you are complaining that people are still describing Stamina as Effective Health, even though Armor isn't too incredibly important anymore (relatively speaking).
The Giant Nov 27th 2009 5:03PM
Swing..! and a miss
Allison Robert Nov 27th 2009 5:10PM
Frankly, I'm not happy with how this particular installment of the EH series came out and will be revisiting the point at some point next week or in the near future, but the underlying problem is that players on the Tanking forum chirp endlessly on about a stat that they refuse to place in context. EH by itself doesn't mean anything, more particularly the series of static numbers that get pulled out on the forums to show that bears need to be nerfed and warriors need to be buffed.
This may well be (and I think there's some truth to the claim that warrior health needs to scale better), but left unsaid is that how cooldowns are deployed during the encounter -- and how intelligently the raid plays -- winds up impacting tank survivability more than anything else. Saying "I died because I didn't have enough EH" or "I can't tank that encounter but this other person can because of EH" is foolish. Assuming the required level of gear, all tanks have tanked ToGC just fine (with the exception of the Death Knight on Anub'arak, they're kinda screwed there when it comes to doing either the MT or OT job) and the burden of proof is on the player insisting that it can't be done because they're playing the "wrong" class.
I don't like stupidity, but I hate mulish stupidity even more. I don't think I made the observational transition in this outing very well, but I'd much rather that people on the forums asked better questions about how people die on these encounters. If you're, say, a tank who's consistently dying on ToGC Gormok-25 and howling on the forums that your class needs a buff, and your raid reports show that there's no improved AP reduction on the boss, your hunters aren't using Scorpid Sting, your priests/shammys are letting the 10% physical damage reduction buff drop off you regularly, you're not consistently Earth Shielded, etc. etc. -- then all available evidence points to the fact that your raid isn't working to keep you alive.
These articles are ideally about how an otherwise good concept (EH) is getting twisted into a catch-all excuse for tank failure. I hope the cooldown article makes the argument a bit better.
Karilyn Nov 27th 2009 5:34PM
They chip about it constantly because it's the right thing to do for Warriors, Druids, and Death Knights.
I understand you are a Druid, and Agility is an amazing stat for you.
But for the other three tanks, there is literally no stat that compares. You survivability goes up if you get more stamina, and nothing else even comes close. Only in the most extreme situations would you chose avoidance over more stamina; generally in the orders of a piece of gear. To be equal, it often requires as much as 3-4 times as much equivalent avoidance as the stamina you stand to lose by using the lower stamina piece.
9 times out of 10, if a Warrior, Paladin, or Death Knight is dying on a boss fight (assuming they aren't "standing in fire"), and they switch all their gems and enchants over to stamina, they will miraculously stop dying, it makes THAT much of a difference.
I know you mean well, and I normally like your posts Allison, but the only thing this blog post has seems to have done, is caused several newbie tanks to post "Oh I knew all this EH stuff was a phony, I'm going to socket and enchant and gear pure avoidance now"
You've mislead many newbie tanks, who are now on the wrong foot, and will be gearing incorrectly and having a harder time tanking as a result. That's a bad thing.
Bikhai Nov 27th 2009 5:37PM
The fact remains that Stamina isn't the same thing as EH. Stamina will always be important as EH. The problem right now is that even though some bosses do hit like a truck, requiring a lot of stamina and a lot of armor to mitigate it, there is more to the tanking equation than just those stats.
What Allison is saying is that we need to get away from EH as the sole determination of a tank's value and start looking more at the other factors that contribute to survivability. And yes, developers are moving away from EH with ICC. Smaller, more consistent hits against tanks means that you don't have to stack Stam/AC out the wazoo to survive like you do in current content.
tim Nov 27th 2009 5:51PM
@Karilyn
"9 times out of 10, if a Warrior, Paladin, or Death Knight is dying on a boss fight (assuming they aren't "standing in fire"), and they switch all their gems and enchants over to stamina, they will miraculously stop dying, it makes THAT much of a difference."
9 times out of 10, if someone says something like "9 times out of 10" without a citation or a model to back up the claim, they're just making up numbers.
Think a little bit about what you're saying. You're saying that, if a Warrior, Paladin, or a DK tank is dying and their gems/enchants aren't all stamina, in 90% of cases they can get through the encounter merely by switching all the gems/enchants to stamina.
WOW! That's a pretty strong claim! My guild's DK tank will be relieved when I tell him that he's not dying because the OT is screwing up, or DPS isn't burning fast enough, or healers are dropping the ball, or--god forbid!--he just got an unlucky string of hits on him... no, it's because he foolishly gemmed stamina/dodge rather than pure stamina on a few gem slots.
/shrug
Karilyn Nov 27th 2009 6:06PM
Yes, I'm seriously Tim. I have seen more tanks than I can remember, who were gemming and enchanting for standard "maximum stats," which basically means, all socket bonuses, and the "highest rank" enchant they could use.
These tanks, died on Heroic Gormak over and over and over; simply dropping dead spontaneously, when nobody in the raid was doing anything notably wrong.
Those tanks replace all their gems and enchants with pure stamina, gain an extra 3k health, drop around 1% avoidance, and they stop dying.
Also. One of those tanks? Do you know the name of one of those tanks I'm talking about?
I'll give you a hint. It starts with a "Kari" and ends with a "lyn"
I'm not just talking out my butt. The difference is mind blowing. It is the difference between night and day. Don't believe me? Try tanking with "balanced stats" one week, then pure stamina the next, and watch your survivability skyrocket.
Allison Robert Nov 27th 2009 7:34PM
@ Karilyn
I doubt that anyone's going to run out and gem nothing but +parry (The horror! The horror!) because of this piece, but I also think it's important to note that most of the discussion here's centering on our buddy Gormok. We're talking about one particular encounter where stacking the bejesus out of stamina pays enormous dividends for all tanks, with the reason being that the Impale mechanic is a huge damage tick that doesn't care about anything but how much health you've got left or how much you can mitigate by blowing a CD.
Geared and experienced tanks with a competent heal team aren't in especial danger of dying on most other ToGC-25 bosses (I mean, the whole point of the +block set on Anub'arak is how little damage a warrior can take from the adds while wearing it), so telling everybody to run out and gem stam on everything because you'll get stomped otherwise jumps the gun a bit. It's useful and it gives healers a margin for error. It's fantastic on progression content when healing errors are common while the team's learning the fight and you don't have the buffer provided by more advanced gear (which I think was really Ciderhelm's original point concerning EH -- stacking EH on progression content mimics the health bonus granted by better gear).
It's not necessarily required or even desirable on fights where the greatest threat to tank survivability isn't the brute force of a boss ability -- or a boss ability that, at present or theoretical levels of EH, will still one-shot you without the use of a cooldown -- and EH as the new catch-all answer to everything that goes wrong for a tank is a something that I very much want to die.
It's not the concept, it's how it's used, and how it's become an excuse for everything. Tanks need to be more honest with themselves about what's more likely to kill them than "I don't have enough EH, QQ."