Why We Do What We Do (An in depth explanation of EHP and ICC 3.3.3 tanking mechanics)

[Passive]

Agg, I was generally with you until you said 172 dodge rating and 6.4k armor proc was worse than 111 Stamina. That's taking it a wee bit too far, IMO.

(Although I would agree the H:CTC is not a very great trinket, neither is Essence of Gossamer at this content level.)

Essence is better for effective health though. When CTC has proc'd and is on cooldown, the 1500 or so health from essence is still going to be there. That's why is effective health. An armour proc is no more reliable than avoidance, especially one with an internal cooldown. Thats why, for the worse case scenario (that's what we're all gemming full stamina for, right?) I would chose essence of gossamer every time.

Edit: just as a point of interest I thought I'd point out that some well geared disc priests and resto druid have gone back to using the Illustration of the Dragon Soul trinket. It is item level 213, but once fully stacked provides 200 spell power. This is second only to the 277 gunship heroic healer trinket (201 spell power?). They use this because after a certain point spell power is so much better than anything else, all their other stats and mana regen just come from gear. This is like endgame tanks gearing for EH because crit immunity and avoidance high up on the DR curve is inherent in their gear. Even a small increase in our best stat is better than a large increase in one of our not so great ones.

[Kojiyama]

Saying something is better for EH when it is the only one for EH is a bit of a no-brainer--but that does not, by default, make it a better trinket. I think that is an example of taking the 'EH > everything, regardless of quantity' a bit too far.

The comparison to mana regen for healers is not the same, as mana regen is something that specifically becomes of zero use once a certain gear threshold is met. 172 Dodge Rating never becomes of zero use to a tank, nor does the armor proc.

Although EH certainly has major benefits, Dodge will contribute to you avoiding death in many cases. When the quantity of the values is so disparate, it's certainly worth considering dropping the 'EH > everything, regardless of quantity' mentality, IMO. Comparing 20 Dodge Rating gems to 30 Stamina gems is totally different to comparing 172 Dodge Rating to 111 Stamina. That is even before the armor proc, which has pretty high uptime and is not particularly 'unreliable' (if you are in a situation where 111 Stamina would help you considerably, you are in a situation where you will drop below 35% and trigger the trinket, it is not random.)

Unless the fight is purely magical, you are probably making a choice that is detrimental to your long-term chance of survival to take such a statistical hit just for a bit of extra Stamina.

If you said Flawless Fang, Skeleton Key, Glyph of Indom, Balebrew Charm, Satrina's Scarab, The Black Heart (considering the armor proc), or even the Heart of Iron I would most likely agree in all those cases... but Essence of Gossamer seems a bit over the top.

[Aggathon]

the armor proc is just as unreliable as TBH (though I think overall TBH is probably a better trinket too). I really don't factor in the armor proc difference especially because if it happens again within the cooldown period it can't be relied upon, though admittedly a proc at a % is probably better than a random proc.

172 dodge rating is ~2% avoidance though. In ICC 111stam = 111*10*1.09*1.1*1.3(as of today I think) = 1730 hit points.

I'm going to take 1.7k HP over 2% avoidance any day for progression content.

And Thaurassian now I know you're trolling. You seem to really like taking what I say out of context and making it personal. I was almost benched because I was doing it wrong. It had absolutely nothing to do with my class because once I fixed my problems I was able to do everything any other class could do, and some things that not every other class could do. Furthermore you never even responded to my counterpoints, or anyone's for that matter.

Little side note down here: You could probably make an argument that the armor proc might be better than the relativistically low stam on essense of gossimir, but the argument is about dodge vs. EHP not the armor proc and it's his contention that armor is far more worthless than dodge anyways.

Edit: Also Jayde, where avoidance really shines is when MANA becomes a major consideration. As long as I'm smart about my cooldowns, and sometimes even when I'm dumb about them, mana is almost never a problem for me while healing. I seriously believe based on the results of the simulations posted in section 6 that with the ICC HP buff and the avoidance debuff that causes that curve that 2% dodge is statistically irrelavent and I would take even relativistically low EHP increases over it, and I think essence is a far as you can go.

Furthermore, where is your brightline? Glyph of indom is only ~ 40 EHP worth of stam better than essence of gossimir. How low do you have to go. My contention is that when you can't make that brightline (which modeling effort after modeling effort has thus far failed to do, although Hypathia is currently doing some fascinating work and if we can use those simulations to come up with a brightline then I'd be thrilled) so then you're faced with either a marginal gain based on a random number generator that may or may not be there for you or a marginal gain of EHP that IS always there for you.

Edit2: I'll conceed another point though, that both CTC and essense of gossimir are both pretty bad trinkets at current progression levels and frankly there's no excuse for a tank that has more than 10 hours played at 80 to not have both the black heart and glyph of indom.

[Kojiyama]

Saying X% avoidance is not a meaningful distinction. Avoidance has increasing returns. The more you have, the more X% is worth. 2% to someone with 50% is worth less than 2% to someone with 90%. The DR curve on ratings makes the rating increase linear, which is why I compared rating to stamina not HP to avoidance%. The value of avoidance also increases with your current health and your current armor. So, broad rating equivalence is all that is important.

If you are saying that in all cases, Stamina is 2.5 times better than Dodge Rating, I personally believe that is taking the Stamina argument a bit too far and into unsupported territory.

You seem to be making that particular trinket distinction via 'feel', not so much via math. I can almost 100% assure you that other than in extreme fringe cases, using the H:CTC would result in a lower chance of dying than using the Essence of Gossamer. (Not the least of which due to the proc, which I'm not sure why you are dismissing. It is 10s of extremely boosted EH when you are in a low-health scenario.)

In a Burst Time model, the value of the H:CTC is over 3 times that of the Essence of Gossamer, without even so much as factoring in the Armor proc whatsoever--which I still contend is a fairly powerful proc, due to its mechanics. Getting to the point of taking something so terrible in both budget and contribution just because it happens to have the word 'Stamina' on the tooltip seems to be taking it into zealotry territory, which I really can't get behind.

Pretty much any other trinket you named, I would agree. But that is an extreme example. Where do you draw the line? Is http://www.wowhead.com/item=30620 better than H:CTC? What about http://www.wowhead.com/item=31080? http://www.wowhead.com/item=21568?

Not sure about the armor vs. dodge argument trying to be made elsewhere. Armor is awesome, but has a fairly low availability and low cap. It's not a scenario that comes up very often, either way.

[Aggathon]

I'd draw the line at essence of gossimir, being a level 80 trinket and all... I also tossed in an edit there you might not have seen.

[Aggathon]

A little more detailt hough.

Saying X% avoidance is not a meaningful distinction. Avoidance has increasing returns. The more you have, the more X% is worth. 2% to someone with 50% is worth less than 2% to someone with 90%.

Which is exactly my point and is shown very clearly in the graphs in section 6. Now if we were talking outside of ICC where it's easy to get 70%+ miss/dodge/parry (would probably have to alter gearing/gemming a little) then yes, I would completely agree that CTC is the better trinket because the avoidance actually has really good returns. But in ICC where the average avoidance is closer to 43%, 2% more dodge has relatively bad returns.

If you are saying that in all cases, Stamina is 2.5 times better than Dodge Rating, I personally believe that is taking the Stamina argument a bit too far and into unsupported territory.

I really don't think that's unsupported though, at least in the confines of the post, that being ICC progression content. The runs calculation simulator pretty clearly shows that until ~57% avoidance or so (rough eyballing it number) you're not gonna get great returns in the long run.

Not sure about the armor vs. dodge argument trying to be made elsewhere. Armor is awesome, but has a fairly low availability and low cap. It's not a scenario that comes up very often, either way.

Which is our entire point and that's the point Thaurassian is contending.

[Thaurissan]

Furthermore you never even responded to my counterpoints, or anyone's for that matter.

I fully understand what you're saying, as well as those who are mostly mindlessly parroting what you've said (which is a big part of the concern with the extreme position you've taken). I don't think much thought is being given to my arguments at all. The response is basically "the original post says so"...when the original post itself admits that it is "damned near impossible" to answer my question.

it's his contention that armor is far more worthless than dodge anyways.

And that is not remotely what I said.

[Tinytankz]

Parroting is one thing but concurrence is another. You cannot simply say X dodge is better than Y armor when the levels are near each other (in terms of itemization costs) However You cannot say that dodge or parry or Avoidance a a whole is better. Yuo simply cannot plan around not eating Big hits. You have to plan on eating them and eating them often. Armor is a Guarunteed amount of REduced uncoming physical damage. Dodge is a possible chance to not take any. Adding 1% or 2% to your dodge does not mean that any dodge you then get is from this new source you simply gave yourself better odds. The arguement in question is a Blue Stam trinket with a Epic Dodge trinket and true story the Stam one wins more HP for more damage you can take. Rather than simply Looking at the math and saying Yes Stam >Dodge or Armor > Dodge your spending time on a wierd logic train.

We cannot Equate when dodge or avoidance as a whole outvalues EH. However since you cannot either your insistent posting that it is or can be is a logical falicy. Your contending that since we cannot prove a quantitative comparison that you must have a valid and solid arguement. You do not.

[Thaurissan]

We cannot Equate when dodge or avoidance as a whole outvalues EH. However since you cannot either your insistent posting that it is or can be is a logical falicy. Your contending that since we cannot prove a quantitative comparison that you must have a valid and solid arguement. You do not.

It is not a logical fallacy. We all seem to agree that there must be a point where X amount of dodge is > X amount of armor. We also agree that it would be extremely difficult, if not impossible, to calculate.

The other part of my disagreement here has to do with the lack of accounting for class mechanics. If the issue is the "5 swings and you're dead" worst case scenario, you cannot make calculations or claims about it for all tank classes without taking their mechanics into account. In the case of Ardent Defender, the 5-hit scenario automatically becomes 8 if the proc is off cooldown, because you are saved from one killing blow and returned to 30% health, which means the next hit will not kill you either.

Let's say you have 60,000HP. That means that 5 hits of 12,000 damage (after your other mitigation) will kill you. AD's 20% damage reduction automatically kicks in on the 5th hit, reducing it 9600. That leaves you with 2400 HP. The next hit (6) procs the CD, which restores you to 30% health (18000 HP). The next hit (7) is for 9600 because you are below 35% health, leaving you at 8400. The next hit (8) kills you...assuming you got no heals at all during that period of time. But wait...you're using Judgment of Light, which means there is a very good chance you will get healed for 1200 at some point from your auto-attack even if no healer casts a heal on you and there are no HoTs on you. If you get that heal twice, you survive another hit (9). That's without using a single active CD.

In other words, AD is guaranteed to turn the first "worst case scenario" in a given fight into 8 or 9 hits instead of 5. If another "worst case scenario" happens before the 2 minute CD is up, that one becomes 6 hits instead of 5. Considering that most fights last at least 5 minutes, the AD proc will be off CD again later in the fight.

If we're talking about "time to die", you can't leave out that kind of mechanic and call it a day.

[Gape]

I think Agg and several others have stated that they would take a small amount of guaranteed EHP over 99.9% avoidance in ICC.

And the point where avoidance is greater than armor is when you're armor capped in ICC.

At the cap, armor no longer gives EHP, so if you have the option of Armor or avoidance, then you take avoidance. But Stam is still there, and Stam increases EHP (EHP>Avoidance in ICC) So if the option is Stam, avoidance or armor (that is capped, so is worthless) you take Stam. (but no one is going to stay at the armor cap for the entire fight. They will most likely be capped for 10-15 seconds every minute-ish. So Armor is still going to give EHP in a real scenario).

And as far as I know, there is never "too much stam", so if there was some way that anyone could get armor capped for the entire fight, you'd still take Stam over avoidance in ICC. Unless you overgear the content.

I used a scenario where you are armor capped so that you would see that even when armor is capped, avoidance is still beat by another stat.

[Kojiyama]

Armor generally wins due to it being both EH and damage reduction sumultaniously. It's not difficult to beat Avoidance or Stamina due to this fact--Armor is, until capped, generally the highest-value stat for its item budget.

All this is quite moot since the opportunities to choose between the them are so few.

[Kazeyonoma]

Okay, so here's the /end of argument post

Yes, there is a point when avoidance overtakes armor in value, what value that is, we don't know. In current content, EHP is king, so much so that armor is arguably the best stat possible thanks to things like +30% health, however, armor stacking is unfortunately limited based on a few gear selections. Hypatia (the original Hypatia not the new also awesome Hypathia) already mapped out the awesomeness of avoidance when it gets to high enough levels to become "reliable", but due to the mechanics of current in game content, these values are inapproachable, and because healer mana isn't an issue, EHP is, and has been for a long time king, you can't go wrong with it in almost any situation when comparing EHP values vs. Avoidance values. YES there is a breaking point, but it is impossibly difficult to calculate, so for REALISTIC purposes, stacking EHP is a more tried/true/valuable gearing method for us to go by, rather than HOPE that we're in some magical way, surpassing the aforementioned break point.

Conclusion: Is avoidance good? Yes. Can it be better than Armor? Possibly. Is avoidance better than EHP? In current content no. Will it ever be? Who knows, with cataclysm, many things are changing, and guides like this, can and will be re-written to adjust to the tides, but just like Ciderhelm's original guide into tanking (Fortitude, and/or Holding the Line), the core elements often are not lost, just merely adjusted. Now can we stop arguing about Armor vs Avoidance in a thread that is supposed to explain WHY we do things in CURRENT content? Theoreticals are fine, but they don't necessarily belong in this thread.

[Aggathon]

I think Agg and several others have stated that they would take a small amount of guaranteed EHP over 99.9% avoidance in ICC.

I've never said it to that extreme (though others have), based on the curves and simulations once you dip down into 70%+ avoidance more added avoidance would be very very very good. I'd probably take into consideration avoidance more if our base values were closer to 75%+ because then the benefits from avoidance are very good and even at those levels it reduces the odds of even getting hit 5 times in a row is like 10% for a fight (pretty sure you could afford to just use a cooldown here too). The problem is when you're up near 40-50% avoidance the odds of getting hit 5 times in a row in a fight is above 90%. In order to get to the point where you could push those statistics far in your favor you'd lose so much EHP that it would become a matter of being hit 3 or even 2 times in a row and the stats are thrown off of your favor again.

If it wasn't for CotT this would be an ENTIRELY different discussion and I might even agree that CTC would be clearly better than EoG. As it stands though, I don't see it (edit: and by it I mean the dodge portion, you could make reasonable arguments for the armor proc vs. the static stam since it's SO MUCH armor, but that comes to personal preference/current gearsetup imo, and is fairly ludicrus anyways because you could just get TBH and Glyph of indom, both really easy to get and better than both EoG or CTC as even admitted by Jayde, so this point really does become moot) making a big enough statistical difference for it to clearly affect one's survivability over the course of a fight.

Also:

My biggest question is not about stamina vs. dodge. It's about ARMOR vs. dodge. Since armor and dodge both protect only against physical damage, I think it is a much closer call than it's being made out to be. Lumping armor in with stamina as "EHP" is misleading in that sense.

[Gape]

Thats why everything I said had "in ICC" tagged onto it. Seeing as you can't reach the values of avoidance that you mentioned, then armor is better.... in ICC.

I agree that outside of ICC (or without CotT) avoidance stacking has it's place, but there isn't much room in ICC for it, like your post says. But we're not discussing what works/doesn't work outside of ICC because it is irrelevant.

[Mend]

I can't help but notice that almost every comparison made in the past argument is unrealistic.

Comparing 10 armor vs. 20% dodge gain, Or comparing a starter raiding trinket ilvl 200 to a high end ilvl 277 just doesn't seem to add much truth to the argument made by Thau.

No tank in current content would follow what Thau was trying to say, and if that tank did he would than experienced what Agg said about being "The balanced tank" and would have figured out that EHP is far more beneficial at this time.

There just isn't enough avoidance in current content or even in this whole expansion to reach the numbers where avoidance starts becoming more beneficial than EHP.

You have 2 trinket slots, if you can get the corpse tongue coin I'm sure you can get Satrina's Scarab and Skeleton Key even through PuGs. Using a trinket such as CTC in current gearing demands will not prove to be very beneficial.

[Satorri]

Using a trinket such as CTC in current gearing demands will not prove to be very beneficial.

Do you have anything to support that other than your feelings?

[Mend]

Do you have anything to support that other than your feelings?

No, nothing more than comparing it to its rivals of the same ilvl.
I'm not gearing towards avoidance at this time, if I was than CTC would be BiS in that case.

CTC vs. Skeleton key, Flawless fang, H-Satrina's scarab, Organ

The later 4 would win hands down if you are after max survivability. Each providing over 2000 health modified by talents or a healthy amount of armor and useful proc.

[tlitp]

Surprisingly, it seems that no one has taken into consideration the pattern of the boss attacks. f(n,k,p) can be decomposed into three distinct factors :
p - accounts for player gear and raid environment
k - accounts for the size of the incoming hits (and the healing pattern, to a degree)
n - accounts for the pattern of the attacks

A Patchwerk-type boss can be, indeed, modeled using a large value of n. Patchwerk doesn't change primary targets (unless death occurs); Patchwerk doesn't move (unless kited);Patchwerk doesn't "waste" GCDs on spell casts (with the odd exception of Slime Bolt). What's the catch? Well, most bosses aren't Patchwerk.

If a given encounter can be modeled using m distinct sets of events (n'=n/m, the sets are supposed to be equinumerous for the sake of simplicity), the (total) probability of getting hit at least k times in a row is : f'=1-(1-f(n/m,k,p))^m. One can draw two conclusions :
- f' decreases monotonically as m increases
- avoidance itself gets better and better as the attack pattern is increasingly more coarse-grained
(For those which visualize better the theory when confronted with numbers : run the simulator provided by Agg, comparing f(100,5,0.4) to 1-(1-f(25,5,0.4))^4.)

Of course, real-world practice is significantly harder to model - the assumption of distinct, equinumerous sets isn't valid in most of the cases. Still, the conclusion is the same : given m>1 subsets of events, f' is always smaller than f as long as 0<p<1 holds.
The goal of this objection ? People should realize that choosing large values of n leads, most of the time, to abstract results. Making further inferences on the basis of these results should therefore be avoided.

[Kojiyama]

Well, a perfectly accurate simulation of raid environments could well produce output data that could, more or less, conclusively say that in X situation, adding Y Stamina is better than adding Z Avoidance.

Making such a simulation is exceptionally difficult, however. The amount of input factors are quite massive and each one could dramatically change the results. For instance, a result that may be valid in one scenario could be completely the opposite scenario with a different class distribution of healers or even different gear levels in one of your primary healers.

The subtle ways that Stamina, Avoidance, and Armor interact within complex, in-game scenarios are not those that are easily observed or quantified by anecdotal evidence.

[Hypathia]

You're right Tlitp, taking into account granularity seems to be important. Elsewhere, The RAWR protect warrior is modeling that although I don't really understand how it works (look at Kojiyama for an answer ).
Determining X Y Z as defined above is still really complicated, as X has many configurations. If we want to be precise, we must define :
Boss

• granularity of damages
• Intensity

Healers

• class
• Lvl of gear.
• And I would add the skill too, see *below.

Tank

• Lvl of gear
• Class

Encounter : Scripting it is interesting. It's trivial on Festergut as everything is known by advance on normal mode, heroic needs much more work, as the healers are interrupted. Affecting probabilities of interruption or things like that is interesting too.


*Basically, the more there are interruptions in healing, the best is avoidance. And it is very significant. I extrapolate that with really skilled healers, avoidance become -even- less interesting.
Determining XYZ is what I'm doing (Publicity) to try.

(On a side note, I finally realized there was a Hypatia before me, working on avoidance too with great investment. It makes me both smiling for the coincidence and unhappy as I may look as an usurper. We surely have nothing in common unless our love for the real Hypatia.)