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

[Tinytankz]

I am not worthy.
/bow

All joking aside my tanks right now either look to me as thier tanking god or scoff at my ways. I preach EH all day to all my tanks and only one listens (guess who my OT is every raid) this gives me a large tool in swaying my other guys meaning i may actually be able to dps one of these times. Thank you Agg this is deeply appreciated.

[felhoof]

Mess around with the survivability spreadsheet, it covers this.

Well, it implies this. But it doesn't cover the strengths and weaknesses of armor. One of the arguments you see often around these parts is that EH == stam, and stacking stam is the way to go no matter what. And barring that, armor sometimes has value simply by being mitigation, even when it's not the optimal EH choice (think armor trinkets vs stam trinkets as an example) due to allowing lower HPS on the tank. That does get shaken out by the spreadsheet, but most folks aren't going to look at the spreadsheet - and they're certainly not going to look at it in that kind of detail.

Going back to the discrete sequence of bad events vs overall damage for PEH+MEH - I guess I didn't say it well enough. The implication from your paragraph is that sometimes you'll want to do that (figure out what the worst sequence of events is) but in my experience that's about the only value in EH calculations. Figuring out how much overall damage you take at a time and what percentage it is doesn't really help when determining targets for me, especially on hard fights. Tanks nowadays don't die because of attrition, they die because of burst.

[Eoika]

So.... Test is in a week, and can we use notes?

[Thegreatme]

So.... Test is in a week, and can we use notes?

No, you have to figure out and memorize the equations we created to determine TTD, and the survivability quotients.

PS. agg, why didn't we just do something simple for SQ like MinTTD+avgTTD?

[Aggathon]

I seem to remember thinking of it, then thinking that there was some reason why not, but that was after 10 hours of bashing our heads against the spreadsheet wall and I couldn't think straight and had nightmares about spreadsheets.

Felhoof: I don't get it, in my TEH paragraph I talk both about total damage and possibly burst damage scenarios and how it can change between fights about which you'd use, what more do you want from me?

Also, I'm pretty clear that EH = Armor + Stam, and that HPS needs to be > DTPS. But frankly, our main holy pally healer, while just spamming holy light (no val'ynir procs, no sacred shield, no glyph of holy light spash tricks, no bacon of light, but also no divine please, but while keeping up his necessary buffs like judgements) can sustain 13.3k HPS. Now obviously this is a rough number, but it's probably a good average and baseline to work from.

That's one healer, put 3 healers on a tank and you can easily get to > 30k HPS. The only boss that hits that hard is H-TLK 25, and due to fight mechanics you really do want to stack stam on that fight. I am still a believer that until healer mana is a problem, stacking both armor and stam is critical and it doesn't really matter which one as long as it is more EHP (obviously stam matters more for magical fights, blah, but you get the point) and it just so happens that with current gear selections we have a lot of armor based choices that provide us with more EHP. If you want more of my opinion on this matter, go search for "misconception about trinkets" or something thread started by me.

And also, there is no possible way that I can include every single last aspect and hold people's hands through this. I wrote a ton as it is and still didn't including everything I wanted to. The entire point, and main thesis in the introduction, is for people to understand WHY the mechanics work the way they do, not for me to just continue to tell them, "yo dawg, stack armor broah."

I can't break down every fight for people, but hopefully they can look through this guide and understand what they should do, and it will make them better tanks.

[felhoof]

Felhoof: I don't get it, in my TEH paragraph I talk both about total damage and possibly burst damage scenarios and how it can change between fights about which you'd use, what more do you want from me?

Sorry - I didn't mean to put you on the defensive.

I'm saying that the implication I got from reading that paragraph was that sometimes it's good to just take the total damage in a fight, figure out what's magical and what's physical, and then maximize your EH based on that. My point is that this is never a good strategy for figuring out what EH you might need or how to maximize your useful EH in a fight. I can't think of a situation where this is the case in LK. Can you? I'm genuinely serious here; I don't know of anything that looking at total damage would be useful for in terms of EH.

Also, I'm pretty clear that EH = Armor + Stam, and that HPS needs to be > DTPS. But frankly, our main holy pally healer, while just spamming holy light (no val'ynir procs, no sacred shield, no glyph of holy light spash tricks, no bacon of light, but also no divine please, but while keeping up his necessary buffs like judgements) can sustain 13.3k HPS. Now obviously this is a rough number, but it's probably a good average and baseline to work from.

That's reasonable. At the same time, there are plenty of guilds out there without a single holy pally. And there are plenty of situations where for whatever reason figuring out exactly what HPS you had and how to minimize the drain on the tanks was important; think about Anub'arak or Yogg0 as examples. My point on this was only that you didn't talk about when armor is better than stamina for physical EH purposes, and that's a shame.

If you think I'm riding you hard on this, it's because you've written a damn fine document that is in all likelihood going to be quoted and misquoted for a long time. Which means in the grand scheme of things people are going to take a lot of what you wrote and use it in ways that you didn't want, or believe certain things that aren't accurate.

[Satrina]

If you think I'm riding you hard on this, it's because you've written a damn fine document that is in all likelihood going to be quoted and misquoted for a long time. Which means in the grand scheme of things people are going to take a lot of what you wrote and use it in ways that you didn't want, or believe certain things that aren't accurate.

This is as certain a truth as you will ever find. Once upon a time in about 2005 I said "or maybe there is just always a once-in-a-blue-moon chance of getting crit regardless of how high your defense is" =) How much have I regretted adding that bit in? Yeah.

[Kazeyonoma]

Lol, and just like that Satrina has forever built in the "super rare chance to be crit" in the minds of the uninformed forever ;P

[Satorri]

Amg, it's so true, I got crit this one time, and I had lots of defense, and I know it must be true!!!!!

/fan flames

[Fathom]

Solid piece of work, my only major complaint about it is that I kind of wish you'd laid out more explicitly up top what assumptions the body of the text is working with, in particular what a Progression Raiding environment is. In particular with regard to healer behavior and other demands for healing, i.e. healers have no significant mana limitations (which admittedly is noted in the counterarguments to pro-avoidance lines of reasoning), healers will be topping off tanks between most if not every hit, and dps that isn't putting additional healing requirements on the raid beyond unavoidable raid damage (since it impacts the assumption that healers won't have mana limitations. (At least all that's what I assume is what you mean by progression raiding).

I only mention that because a good chunk of the people who need to have the "Why EHP?" question answered for them may not be raiding under the same assumptions (e.g. puggers, casual raids, etc.) that the rest of the math assumes.

[Aggathon]

Well, I use the term for "progression content" and "things you haven't downed yet and don't explicitly out gear." Given the mechanics in WotLK, either: you are going to want to stack EHP to give healers, good or bad, a better margin of error for when maxDTPS occurs, since it will occur, or you will flat out outgear the encounter. Now if you out gear a fight and your healer can pretty much play peggle while healing you, then frankly none of this matters, you could gem avoidance, threat, spirit, whatever tickles your fancy.

[Bodasafa]

Nice job Agg. I'm only sorry I did not have time to help contribute as originally planned.

[Aggathon]

It's all good:

Also, I'll edit the post where I link the Survivability spreadsheet, but WarTotem caught a few errors in it and gave me a better way to calculate minTTDwHeals. I will be updating this and giving a more in depth explanation of the spreadsheet soon, probably sometime tomorrow.

Simulator is fixed and is now posted.

[Kazeyonoma]

I'm offended you never came to ask me for help. Deleting post....

[Aggathon]

I'm offended you never came to ask me for help. Deleting post....

Hey, every other time I asked you you turned me down! Who's cryin' wolf now?!?!? Or... something...

[jere]

Sorry - I didn't mean to put you on the defensive.
I'm saying that the implication I got from reading that paragraph was that sometimes it's good to just take the total damage in a fight, figure out what's magical and what's physical, and then maximize your EH based on that. My point is that this is never a good strategy for figuring out what EH you might need or how to maximize your useful EH in a fight. I can't think of a situation where this is the case in LK. Can you? I'm genuinely serious here; I don't know of anything that looking at total damage would be useful for in terms of EH.

There is definitely a common misconception that TEH is only used via total damage. You can use TEH for burst as well. You are not forced to use total damage at all. What you typically do in those scenarios is figure out what your burst scenarios are, define the magic and physical sources of that damage, determine their max values (if they have a range), and use those values to apply the MEH and PEH percentages. There isn't any need to average them or use total damage in any way, unless that is something a person is interested in for some reason. EDIT: (I realize you might already realize this, but when I read you posts they come across as if you believe TEH is only used with total damage, so it is probably just a misconception on my part).

In terms of when using total damage might be worth anything is if you need a napkin math figure to compare two fights for a general idea. It can sometimes be easier to look up an old parse, take the totals and do a quick calc. It isn't a valuable result for any serious theorycraft, but not terrible for offhand calcs. I agree though that those times are limited to almost never probably.

[felhoof]

I understand that, Jere - but the implication I got from the statement was that you should be using total damage sometimes - otherwise, why clarify it as sometimes you should use burst damage? Guess I should have stated that it would be better to make a definitive statement of how to use EH and figuring out PEH/MEH, and that you should use it by defining a dangerous situation and then determining what the burst damage in that situation could be - NOT to use it by looking at the total damage over a fight and averaging what sources it comes from.

I'm still curious if there are any situations where using total damage over a fight is at all useful.

[Insahnity]

Agg, really nice work. I still don't know why your name is blue.

There are a few things I would like to cover.

You mention that this is targeted towards progression, and this (I hope) implicitly implies a functional raid doing their job (healers healing at the right time which I think you take a bit for granted, and DPS having enough DPS to beat the boss before healer backed tank collapses, DPS watching threat and not distracting healers, etc.). But the fact of the matter is, people in struggling raids who are nowhere near this level may read this and stack stam (and not EH, because they just don't read everything), and then land in trouble as their threat stats go down the tubes, and they *do* become squishy as they can no longer hold threat, boss wanders over and eats the top DPS or a nearby healer, and everything goes downhill from there. Any advice given to address threat is a bit necessary, because either the DPS doesn't wait or doesn't use the most obvious tools built to address this, namely Misdirects and TotT. Also, you mention healer mana is a non-issue, but I can recall a lot of raids where there are zero to one source of replenishment, and if that source dies, healer mana ceases to be trivial (which, BTW, helps explain your Holy Pally sustaining high HPS without Divine Plea). Simply put, a raid with proper DPS (including threat redirection talents), proper composition to ensure all buffs and DPS styles (melee vs ranged) are covered, Replenishment and geared healers, basically are pre-requisites for any of this to apply.

Another topic is the heals themselves. When calculating TTL factoring in healer heals, the healing style needs to be considered. Ideally, you want a raid with 5 healers, one of each type (Holy Priest, Disc Priest, Tree, Resto Shaman, Holy Paly), to get the spread that you need. Holy Light for baseline 2-tank healing, Disc supporting Holy Paladin during tank damage or Raid heals while providing 3% Damage Reduction, Shaman for strong melee heals (who inherently may take more damage), a resto druid to roll hots on a tank and raid heal, and finally a Holy Priest for raid heals (with option to switch out to shadow if less healing required). However, assuming this kind of heal structure, the sequence of healing is important. Healers are ALWAYS healing (unless motivated by movement fights), which was always true even back in TBC days, but I find myself doing more cross-functional healing (Holy Priests and Druids contributing to tank heals, Paladins using HL glyph Splash to help the raid, etc.).Remember, back in TBC, spirit based healers (i.e. raid healers) would raid heal when called for, otherwise they would rest and take advantage of 5SR, while tank healer would never dare raid heal because they just didnt' have the tools (Pallies didn't have splash heals, CH had a small range, Disc Prayer of Healing was limited to only their party, and so on). The net effect of the shift was to smooth out the heal profile, as more healers were on the tank for smaller amounts (Druids now rolling HoTs on tanks even when assigned to raid healing, Holy Priest mechanics necessitating the use of Flash Heal, Resto Shamans bouncing CH off the tank) and while maintaining the pressure on heals by giving raid heal tools to everybody and putting massive raid damage forcing everybody to chip in. So comparing a Brutallus fight today, if you were to hypothetically gather 25 x 70s and re-run brutallus in 3.3.3, even setting aside DR on avoidance, people would still stack non-integer EHP values as you would have all 5 healers contributing to the smooth tank healing profile, not just 1-3 tank healers hopped up on cast or cast-canceling like clockwork. This is why I firmly believe it's possible to run a raid with 4 tree druids rolling HoTs even under mobility fights, with the token Disc Priest just to throw up Inspiration. The sheer rolling of HoTs with the occasional Nourish further reinforces the power of EHP, whereas a raid of 4 Holy Paladins or Holy Priests would simply results in a lot of overheals and greater risk of wiping. I'm not saying go-go Tree druids, I really love my resto shaman the most, but the fact of the matter is that constant predictable incoming heals are pretty critical to an EHP centered tank, and a HoT centered healing style is just best suited for accomplishing this. Incidentally, this is why everybody else (most especially paladins) are going towards either HoTs, except Disc Priests (but shielding is just as effective as a HoT).

[Insahnity]

I understand that, Jere - but the implication I got from the statement was that you should be using total damage sometimes - otherwise, why clarify it as sometimes you should use burst damage? Guess I should have stated that it would be better to make a definitive statement of how to use EH and figuring out PEH/MEH, and that you should use it by defining a dangerous situation and then determining what the burst damage in that situation could be - NOT to use it by looking at the total damage over a fight and averaging what sources it comes from.

I'm still curious if there are any situations where using total damage over a fight is at all useful.

Felhoof, I'm kinda with you on this, but let me try a different explanation.

In WOTLK, burst damage kills. We saw the epitomy of this in ToC, however, Ulduar started this by forcing tanks to time cool downs based on fight knowledge, while ICC is more of a relentless assault at near ToC levels more frequently to counter the upgraded gear. There are only two things that help survive burst, which Agg covers, EHP and proper defensive CD timing. Therefore, you gear for burst, and ignore the rest, as "the rest" should be covered by your healers if they have a pulse and are not AFK or disconnected midfight.

Total damage is not a good indicator only because it "dilutes" the burst damage figures, which is why you shouldn't use it for TEH. But there is value in Jere's comment that total damage can tell you something. If the total damage is high, tank survivability is an issue. If total damage is low, your ability not to stand in fire/poison/etc. is the prime concern (like Ulduar days, where everything bolied down to that).

[Aggathon]

For the record: I think that total damage as far as the TEH calcs go is valuable for fights with various phases that can still kill a tank. I think putricide is a great example, I would lump blood princes in with that, obviously that depends on which prince you are tanking though. Dreamwalker might be one, but frankly, that fight is easy as hell and tanking really isn't the main problem there, it's all about healing up VDW. Marrowgar might be one if you calculate bleeds and the fire damage, but frankly I think you can go all in PEH on that one. The only magical burst fights I can think of are TLK, LDW, and Sindragosa.

I'd like to add in again though: I TALKED ABOUT USING SEPARATE BURST CALCS.

If people want to mis-read and misinterpret my work that's their own damned fault.

Insahnity: I didn't go into threat mechanics for 3 reasons: 1) It was long enough as it is 2) it has been beaten to death and 3) I am FIRM believer that regardless of what your threat stats are if you are having threat issues you are doing something fundamentally wrong. If people mis-read my information, then (again) that's their own damned fault. There is also absolutely ZERO reason why a tank that stacks EHP would somehow become squishy, even in a "non-optimal" raid environment. And if their healers are doing something wrong, well then that's the healer's fault, not the tank's, and a tank gearing for something other than EHP will not help that and in fact may hinder it.

As for that second paragraph, I have no idea what you are talking about, or it's relevance in this thread. That and I also think you're completely wrong about a progression raid being able to function perfectly with 4 resto druids and a disc priest. I also do make the ending conclusion that the math breaks down at some point due to the granularity and discrete nature of healing and incoming damage respectively. Even HoTs are granular. That's why I say DPS is Science, Healing is Art, Tanking is Strategy.

I will admit I know the least about healing mechanics, however I have talked this over many times with my Holy Pally GM and he has confirmed every word of it.

Also: if you have an argument or supposition as to why anything I've said might be wrong, please provide evidence, not just anecdotal thought experiments that don't quite fit. If you have a valid concern over something I legitimately either did not address or was wrong about, I'll listen to it. However, I really don't want to go after a whole bunch of "what if" arguments that are barely applicable. Also: I tend to respond better in PMs, and I realize this is kind of the pot calling the kettle black, but please try to break up paragraphs for more coherency.