Nightcrowler, i've got some questions regarding your latest Bear simulation results on this page.

First I try to understand the practical sense of the HP soft cap. I can understand the meaning from the description at the end of your post. In a real raid environment though, would you say its wise to shoot for a total raidbuffed HP as low as 42k when doing Ulduar?
Your best in slot gear has a total of ~49k HP and thats also pretty much in line with my assumptions for a "good" HP value. So I'm a bit lost how to interpret or practically use the HP soft cap number.

The weapon choice in you best in slot gear is also a little bit disturbing for me. From the discussion at this board i was pretty much thinking that common understanding in the community here is that the weapons dropping in Ulduar are no real upgrade from a progression point of view.
With the softcap given at 42k HP i can understand that [Twisted Visage] or [Lotrafen, Spear of the Damned] is an upgrade to [Origin of Nightmares], but from a pure survival point of view in a progress scenario I'm not sure if it makes sense to swap out the Naxxramas Staff for one of the Ulduar weapons. The Naxx25 staff has armor, which helps quite a bit with physical damage after our nerfs, and it also pretty much has the stamina edga over the others, when fighting bosses with magical damage.
However do you think that one could really get into threat problems when gearing to much into the survival and not using the dps weapons or is it just me being lost in the old "Armor > all" paradigm?

The third and last question would be regarding the weapon enchant. Have you tried to model the new Ulduar enchant ([Formula: Enchant Weapon - Blood Draining]) in your simulation or not?

Thanks in advance for taking the time to answer my questions if possible.

@SS_Keera:
Yes HP soft cap should be probably remade, 42k seems a little low for Ulduar (but enough on most fight), the soft cap as I think about it is a "minimum HP" more than a "maximum HP", actually I'm shooting for 50k buffed for my gear.

As for weapon choice I decided to use a "single mathematical method" for gear chocing. Actually in real game I'm swapping between Origin and Twisted depending on the fight. For istance I usually tank adds in the arena at Thorim and then I tank the boss. I use Twisted for adds and Origin for the Boss. For fights were bosses doesn't hit really really hard and/or trashes Twisted extra agility and extra shield value and time it's better than Origin extra stam/armor also from a survival PoV. In the link posted, only 7.5% of the average item value goes for threat stats so the survivability value is pretty high. It's obvious that you should focus on different stats depending on the fight. For istance extra armor gear is not used so much in my sim/report but there are fights were I'll prefer some extra armor instead of avoidance for example. Tanking (differently from dpsing) is more about changing gear depending on the fight and there is an higher degree of freedom.
As for Blood Draining I can't simulate it without changing a lot of code, actually I've not simulated "healing" so effects like frenzied, blood draining and ImpLoP aren't simulated. Tossing out numbers about them is pretty difficult because healing depends too much from your surranding.

a couple of questions for you nightcrowler;

What is the gem priority now? from 3.0-3.1 I was gemming Str and Str/Stam for Meta activation. Are we back to Agi or ArP as the definitive stat or is Str still the way to gem and why?

also, how do the engineering glove enchants handle in PvE compared to Crusher? there is both the Hyperspeed Accelerators and Hand-Mounted Pyro Rocket. The new Hand-Mounted Pyro Rocket no-longer triggers the GCD and is a 45 second cooldown; and the Hyperspeed Accelerators are 340 haste for 10 seconds with a 1 minute cooldown.

Thank-you ^_^

Thanks for the quick response nightcrowler.

@HP cap:
If you really want to have a minimum value mentioned you can leave it at 42k i suppose, because i think for entering Ulduar it is possible to tank with this amount if your healers are not slacking.
I was just under the impression that the wording "cap" suggests that 42k HP is enough for tanking everything in Ulduar and everything above is "wasted", which would be a bit disturbing as even your BiS gear gives nearly 50k. Therefore the question how you interpret this value.
In the end as with all stats for tanking HP is also very much a preference thing than there is a cap that is set in stone.

Regarding weapon choice. I see you point and can agree on this. Of course its quite difficult or basically impossible to have the one and only best tanking gear. As you said this is very much depending on the fight your doing. Of course I'm using several different sets for different bosses myself. Weapons have anyway a special role in this gear sets as they can be swapped throughpout a fight as you mentioned.

For the enchantment its sad. As I'm trying to find some numbers for quite some while to quantify the difference between Mongoose and Blood Draining. Most of the community seems to be still using Mongoose and suggesting it as best enchantment. Although as with the weapons itself in the end I suppose it makes sense to have also different weapons with different enchantments to maximize efficiency.

Anyway thanks for your effort again and keep up the good work with your simulations and FBN addon.

PS:
@Incontrovertible: the first post in this thread contains BiS gear suggestion for Cat as well as dps stat comparison and everything else your looking for. Maybe you should take a look there. For the engineering stuff you could ask Rawr to help you.

I *knew* I missed a part of the first post regarding stat weighting, silly catfase, thanks =) And since I run Mac OS X I am unable to use Rawr =/

I'm not planing to introduce gloves on use enchant (at least not soon). But basically we are talking about 20 agility vs. those 2 enchants. Actually 340 haste for 10 second every minutes average out at 57 haste. Probably it's a bit better than that because you can use it at full energy and haste stack a bit with itself. Said that, let's have a nominal value of 60 haste. 60 haste gives 1.2*60 = 72 dps. 20 agility gives 1.53*20 = 30.6 dps. So it's pure win.

30.6 dps are 1836 damage on 1 minute. the other enchant you mentioned is 1440 to 1760 damage every 45 seconds if I understand it rightly or 35.5 dps, the difference in this case is minimal and probably will be better the agility enchant for sinergy. So I think that without many calculation the right enchant for an engeenier should be Hyperspeed Accelerators.

P.S.
I also use Mac.


@SS_Keera

As a personal feeling I'll use Twisted Visage + mongoose on adds and low hitting bosses, origin of nightmares + blood draining for hard hitting / silencing bosses, baddly blood draining doesn't dropp for us.

He was hit by 9 flame jets. If he had a 0.75 second auto-attack swing time, 2.5 seconds of being out of melee range per flame jet would add up to 30 lost swings. That seems reasonable from my experience.

What kind of rip uptime would you need for agility to be superior to armor pen? I'm also wondering if a zero FB rotation is better if you're stacking agility.

I don't have much of a choice since I'm a tank first and a dps second, so I have to gem agility if it's something I want to use for both. I'd like to optimize as best as I can though.

it's my first post here so i would like to say hi:]
FeralByNight is a great addon ur doin great job Nightcrowler

i'm wondering that any of u have seen that post:


MMO-Champion BlueTracker - Feral dps supposed to be this high? 8-9k dps?


what do u think about that??
anyone knows that is only a simmulation and dps which is shown on that simulations is almoust unavlible to manage on many of fight on naxx and ulduar
im plain feral dps and highest dps i manage is 8k on thadius, my normal dps is something about 5k maybe my gear isnt as good as in simulations but im sure that geared druid want each that high on normal fights...


btw this is my armoury link:http://eu.wowarmory.com/character-sh...indoun&n=Desta
i will be glad if u will be able to advise my what i sholud change first in my gear


sory for my english:]

I have to say that i'm extremely impressed with the work you've done Nightcrawler. I especially like the work you're doing on cat-dps. Your addon is fun to use, and the suggestion-box is fun to use when you dont want to think for yourlself.

However i did some testing on target dummys using the suggestions FBN addon will propose. Sometimes I felt, that i had too little rip uptime (i've read your thoughts about rip uptime though). Other times I saw that i had 0 debuffs (except fearie fire) on the target dummy. Then the addon suggestion building up the debuffs again, but is it really supposed to be like this? It just feels, wrong. The DPS recorded on recount (selfbuffed) was 4139 over 2,100,000 damage so thats pretty decend.

//Ayriss

@Ayriss:

Actually I've done some tests on dummies myself today (about 40 minutes) and I've recorded 4120 dps, I've about BiS pre-ulduar gear, chest from hodir, staff from XT but I don't have Noble deck nor KT cape, so it should average out and I'm about 100 dps lower than simulated values.

I should tell the thruth to you. I don't know if it's ok that much downtime. Actually inside the simulator (and the addon) I've pushed some pseudo-AI who compute the best down-time of rake/rip/mangle based on your crit/ap/haste/etc.. (so it should works fine also on dummies) but to test if it works correctly I've used full buffed stats, so I don't know if unbuffed on dummies for istance is better to have an higher dot/debuff uptime or there is a different stats allocation and variables unimportant for 25-men becames important (for istance I know I should add a check for bleeds) but I should also say: "Is it really important to do let's say 100-150 more dps on a dummy?" what really matters are raid-dps and is for that kind of things that the code is optimized.

Night: I disagree about your assessment of HP. It only affects your minimum TTL in absolute values. What's the difference between 46k and 50k hp if vezax will still splat you in 3 hits? All it really does is increase your healer's efficiency a small bit by reducing overhealing. Not nearly as much as mitigation & avoidance will. I think healers THINK it helps because they see you take two big hits, an extra 7k on grid "looks" better. But if that next swing connects, you're dead either way.

Yes, if healer mana isn't a concern, why not gem for sta anyway since it works on spell damage too? Avoidance and mitigation will reduce the amount of times you have to test your minimum TTL. Fewer times I have to make my healers go "oh shit" and divert attention away from someone else that needs healing, the better.

That said, I have a high hp gear set for fights like mimiron and high physical damage bosses. But the extra hp is only useful if you reach the next plateau of consecutive hits taken.

The difference between 46k and 50k is that the minimum your healers have to heal you to prevent you from dying is raised by 4k. While "I took 3 hits for 60k and got no heals" situation occurs (and more stam wouldn't help), you may also run into the situation of taking 3 hits for 60k and you got 10k-14k healing.

You look at the first situation when you die on parses, you never look at the second.

10k-14k healing over 3 boss hits sounds more like a healer fail than a "tank geared wrong" fail. Or, it's a fight like Vezax where the healers can't spam the tank. But that's a known situation and it's a fight where I would wear more sta.

Plus, as I said, sacrificing mitigation for sta means you will put yourself in those situations(consecutive boss hits) much more often. So, you can look at it

-60k damage max with only 10k healing and live, occurs 10% of the time
-56k damage max with only 10k healing and live, occurs 5% of the time

Which is favorable, assuming they are completely random events? I'll admit the 10% and 5% numbers are completely fabricated, but I wouldn't be surprised if the difference is that pronounced since the dodge difference is compounded over many hits.

I was looking over the wowhead gear ranking list, and to me it seems like the trinket section is all wrong. I mean, caster trinkets are in the top four, and the novelty tentacle trinket made the BiS. That doesn't seem right at all.

New Rawr version out--
Rawr - Release: Rawr v2.2.3
"Rawr.Cat: Fixes and improvements to rotations"

... just checked my gear and it finally lists ArP over agility:

Armor Penetration Rating 1.29
Agility 1.22
Strength 1.22
Haste Rating 0.93
Crit Rating 0.91
Attack Power 0.52
Hit Rating 0.30

I should have figured that out. Thanks for the help.

So it will be useful in exactly the scenario you first described. But in the period of 2 boss auto-attacks and a special (2 seconds), you might only have one or two heals land on you, depending on the boss.

You're taking 2 situations and defining them as the whole world: either the healers are 100% effective and will heal you through anything (in which case you stack avoidance because the only thing that can kill you is unfair RNG), or your healers are 0% effective and it's not their fault(in which case you use TTL math). The reality is, some heals will land (even just HOTs) in between a first hit at 100% health and a second hit. Some healers will be casting on someone else because you hit an avoidance streak right before the damage streak.

Your gearing job as a tank shouldn't be to make it so it's the healer's fault you die, or the RNG's fault you die. It's to minimize the times you die. Maybe 90% of the time they sustain throughput to get you through 3 hits; but 10% of the time, they fail. If you have 55% avoidance instead of 50% avoidance, you're looking at a 9.1% chance of taking 3 consecutive hits (given a random selection of 3 hits) as opposed to a 12.5% chance. So, 3.4% advantage (ultimately, 27% of the chances of 3 consecutive hits are eliminated). But to get that 5% avoidance post-DR, you're giving up a ton of health- I'm going to guess in the 5k range, if you're converting AGI gems to STA gems.

Finally-- assuming those events happen randomly in the FBN simulator, and the numbers work out right. I don't really think you can attack the basis of the values unless you can somehow dispute Nightcrowler's implementation of boss damage/healing (and since most Ulduar tank damage healing is "autoattacks like a truck", it's pretty accurate at least on the boss damage side, although testing a variety of possible incoming damage profiles might be more necessary.)

The simulator is a lot more evidence to stam being useful even without healer failure, so perhaps you might need to find some more evidence beyond "I don't think it's right".

If a boss has a melee special that is that lethal(which is...Thorim only?) you plan ahead for it. Didn't think of it until you just mentioned it, but Melee-Special-Melee combo is almost non-existent in Ulduar. If you're referring to ANY special that hits hard enough to matter, those are all on a specific timer or close enough that you can cover them most of the time with cooldowns from the tank, or cooldowns from the healer(s) for the ones the tank can't get. And the ones you can't get with healer and tank cooldowns probably have a cast time.

So, yes, I am defining that those 2 situations as the whole world. You cover the spikes you know about with tank and healer cooldowns, all that leaves is consecutive melee hits. 3 melee hits to kill a tank is at least 6 seconds. If I can't get a heal in that time, it's either a fight mechanic(loatheb) and we plan for it(more sta) or the tank healers disconnected/incapacitated somehow(stone grip, which might be another case to wear more sta for that fight). But since we've established the assumption that healer mana isn't a problem, you have assigned tank healers. They should be healing you all the time. Even if they switch off to heal another target(which they shouldn't), that's some poor reaction time if they see you take 2 melee hits(4 seconds) and don't cast an emergency heal.

So, I'm not disagreeing that reducing your chance to die should be a priority as a tank. I'm disagreeing that it's a minutely small chance to die from random consecutive auto attacks, thus sta shouldn't rule the tank-gearing paradigm. It should rather be the "kolo could stone grip the MT healers and cast oblivion at the same time, I'll put on a bit more sta to help the backup healers react". I guess what I'm saying is you should gear with a mind towards something that's always good to have: avoidance and threat(agi in this case) and then work the stamina back in as needed on a case by case basis.

I've read the bear section of the simulator over and can't find any mention of how healing is dealt with in the code. He sets an arbitrary HP soft cap(which is fine) but then goes over it by a fair amount because of his own arbitrary scoring system. Which, again, is fine.

What I disagree with, or can't find anywhere, is how the simulator handles how the tank is healed, which is half the equation. From what I understand, there are no parameters(specified at least) as to how much healing is incoming and how often. What kind of healers? Basically, the simulator leaves out the intelligence of the human player to counteract those bad situations with quick heals or queued-up heavy throughput.

No, the simulator, or at least the number system he used to rate the stats, does not show the true usefulness of stamina beyond the hp softcap. It would be nice to see how increasing sta reduced overhealing compares to increasing avoidance decreases incoming damage. Or the comparative value of both if you have a "lazy" healer(increase his cast time). As I said above, until you can get a "smart" healer into the simulator, you're never going to see the true benefit of stamina or mitigation. At least not from this numbers standpoint.

Is this just a typo or did you mean 50k HP without any temporary health increases? I have plugged both, your BiS gear and the highest stamina gear I could find -mostly furious-, into RAWR and achieving 50k HP was impossible without fortitude, kings etc. (I could not switch gem though (anyone else got this problem with the latest release?) but i doubt they would make up such a huge difference)

I would say 50k unbuffed is not possible, im currently nearly BiS Ulduar + pvp weapon and i'm at 40k unbuffed. Fully gemmed sta (and i'm jc).

Health vs Avoidance is a pretty cool matter to discuss.

Actually when I told "I'm shooting to 50k hp" I mean buffed obviously (It was a typo) and simulated value are fully raid buffed.

Just a little digression: a poster above said "if you are not JC?". Actually there is any change with professions in both simulator and addon. The addon takes your stats to compute the optimal cycle so it automatically take into account your profession, the simulator takes gear/spec/buffs into account, if you want to test "your value" you should only put them into the simulator, for the sake of simplicity I've runned and pubblished simulations with BiS gear (with some exception due to hard modes) and obviosly the best professions (JC+BS for cats, JC+BS or Ench for bears), actually I'm herb+alch but the minimal change due to profession doesn't change a lot the stats output (it doesn't change a lot if you go from T7 to T8 gearing).

Coming back to the main topic.

Gearing and playing a tank is not the same as a dps. If you want to look at dps you can find out right numbers and simply use them to gear up, you can also argue something about movement vs. static fights or adds vs boss fight or still you can prefer stable dps (lower dispersion) than an higher average dps, but the difference is usually so small or so high that unless you are really cookie cutter it's not a problem (changing your gear from agi to arpen or viceversa shouldn't change your dps-meter ranking and any other stat is really lower than them).

When you play a tank, your skill and your healers skill plays a greater role and, in the same way, fight specific issue play a greater role. If the tank die = wipe in 90% of the situations, if a dps does 100 lower dps usually it's not a problem (unless you are touching boss enrage for 2 seconds...)


In my simulation I've not introduced a smart healer directly, I wasn't interested in that (but I'll prolly do that in the near future due to my guildy healers, they want a simulator too and implementing them into FBN sim is the better way to do that).

Basically at the moment you should set the following parameter for boss:

swing speed: (I set it to 1.7 seconds -> 2 seconds with infected wounds)
pre-mitigation damage: (Is set to 70k pre shouts/debuff,armor, etc.. remember that improved demo roar cut it by about 15%)
boss hit/miss/etc
boss parry haste (you can disable it, but many bosses except Kologarn have them in Ulduar, for example the general hits like a trunk and also can parry gibbs you) and it's set to 14% (I've done an extended test about it and you can see it into the combat rating post in the class mechanics forum).

No special attacks are simulated (but the simulator support them).

Actually I was interested into healing only to compute the "chance to die". You can see it into the report.

the time could be set.

Basically it works in the following way:

the time interval is divided by xxx second steps. At the begining of each step the tank is healed to full then it's not healed for the next xxx seconds. If the tank dies add 1 to the counter.

Why I do that?
Because from my expirience in wotlk raiding the only things that could kill you are "Oh shit!" situations. You can't gear up based on your healer spell power or you healer classes. Does you raid leader puts a disc priest on you? 2 pallies? 1 priest, 1 pally and 1 shaman? You can't know that. How much crit they have? How much SP?

In the rating I've used a "modified version" of tank points. Basically my simulated tank points are as follow:

HP*(damage done by the boss)/(damage taken by you)

this is actually the average TTL. Before running the first simulation I was sure that avoidance was the king with this system because this system favour AVOIDANCE. Average TTL doesn't take Oh Shit situation into account and usually (at least it was this way in TBC) avoidance increase TTL more than stamina. With DR on avoidance this is no longer true.

As I said a lot of times, if you look at dps ranking you could be pretty well using it mind-numbing, if you look at tank ranking you must use brain.

As a general discussion you have two kind of thing to worry about as a tank (talking only about defensive problem):

taking less average damage
surviving spike damage

The damage TAKEN can be of 3 kind:

predictable spikes
random


Avoidance (defense, agility, dodge): reduce your average damage taken in an unpredictable way.
Shield (ap, crit, str, agi, hit, etc..): reduce your average damage taken in an unpredictable way (but they are more predictable because usually the avoided damage is low, around 2k, and the uptime is high, around 90%)
Stamina: doesn't reduce your average damage taken but increase your TTL.
Armor: reduce average damage taken in a predictable way.
Expertise: reduce average damage taken going versus unpredictability.

Looking at gear ranking I've used TP because you need numbers for gear ranking. If you want to know how I personally gear up: I've gear throw naxx and I'm now gearing in Ulduar for stamina, expertise, armor.

Reducing average damage taken is important only for healers mana, if the healers can't go oom you can have very low avoidance and be ok. As said by the previous poster, you must have 2-3 healers spamming heals on you, so? Avoidance is only overhealing.

Armor reduce you damage taken in a predictable way, for this reason it's particulary good, stamina is not only the best stat to increase TTL (100 stam = 129 agility from a ttl PoV but 3 stam = 2 agi from a item budget PoV) it's also the only stat you can stack to survive spike damage. The 42k value I've reported as a soft cap is a lower bound value based on the average damage you will take in 5 seconds. This is completly different from the maximum damage you can take in 5 seconds (or any random interval). For istance, suppose that the boss has 2 sec swing.
In 5 sec. you will take on average 2.5 swing. But you can easly take 3 swing (time 0, time 2, time 4) or 4 swing if you are not expertise capped (time 0, time 2, parry -> time 3, time 5). If the soft cap to resist 2.5 swing is 42k, the hard cap to survive 4 swing is 67k, and you can't always use cd in that situation. What you can say is that you'll never be able to stack enough stamina to reach 67k, that's true. But I can also say that in that 5 second time could be the case that a single healer is able to toss you a 18k heal or some hot ticks. With 42k you are still dead, with 50k you can survive.

When I play tank I must assume that other raiders are numbs. I must assume that an healer can DC or simply recive a phone call, I must assume that the dpser doesn't use omen and so on, you can't play basing your game-play around other people skill. Something that I always tell to our RL when he blame the healers for tanks dead: "If a tank die is always the tank blame" (also when I am the tank).

Expertise has a particular role. First of all it's a really good threat stats and the best TTL stats after usual defensive stats. But it's particular role is the following: "it reduce randomness". For basically every boss you can't reach enough stamina to really survive 4-5 seconds without a single heal. But if you are expertise capped (56 exp) you can. Expertise is the only stat in game that let you prevent RNG. In this way it's the complete opposite of avoidance and it's a god-send.


Stamina gemmed/enchanted + gladiator on shoulders:


Boss crit: 0%, Boss hit: 36.6546%, Boss miss: 63.3454%, Boss white dps: 2450.12, Boss average weapon speed: 2.02798 sec., Boss average parried attacks x sec: 1.37913%, Average Boss White DPS increase due to parry: 0.589146%
Chance to die: 0.0160833% without heals in: 5 seconds

_____________________________________

Defensive Stats after the buffs (vs. a raid boss, full stacked raid + Devotion Aura)
_______________________
Bear armor (before inspiration): 31707
Physical Damage Reduction Bear (before inspiration): 70.6267%
Defense: 433
Chance to be critical hit: 0%
Dodge (bear): 49.9781%
Miss: 10.0518%
Total Avoidance (bear): 60.0298%
Total HP (bear): 49455
HP soft cap (bear): 42352.3
Shielded hits: 90.7053%
Inspiration uptime: 99.5267%
Tank Points (hp equivalent with 0% avoidance, 0% Damage Reduction and without parry from the boss) (bear): 699780 +/- 243.94
_____________________________________


Agility gemmed/enchanted + def/dodge on shoulders:



Boss crit: 0%, Boss hit: 32.703%, Boss miss: 67.297%, Boss white dps: 2149.41, Boss average weapon speed: 2.02797 sec., Boss average parried attacks x sec: 1.37781%, Average Boss White DPS increase due to parry: 0.589583%
Chance to die: 0.228933% without heals in: 5 seconds

_____________________________________

Defensive Stats after the buffs (vs. a raid boss, full stacked raid + Devotion Aura)
_______________________

Bear armor (before inspiration): 32355
Physical Damage Reduction Cat: 80.6462%
Physical Damage Reduction Bear (before inspiration): 71.0152%
Defense: 436
Chance to be critical hit: 0%
Dodge (bear): 54.1542%
Miss: 10.1559%
Total Avoidance (bear): 64.3102%
Total HP (bear): 43831
HP soft cap (bear): 41792.1
Shielded hits: 93.0909%
Inspiration uptime: 99.5231%
Tank Points (hp equivalent with 0% avoidance, 0% Damage Reduction and without parry from the boss) (bear): 706968 +/- 268.956



Difference regemming from stamina to agility:

TTL/Tank Points: +1%
Chance to die in 5 seconds: +1323%
Average damage taken: -12.3%
_____________________________________