Tamral

Here is something I wrote to go along with the whole avoidance/stam concept. The simple fact is avoidance alone isn't enough, you need avoidance/expertise/hit.

Currently I run around 16k unbuffed, 27 dodge, 18 parry, 450 block value (a limitation to my gear based on the fact that we are currently only 4/5 and 5/9 right now), with 127 hit rating and 20 or 25 expertise rating (depending on which gloves I wear).

On bosses like KT, where the hits have large gaps between, HP stacking seems to work best. But for bosses like Rage, Archimonde, Kaz'rogal, Azgalor, Gorefiend, or Najentus, which can hit hard and often, some levels of avoidance make life a lot easier. And considering that 1 expertise rating is worth about 11 block value in terms of threat generation, stacking some avoidance with expertise does the following.

1) Limits your parries. There is no faster way to die than to go on a string of parries.
2) Lowers the chance that you will be subject to a crushing blow after a parry because odds are you're avoiding enough that a third free swing in the 5 second period will not catch you without a shield block charge up.

The difference between 18k and 16k is absolutely nothing when the boss is hitting you for 7-9k a pop. Fully raidbuffed, you'll have 23k or so? I've got around 21k.

If a bad silence occurs, the odds of you absorbing hit, hit, hit with a parry in between the speed the hit up is far higher than someone with appropriate avoidance. Neither of us has even a prayer to survive 3 hits anyway, so the extra 2k HP mean antpiss.

But in reality, the greatest stat we have now is expertise. Taking parries off the table is the fastest way to avoid spike damage. Unless your healers suck, the only thing that is going to get you in the long run is the spike damage. You take a nice combo of hit, parry, crush, and its not likely you can survive it.

Avoidance Stacking:
By general theory, avoidance stacking past a base of 62% raidbuffed (which is about 60% avoidance on a boss 3 levels above) is detrimental to tanking, especially considering usually thunderclap and either a demo shout/CoW is on the mob.

But in terms of actually looking at stats and how they affect theoretical survivability and threat generation:

As an example in gearing I will use the following:

Base:

500 defense
20% dodge
20% parry
30% block
700 block value
6 expertise (from spec)
0 Hit Rating

In terms of SURVIVAL, there are a couple of main goals.

1 ) The first is to limit the probability that you are hit 3 times in succession. This, coupled with any parries to speed up the incoming damage, greatly increases the chance of a crushing blow landing. In other words, reducing the spike damage probability cycle.

Given the above base, the overall avoidance is just under 50%. The odds of being hit 3 times in succession are 1 in 8, or 12.5%. This is given more directly by (1 – x)^3, where x is the raidbuffed avoidance against a mob 3 levels above you.

The marginal gain at this point of 1% additional avoidance is around .8% (meaning 51% avoidance results in the chance of being hit 3 times in succession is reduced to 11.7%). We’ll save this formula for later. Because the gain/loss is NOT linear (it’s a quadratic gain curve (the dertivate of the basic cubic function), the only thing that means anything is the base formula.

2 ) The second is to understand that on average, you will be attacking the mob at an approximated rate of 1.25 times per second (one warrior ability every 1.5 seconds with 200 ms lag, and one normal swing/heroic strike every 1.6/1.7 seconds), so you need to limit the number of cycles that result in parry/parry on successive swings (which results either in 2 sped up swings from the target, or potentially one insta-swing/one 40% hasted swing). Either way, it results in the warrior taking one additional attack in the 2 attack cycle (you get swung at 3 times in the time it would normally take a mob to swing twice).

It’s been debated, but the number I see make the most sense for boss parry rate is 10.6%. Assume E is your expertise rating. The boss parry rate is best approximated by (10.6 – E/4)%. The chance for a parry/parry with the base (6) is .83%, or given the swing state of 1.2 attacks/s, you will incur a cycle approximately once every 100 seconds. We’ll use that as the base for later as well.

Over the course of 100 seconds, it is expected that with a base expertise of 6, you will incur 1 parry/parry cycle and a total expected number of parries of 10.

The formula for this is:

Expected Time per parry/parry cycle = P = ( 1/ [(.106 – INT(E/3.9425)/400) * 1.2)]
Expected Number of Parries per cycle = P * 1.2 * (.106 - INT(E/3.9425)

3) Stacking certain levels of Stamina is just important period. There is no number that isn’t enough. Same with armor. Fortunately, most of the itemization points that you get are stuck in these two stats, so generally, the upgrades themselves will stack these two stats in abundance for you anyway.

Threat Generation:

Based on known formulas:

1 Weapon Expertise rating generates approximately 1.7 TPS up until weapon expertise = 22 (all block/dodge off the table), and approximately .7 TPS afterward until weapon expertise = 42. After that there is no more benefit.

1 Block Value provides a static TPS increase of .16

1 Strength provides a roughly static increase of .2 for prot builds. Yes, strength does have a scaling element based on hit rating, crit rating, expertise, and weapon speed (for devastate calculations), but barring something stupid, its in the ballpark of .2.

1 Hit Rating provides .7 TPS until you cap at 142 hit rating. After that it provides 0.

With 6 expertise, 12% crit and 0 hit rating, the hit chart is as follows:
9% miss
4.1% dodge
9.1% parry
4.1% block (when applicable)
21% glancing
40.7% hit
12% crit

[top] 26.2% miss

With 22 expertise, 12% crit, and 142 hit rating
.1% dodge
.1% block
5.1% parry
21% glancing
61.7% hit
12% crit

5.3% miss

Just as you want to minimize the chance that you incur a hit, hit, hit, you ALSO want to minimize the chances of being missed three times in a row as well. Given a base cycle of SS, Revenge, Devastate, Devastate, 4 consecutive incoming misses leaves you rage starved, resulting in either you having no threat generation until you are hit or worse, you have no rage to hit the shield block button, risking getting 2 shot.

The odds of 3 misses in a row = x ^ 3, where x is your avoidance. Now, also with this it is important to note that this rage starvation is generally mostly mitigated by by ensuring that you are not also consistently missing attacks. (giving you a more reliable second source of rage). The estimation most relevant is that you want to limit the chance of 3 consecutive incoming misses combined with you missing 2 consecutive attacks. (that will allow for at LEAST shield block to be active)

In the base stats example, this chance would be:

(.5 ^ 3) * (.262^2) = .8%

If that % is anything over .4%, its bad, because that translates into more than 1 expected such moment per 5 minutes.

Assuming you stack just avoidance to 60%

(.6^3) * (.262^2) = 1.5%

Clearly, stacking JUST avoidance leads to bad results. 1.5% means about every minute or so, the tank will either be forced to generate 0 threat from abilities for a few seconds and/or not have shield block up until after the hit. Either way, that tank will find him/herself SOL a lot.

Rage Generation:

Quite Obviously, a Warrior needs rage to tank. Considering a typical big hitting boss whacks on plate for around 7-8k (with debuffs up), and swings approximately every 2.4 seconds (with imp thunderclap up), the base incoming damage to the warrior (assuming no crushes, because I don’t have the time or inclination to deal with a first order partial calculation), of 3k DPS.

Given the known rage formula, that is about 30 rage per second incoming absent avoidance. At 50% avoidance, that number becomes 15 rage per second incoming, and at 60%, 12 rage per second incoming. The formula to remember is approximated by:

Incoming Rage from Damage Taken = 30*x per second

Since the cycle for an average tank is 6 seconds, (SS, Rev, Devastate * 2, Shield Block * 1.2) The base rage required to sustain the cycle = 49. (assuming imp sunder armor).

Now, there is also the measure of net damage from the heroic strike/autoattack cycle.

Every successful auto-attack results in 6 rage being generated. Every heroic strike results in a loss of 9 rage (assuming imp heroic strike)

The net rage gain/loss is:

autoattacks *6 – heroic strikes *9
Subject to
Autoattacks + Heroic Strikes = (6/weapon speed)

So, every 6 seconds, we have to balance the net:
180 * x + autoattacks (1 – miss%) * 6 > 49 + H.S (1 – miss)* 9 + H.S (miss) * 2
Subject To:
Autoattacks + H.S = 6/weapon speed

Hope some of that helps.

If anyone wants to tackle the specific math (I would recommend using a program like Latex, MathCAD, or anything else that can solve partial matrixes with a supplied determinant) themselves, I’ll be happy to give you the format of the n-bound matrix to plug the numbers in.

Keep in mind, solving this via matrix leads to DIFFERENT results for every boss. (the final column needs to contain boss swing speed, average damage, average crush, 0 and relative parry rate) The primary input (since you are solving an n-bound nonlinear with a known determinant (and yes, the equation is ALWAYS a second degree exact equation), you will get a different optimal set of numbers for avoidance, expertise, hit, shield block, and stam. It will give you the relative importance levels of each stat from 1-100. (you get this by basing the matrix subject to those variables all being greater than or equal to 0 and adding up to 100.

If you want to calculate the determinant itself, it is based on the itemization base for each extreme (nothing but stam, nothing but avoidance, nothing but shield block, nothing but hit, nothing but expertise, all sitting over the base of 15000 health, 50% avoidance, 300 shield block, 0 hit, 6 expertise).

To give you the answer in layman's terms for any hard hitting boss (average of over 8k per hit, 12k per crush), it will yield ballpark 60% avoidance, 43 expertise, 300 shield block value, 91 hit rating, and 17.5k ish unbuffed health)

Since those numbers are theoretical and generally not easily acheived, many people substitute a bunch extra + hit for the expertise that cannot be gotten, get some of the natural shield block value that comes with some of those pieces (Gauntlets from Teron, Myrmidon's Treads, etc,), and stam whereever else they can get it.

But the bottom line. At base damages that are low, shield block value is more important. At a base damage of say, 3k per hit, the raw shield block value number is 3.8, meaning 3.8*350 = 1330 is the ideal number. Once base damage starts to creep up over 5-6k, there is not a single instance where weighted shield block value would be over 0. At the 9k level of base damage, the raw number yielded is in the -1.8 ballpark of shield block value, meaning ideally, you would lose 630 shield block value ( to have a -280 shield block value), just so you can use that to get more weighting elsewhere. It is all about regulating the damage that comes in as best as possible, and limiting the spike damage scenarios. The risk of spike damage being catasrophic on a boss hitting for 3k is next to nothing unless under extrreme duress (like RoS phase 1), whereas the odds of it being catastrophic on harder hitting bosses is much much more likely.

 

• Jamor

Yeah, I wear my current max expertise on that fight (-6.25%) - also have warriors doing Demo shout. Do your warriors intercept back to him on air burst or take it? That might be where the difference lies.

 

Tamral

Ultimately, skill always > gear. Gear will greatly enhance your ability to display skill, and the combination of both leads to ezmode raiding. Ultimately, you can try to tank all bosses in a single set of gear, but that ultimately results in making some bosses for which your gear is not perfect much harder. The best players will ultimately overcome it, but just because you’re that damn good and you can eat soup with a fork, doesn’t mean soup is meant to be eaten with a fork. You’d be amazed how much easier it is with a spoon.

 

• Jamor

Not exactly sure what post this is in response to. Was it mine above, or just some random thoughts from you?

 

Tamral

No, it was a follow up to my own previous post above yours. It's just that when I started typing it up, your post below mine didn't exist yet. One cigarette break and the submit reply later, you have a statement completely out of context.....

 

Ihmes

Nice wall of math in your former post, crit me for 9001 =D

Anyhow, about skill...

Is it just me, or is tanking really that hard? What you need is just proper UI, some situational awareness and not sleep on your keyboard. I have a confession to make though: I use macros, and I use them a lot. It just makes things so much easier, even if they're not always the optimal way to do things. For example:
Binds both important debuffs conviniently on one button, less bars etc, helps keep up shield block and heroic strike, while giving me free finger for movement (strafe in Q with mouse). Also applying debuffs won't affect shield block uptime. This is great in limitless rage situations, and what's best, it allows acceptable level of tanking even when tired. We're still humans, right? I know this doesn't offer the best TPS in limited rage situations, that's when I spam it less and use more Devastate and Shield Slam (and those are just the skills, no /castsequence macros). That macro is clickable with less than 10 rage, if GCD isn't rolling, it just makes revenge since there is no rage for SB or HS. So if I'm fully awake, I can limit the effects on TPR with my awareness on my rage. If I'm not awake, I can still get by with simple buttonspam.

That kind of macro is possible since SB and HS are not on GCD, and there was such a macro in Quigon's original post.

Actually, now that I think about it, my whole setup is /castsequence macro. I push buttons 1,2 and 4 in succession, like dancing my fingers on the keyboard. 1 is bound the macro I gave, 2 is Shield Slam, 4 is Devastate. Spam the whole set, it becomes: cast revenge if GCD not rolling, cast SB if enough rage, cast HS if enough rage, cast Shield slam when GCD not rolling (that means revenge is on it's CD), cast devastate when GCD not rolling (Shield Slam and revenge on CD). Then just move little finger for shift and ctrl when it's time to apply debuffs.

Hopefully I won't be flamed by non-macro puritans who raid with default UI :<

And for reference, I *could* bind all those skills on different buttons, waste lots of space on my display and lose some awareness for "purity". But it's my choice not to, this way I can keep stuff simpler and let me concentrate on the game.

 

◊ Quigon

Tamral, I think your approach to tanking is exceptional. It is good to over analyze things as a tank - skill is hugely important, but if wearing that ring in your bag actually is better, and the only reason you're not wearing it is because you're too lazy to do the math, then really, shame on you as a tank - and it is certainly not going to help your progress.
Note, that is not meant to be a knock on anyone in these forums... just saying great post by Tamral.

As to Ihmes.
Is tanking hard?

No not really - but neither is any class. But just because it isn't hard, doesn't mean that everyone performs at even remotely the same level.
Tanking is probably one of the more intricate roles, as mistakes here are much more likely than any other single individual to wipe a raid.

Being good at this game generally transcends your class.
Being good involves awareness:
1. Knowing what abilities to use, and why.
2. Knowing what is happening
2a. What the boss is doing, why, and when, and how often, and what/when to expect more.
2b. Knowing what you're doing (sounds weird, but this is probably the most key point - movement and adjustment!)
2c. Knowing what OTHERS are doing. Many top-end guilds will readily accept raid leaders and guild leaders of other guilds who apply simply because if you can manage others, you can likely manage yourself. Doing well as a raid member involves managing your own contribution, and also understanding the motions and actions of other players... A good healer for instance, MUST KNOW when to compensate for a loss or adjustment! Tanks are no different.
3. Just being alert, not afk, all that jazz.
Probably 4-10 that I'm forgetting now.

But the actual skill of pressing 544542545422444444544522534424443354254434444444425422244444445444452 (4 is clearly HS) is not that great. Adjusting, awareness, knowledge is. MT's should know what they're doing. They have a huge impact on the raid.

And yes, there are good tanks and bad tanks. There are even exceptional tanks. As a raid leader, especially one who had to deal with 8 tanks of different quality on 4 Horsemen - you can tell immediately every level from lets say a rank of 0 to a rank of 100 (Gamespot scale here, Eidos makes it so I cannot score below 70 though). There really is a lot that personal skill can contribute to the class.

PS: I agree gui and macros are important. A great tank can become even better if he/she customizes their gui in a way that matches their own personal preference - if well done you can consolidate information and act more efficiently. Yes, the default gui is sufficient; but just because you do something and it works, doesn't mean it is the best way to go about it. I think about half of the demotivators address this issue - just picking one randomly:
http://images.despair.com/products/d.../tradition.jpg
And we know if it is on despair.com, it is fact.

 

Flurry

On bindings, am I the only one who uses the mouse wheel? I've got four abilities bound to it, using Up, Shift+Up, Down, Shift+Down.

 

• Goatbert

I've heard of people using the mousewheel but I still like it for camera adjustment. Gnome POV sucks.

As for my bindings, mine work great for me but I have had to adjust some for performance. Archimonde killed way fewer people when I went from Alt-F3 to change to berserker stance and Shift-1 to berserker rage to double tapping ~ for instance.

Habits are fine, being unwilling to change them is bad.

Speaking of which, I was big into max stamina pretty much our whole way from SSC to BT and I have to say reading this I think at times I did us a disservice because there are several fights that would be easier on healers had I been a little more open to avoidance. That said, we didn't have a big problem with tank deaths except on Morogrim but it would have been less stressful on them. I'm in the midst of putting together an avoidance set so that my three tanking sets look like:

TPS Set: Block Value, Hit, Expertise. I don't have the Tier 6 chest yet so I'm sticking with 2 piece tier 5 for this set until I can do 4 piece Tier 6 while keeping the Gorefiend gloves on. [Gnomeregan Auto-Blocker 600] for trinket. Thinking of putting my [Drake Fang Talisman] in the other slot for lack of [Romulo's Poison Vial] but not sure.

Avoidance Set: Same as above but with the [Scarab of Displacement] and [Moroes' Lucky Pocket Watch], essentially. Probably a [Rifle of the Stoic Guardian] once I get one.

Stamina: [Faceplate of the Impenetrable], [Praetorian's Legguards], stam trinkets. Mostly for Illidan at this point.

 

Tamral

Just to give you an idea of how wide the variance is for bosses, here are the recommended gearing levels for every boss I have come across, from SSC to TK.

The following are used to create the boundaries:
I’m not including the effects of incoming crits, simply because if you’re not stacking 490 defense already, you’re too stupid to understand any of this anyway. If you’re stacking INT or Spirit and want to see those in the calculations, please direct your concerns to barrens chat in your local realm.

Initial Assumptions

500 defense = Df0
15% dodge = Dg0
15% parry = Pa0
25% block = BL0
300 block value = BV0
6 expertise = EX0
0 hit rating = HT0
10% crit = CT0
200 Strength = ST0
150 AGI = AG0
550 AP = AP0
15000 Health = H0
17000 Armor = Armor0

Other values:
BP = boss parry rate
BD = boss dodge rate
BB = Boss Block rate
MI = Static Miss Rate of incoming attacks
MO = Miss rate of your attacks

Marginal Values:
Value of increasing each value by 1
dDFrating = +.016Dg0 + .016Pa0 + .016BL0 + .016MI
dDgrating = +.051Dg0
dParating = +.042Pa0
dBL0rating = .128BL0
dExrating = - int(Exrating/3.924)*BP - int(Exrating/3.924)*BD - int(Exrating/3.924)*BD
dST = +2 AP + (ST/20)BV
dAG = .033Dg + .033CT
DSta = + 10.5 Health
DST = + 1.1 ST

Using the basis of the following for comparison

1 Itemization point = 1.5 stam = 1 agi = 1 str = 1 crit rating = 1 dodge rating = 1 hit rating = 1 parry rating = 1 block rating = 3.7 block value = 1 expertise rating = 1 defense rating = 12 Armor

Without doing 46 pages of math, I concluded that the full complement of all BT epics (all level 141 items) provides a grand total of 3400 itemization points. Considering the very base of these numbers consumes roughly 2300 of the itemization points (a good 900 of them in stam alone, 300+ in defense rating, +150 or so across other stats)

Therefore:

+.67 DSta + DAg + DST + DCT + DDg + DHT + DPa + DBL + .28DBV + DEx + DDf + .08Armor = 1100
S.T
All >= 0
All <= 550 (Simply limiting you from stacking any one stat past this point (its impossible to do anyway, but it provides a force cap to give better relative values)

That’s the simple part.

Now it gets a bit more complex, so we’re going to summarize some of the data together (and yes, I realize I am combining dodge and parry together under the avoidance hat, and anyone with a brain realizes that parry is better in equal quantities, but it makes the math actually doable)

First, we need to provide proper caps where applicable;
Ex <= 43
HT <= 142

Avoidance = A0 = Pa + Dg – 1.2 + 10 + (DF-500)/25
Boss Damage = Boss Damage * .9 * ( 1 – (Armor/(11960+Armor))

Character hit chart:
{
A0 0
BL Boss Damage - BV
Max(0, 102.4 – A0 – BL - 15) Boss Damage
Max(0, 102.4 – A0 – BL) Boss Damage * 1.5
}

At first glance you may think it best to limit the expected damage, but that is not necessarily the case. We’ll get into that later:

First, we need to measure every stat relative to Avoidance, and Threat where applicable, because we are going to turn the entire equation into a second order exact equation, and second, we want to control the determinant (we want to be able to spit different ideal stats out by supplying the boss stats, not determine what boss our gear is ideal for (because the answer to that will always be an endpoint of 0 damage per swing, infinite time between swings, etc.)

The base values above, as per any normal threat spreadsheet, will be able to give a base of around 900 TPS based on 200 ms lag, using a SS, Rev, Dev, Dev cycle, with 2 heroic strikes per 4. So, using the same kind of calculator, we get marginal values of:

1 Agi = .033 A0 - .008 TPS (Agi leads to a relative decrease in threat because .05% crit provides less threat than .033 dodge removes)
1 Str = 0 A0 + .2TPS
1 Ex = 0A0 + 1.7 TPS
1 HT = 0A0 + .7 TPS
3.8 BV = 0A0 + .6 TPS
1 Dg = .051A0 - .31TPS
1 Pa = .042A0 - .12TPS
1 CT = 0A0 + .44TPS
1 Df = .047A0 - .24TPS
12 Armor = 0A0 - .00045TPS
1BL = 0A0 - .0000081*BV TPS (being it I don’t want to have to use Fourier to estimate this BS, I’m going to revolve this around a static number)
Therefore
1BL = 0A0 - .005 TPS (revolving around 600 Block Value, because this number is so damn small it means next to nothing anyway)

TPS0 = 800 (the value any tank who has l2played can maintain)
A0 = 39.2 (15 dodge, 15 parry, 500 defense)

Now, going back to the damage formula.

First, we need to find the probability within a 30 second cycle that the tank will be subject to a crushing blow.
2 attacks per 5 are off the table

Now, we must determine how many free/extra swings the boss will have.

Using a simple estimation of 1.6 weapon speed, and the fact that parries can occur at any time within a boss swing cycle (2.4 seconds)
It is important to know that once every 60 seconds, regardless of the situation, you will be subject to a crush simply because one swing every 2.4 = 25 swings per minute, while your shield block covers only 24. One 5 second gap per minute is naturally uncovered.

12.8% no effect
25.6% chance swing time reduced by .4 seconds
61.6% chance swing time reduced by .7 seconds
Just multiplying that out, each parry results in the boss getting a swing time reduction of about .55 seconds.

30 seconds = approximately 38 swings.
38 * BP = number of parries.

At the base level, this number is around 3.4 parries per 30 seconds, which results in the boss getting one extra attack per 37 seconds (1.9 reduction in swing time) total over the course of 30 seconds. However, this is misleading. The reality is that the extra attacks are not evenly spread out, and it actually results, in on average, 1.41 swings per 30 seconds that are potentially a third in a 5 second gap.

On average, absent expertise rating beyond that provided by talents, the boss will get one extra swing at the player every 37 seconds. That extra swing is guaranteed to provide one extra attack over that 5 second period. The likelihood that this attack will crush is 15% of the chance that the other two attacks were successful hits, or

(1 – A0)^2 * .15 = .054 at the base level. About 1.41 times every 30 seconds, you’re running a 5% chance of potential catastrophe against a harder hitting boss by stacking nothing but stam (hit, hit, crush in a 5 second gap). Yes, this can be healed through and survived MOST of the time, and you may survive, but congratulations, everyone else’s life is harder because you’re stacking your gear poorly.

At 60% avoidance and 22 expertise, the one extra swing in time is every 72 seconds, and you’re only subject to .5 swings per 30 that fall outside your coverage gap.
(1 - .6)^2 * .15 = .024. So now, instead of a 5% chance of catastrophe 1.9 times per 30 seconds, it’s a 2.4% chance 1 time per 30. it may not look like much, but try asking the guy underwriting life insurance how different those two sets of numbers are. How it affects odds of success over a 7 minute fight is absolutely astronomical.

Assuming starting values:

20000 raid buffed health
Assuming HPS

State 1 Initial: 20000 health
State 1 post initial = return to state 1 health
State 2 = {A0 min (State 1 health +2.4 HPS, 20000)
1- A0 min (State 1 health +2.4 HPS, 20000) - Boss Damage + BV
}
State 3(A0) {A0 min (State 1 health +4.8 HPS, 20000)
1- A0 min (State 1 health +4.8 HPS, 20000) - Boss Damage + BV
}
State 3(BL) {A0 min (State 1 health +4.8 HPS, 20000) – Boss Damage + BV
1- A0 min (State 1 health +4.8 HPS, 20000) – 2 * Boss Damage + 2 *BV
}
State 4 (no 3 swings)
{ 1 Return to State 1 + .2 HPS}
State 4 (chance of cycle having 3 swings)
Subset 1:
(A0 * A0) ) {A0 min (State 1 health +4.8 HPS, 20000)
1- A0 min (State 1 health +4.8 HPS, 20000) – Boss Damage + BV
}
Subset 2:
2* (A0 * (1 –A0)
) {A0 min (State 1 health +4.8 HPS, 20000) – Boss Damage + BV
1- A0 min (State 1 health +4.8 HPS, 20000) * Boss Damage + 2 *BV
}
Subset 3
(1 – A0)^2
{A0 min (State 1 health +4.8 HPS, 20000) - 2 * Boss Damage + 2 *BV)
BL min (State 1 health +4.8 HPS, 20000) – 3 * Boss Damage + 3 * BV
Max(0, 102.4 – A0 – BL - 15) min (State 1 health +4.8 HPS, 20000) - 3 * Boss Damage + 2 *BV
Max(0, 102.4 – A0 – BL) min (State 1 health +4.8 HPS, 20000) - 3.5 * Boss Damage + 2 * BV + 4.1 HPS
}
If health = 0 then, end
Else, return to state 1 + .2 HPS

Boss Damage in this state is the mitigated damage taken based on armor, I simply didn’t write it in because I didn’t feel like typing it out and clustering it with something that is already obvious if you’ve been following along so far anyway.

Your chance of the cycle of having 3 swings is simple. Divide the base number of 3 swing cycles per 30 by 6. That is the chance of any given cycle putting you into one of the subsets of State 4. Subset 3 is basically known as the catastrophic state, because that is the state at which you are in danger of taking more damage per that 5 seconds than you are normally getting healed for, meaning either
a) you just died, gag
or
b) you are going back into state 1 with significantly less than full health, so either catch up time/Loch/etc. is needed before you are relatively safe again.

Assuming you are receiving in the ballpark of 4500 HPS (3.5 direct healers)
Plug in the numbers. For a boss hitting for 8k per hit, 500 block value, and plug into the 4x4 matrix, and knock yourself out. 20,000 health with 22 expertise, 60% avoidance is SAFER (less likely to reach health = 0) than 25,000 health, 6 expertise, 45% avoidance. In fact, at 6 expertise and 45% avoidance, you would need almost 27,800 health to get to that same level of relative safety, due to the sheer number of times spent over 7 minutes in subset 3 of State 4.

Take the above matrix and subject it to the following:

Min P(health = 0, (20000 + sta * 1.15, 39.2% + .033Agi + .051Dg + .042Pa + .047Df , 25% + .128BL)
Max: 800 - .008Agi + .2 Str + 1.7 Ex + .7 HT + .16BL - .31Dg - .12Pa +.44CT - .00003 Armor - .005 BL
Subject to:
+.67 DSta + DAg + DST + DCT + DDg + DHT + DPa + DBL + .28DBV + DEx + DDf + .08Armor = 1100
All >= 0
EX <= 43
Hit <= 142
All <= 550 (Simply limiting you from stacking any one stat past this point (its impossible to do anyway, but it provides a force cap to give better relative values)

Ok, looking over this, we get the following:
A 4 State matrix with not enough static equations to solve (it is a ridiculous partial that isn’t even worth looking at). BUT… well at first glance, we can eliminate several of our variables.

Under any circumstances, the following relative dominances can be made:
Expertise dominates hit and crit. Since 3.8BL is the same as 1 crit rating in itemization, BL dominates crit. Therefore, under all circumstances, crit = 0.
Strength is dominated by everything. It is quite possibly the most worthless stat out there for a protection warrior. It is dominated by everything. Therefore, STR = 0.
None of these stats provide any avoidance, the relative gains in these stats therefore can be measured in a linear fashion against one another.

Until we can plug in boss damage, we don’t know about the relative values of anything that provide both avoidance and threat results. Furthermore, itemization points spent in armor past the basic level of 17000 is such a waste it defies belief. If you get a +armor enchant or kit, go shoot yourself now, because it’s beyond idiotic. You’ll get 18k or so armor anyway just from the items themselves, but the relative value of armor in itemization is literally nothing compared to everything else.

Put relatively, the gains from 100 Stam vs 1200 Armor is the difference between 1050 health, vs, (at 17k armor), a 1.5% decrease in damage taken (60.2% vs 58.7%), or about 350 less per hit. Since the objective here is to limit any possibilities of 3 consecutive hits, a stat that does not pay back dividends until AFTER that happens is pointless. Oh, and the State matrix tells us something else. BL doesn’t even matter until you’re in the oh shit state anyway. It doesn’t otherwise do much of anything. Therefore, byebye.

So at this point, knowing only that armor, str, crit, and BL are equal to 0, we have to actually now solve this matrix. Still seems impossible… BUT, we can write relative formulas as well. Especially since we did that cool thing before writing out avoidance, right? Not to mention, a quick rewrite to make all this shit a bit easier on the eyes:

Min P(health = 0, (20000 + sta * 1.15, 39.2% + .033Agi + .051Dg + .042Pa + .047Df , 25%)
Max: 800 - .008Agi + 1.7 Ex + .7 HT + .16BL - .31Dg - .12Pa S
Subject to:
+.67 DSta + DAg + DDg + DHT + DPa + .28DBV + DEx + DDf = 1100
All >= 0
EX <= 43
Hit <= 142
All <= 550 (Simply limiting you from stacking any one stat past this point (its impossible to do anyway, but it provides a force cap to give better relative values)

Still we are a bit short on equations, and this is still a clusterfuck of a partial to even attempt. But, we’ve got some abbreviations here. First off, because the minimization equation dominates here, and because I know something else you don’t know…… No really, from before since we know that Expertise dominates hit and block value, and that hit dominates block value….
Rewrite it to be:

Min P(health = 0, (20000 + sta * 1.15, 39.2% + .033Agi + .051Dg + .042Pa + .047Df , 25%)
Max: 800 - .008Agi + 1.7 Ex - .31Dg - .12Pa S
Subject to:
+.67 DSta + DAg + DDg + DPa + DEx + DDf = 1100
All >= 0
EX <= 43
Hit <= 142
All <= 550 (Simply limiting you from stacking any one stat past this point (its impossible to do anyway, but it provides a force cap to give better relative values)

AHA! A 4 State matrix, 2 subject equations, and known endpoints! We can now solve this. To make a long long long long story short: Expertise caps first in every circumstance from Boss damage =0 to 99999 and any other parameter you set. It then allows us to rewrite the equation as follows:

Min P(health = 0, (20000 + sta * 1.15, 39.2% + .033Agi + .051Dg + .042Pa + .047Df , 25%)
Max: 800 - .008Agi + 1.7*43 + .7*HT + .16BL - .31Dg - .12Pa S
Subject to:
+.67 DSta + DAg + DHT + DDg + DPa + DDf = 1057
All >= 0

EX = 43

Hit <= 142
All <= 550 (Simply limiting you from stacking any one stat past this point (its impossible to do anyway, but it provides a force cap to give better relative values)

First, solve the relative partials from above with everything else constant to get your variance rates. The reason you are doing this is because you’re not creating the ideal gear set, you’re creating the a means to measure the relative values of increasing stats BY BOSS (I can’t stress the need for different sets for each boss enough here)
Ok, now, you’re supplied multiplier row in the end matrix (6 x 6) x (1 x 6) is (Boss Damage, 2.4, 10.6, 26.2, 0, (your weapon speed).

The lighter the boss hits, the higher the value for things like Block Value (and Block rating, even though I zeroed it out). Stam has an exponentially decreasing rate of return. For example, against a boss hitting for 10,000 a pop, stam has a very high rate of return until about 22,800 HP. For a boss hitting for 9000 a pop, the point at which it gets overtaken is closer to 20600. These are raidbuffed numbers by the way.

So, going off of the original BASE numbers, here is what the ideal scenarios look like by boss (other than Resistance Fights), and they are estimated to the nearest 1/10.

Again, this is what my spreadsheet spits out, and I take into account a lot more than just whats up above, including specific threat dumps, additional attacks, silences, fights where threat is not an issue (karathress) and DoTs. I have not done leo or vashj simply because the number of factors going in is too high)

Lurker: 55% avoidance , 43 expertise, 142 hit, 17000 armor, 850 block value, 16,400 health.
Morogrim: 61% avoidance, 43 expertise, 0 hit, 17000 armor, 350 block value, 17,600 health (I inserted the quakes as an extra physical attack per 30, and decreased his swing timer to 1.9 to reflect his additional attacks)
Karathress: 61% avoidance, 43 expertise, 142 hit, 17000 armor, 350 block, 16,100 health.
Leo, Vashj: Cannot be done, I’m not even going to bother with the math.
A’lar: 52% avoidance, 43 expertise, 142 hit, 17000 armor, 970 Block, 16,600 health
VR: 49% avoidance, 43 expertise, 142 hit, 17000 armor, 990 block, 81 crit, 15,000 health.
Solarian: who cares
KT: 58% avoidance, 43 expertise, 0 hit, 17000 armor, 350 block, 18,400 health
Rage: 61% avoidance, 43 expertise, 142 hit, 17000 armor, 350 block, 16,100 health

I’m not going to go on simply because of the amount of time it took me to create the spreadsheet, and at least as of now, I consider it a general disservice to just spit out numbers in an unqualified manner (because I’m not posting the spreadsheet for a reason, especially since it is somewhat incomplete). I’ll be happy to entertain individuals who want to discuss it and are legitimately interested in helping to tune it.

The main mission here is hopefully 4 lessons:

1) If you are using only one set of gear as your “best” gear for tanking everything, then you should think again, because every situation is different, and sometimes, the gains and losses can be incredible.

2) That stickied “fortifications” post on the warrior forums on the official forums is not only outdated, it’s a piece of shit and a waste of space, and needs to be taken down. That strategy is not only dated, it’s stupid in BC.

3) Hopefully it helps people come up with their own systems of determining whether or not gear is an upgrade not as a whole, but within certain sets. I HATE seeing quality items get DE’d simply because it doesn’t have enough stam, or because it has too much avoidance, or “I’ve already capped this stat”, etc. The fact is, there is not enough plate in the game as is to make ANY of the ideal typesets, let alone all of them.

4) It qualifies what I mean by many warriors right now trying to eat soup with a fork. There are situations gear wise you could run into, where no matter how good a player you are, it won’t matter. You’re leaving too much up to luck and chance, and you’re not giving yourself or your raid a fair shot at the encounter.

 

Reliknom

KT at 58% avoidance and 18400 health seems strange, given that his main damage comes from fireballs and that biggest spike damage comes in the form of the pyroblast. None of these can be "avoided" and his melee attack is kind of girlish.

E: Oh yes, awesome math by the way. I don't say I could follow it through at first read and it's been some time I've had to handle this kind of thing.

"Morituri Nolumus Mori!"

 

◊ Quigon

I don't think much other than morogrim in Tier 5 benefits greatly from gear other than raw unadulterated stamina. Especially KT - where surviving a pyroblast if things don't go perfectly is quite useful.

 

Reliknom

Another thing, in the FLK fight the greatest problem from a tanking viewpoint is not the boss, but his shaman add. So at least some advice in gear selection for this mob would be convenient in the guide. I always went for heavy stamina against him, mainly beacause I've been gearing for Vashj and Kael when he posed any kind of problem for our raid. But I'm curious as to what kind of gear would be optimal in this fight.

"Morituri Nolumus Mori!"

 

Moobinator

I'm getting confused on the +hit, Expertise stats.

This is how I am seeing it:

+hit is your ability to miss an opponent.

If +hit Fails and you miss, then Expertise says what type of a miss it is, (ie. Miss/Parry/Dodge) with the modifier making the chance to miss higher than dodge and parry.


Or is it,

(hit + Expertise) = Hit/Miss ratio?

 

Eyegore

You do not have a set chance to hit an opponent, your chance to get a miss, parry, dodge, and crit are filled out on the combat table, whatever is left over is a regular hit. +hit makes your chance to get a miss smaller, expertise makes your chance to get both a dodge or a parry smaller. Thus both stats leave more space over to be a plain hit.

 

Tamral

Raw, unadulterated stamina is more viable in T5 instances because there is so much greater a margin for error that the benefits of the correct itemization are not as easily seen. The harm in being in the oh shit state (subset 3 of state 4 in most T5 encounters is very limited and generally easily recovered from). However, that's not to say it's ever optimal to be in that state. Given the nature of bosses like Leo and Vashj, with all of the randomness in the encounters, I would agree that most encounters are fine with just stam stam stam. But pure stam stacking for harder hitting bosses is still a mistake, and the math is there.

What is interesting is how much expertise dominates just about everything. To give you the scaling relative benefit, for example, even the encounters where stam can be a huge help (KT, Vashj), expertise off the bat generally has a higher rate of return than the first extra point of stam.

The general rotation of desired is expertise to cap, then some stam up to a point where either hit or avoidance pass it in value, back to stam, then to something else with higher value, back to stam, etc..

Expertise has a very powerful static return of keeping people out of the bad state in addition to threat enhancement. Expertise mitigates spike damage better than any other skill as well.

 

Tamral

Considering he swings really fast, and doesn't hit hard (he overwhelms you with the number of hits, subjecting you to many crushes)

You will be in subset 3 of state 4 a lot, and in that fight, you could have a state 5 and a state 6 within 5 seconds.

In such a fight, stam stacking is also not the best way to do it. You're best bet is actually to stack an asston of block value, avoidance, and of course, expertise. The lower the damage per hit, the more and more benefits you get from block value. The more often you would typically be subject to a crush, the higher the value of avoidance (which keeps you out of the possible crush state more, and is the best mitigation against fast attacks)

 

Fellwraith

If you're talking about the shaman for FLK, he can't crush.

He does have a windfury attack, which is basically 3 hard-hitting attacks at once. You can use stam and AC on him and soak it or you can try to rely on the RNG of avoidance to survive that spike. Without a grounding totem, you will also be subject to 4k+ shocks in addition to heavy fire totem damage. Given the amount of magic damage that you could be exposed to, I really think HP, high AC, and some ironshields are the way to go.

 

Ihmes

We had best results with sta-heavy tank. BV+avoidance tank died to burst when shammy did his 3xwindfury+unavoidable frostshock, 3*~3,5k+4k. I think I had around 19k buffed hp at that time, it did fine. With less than 18k tank died constantly.

Edit: Quigon, just noticed that you didn't name an addon for combatlog. Eavesdrop (ace2) is great, it's basically scrollable combatlog that shows your hits dealt, taken, healing done and healing taken, buffs, debuffs etc. in it's own window. It's easy to configure and it's really good for figuring out what killed you.

 

zork

The shaman is no problem if you use a grounding totem in the shaman tank group (you should keep this in mind for Vashj and Akil'zon too). Keep debuffs/buffs up and use some ironshields. I even use spell reflect here on every cooldown. We do: hunter, pet, shaman, priest, karathress-kill order. So the shaman tank can focus on debuffs/buffs without having to spam max-threat.

Interesting point: On three of our first kills our Karathress tank died while we were killing the priest. What happened was that Karathress ran directly to me and we finished the fight without any further problems. Probably a spellreflected shadowbolt (without damage but threat?!). Other than that it should, in no way, be possible to stay above all the healers.

BLOG [ZIB] _ GUILD [TDMOG] _ INTERFACE [ROTHUI] _ CHAR [ROTHAR]

 

Dynalisia

I seem to remember us giving that a try back in the day, but I don't remember the grounding totem ever absorbing the frost shock. Has this changed or did we simply fail to make it work for us?


How did you know to put reflect up at exactly that moment?

 

Liar

The Shadowbolts don't even hit non-mana users AFAIK. But if you were refering to the Frostshocks: It's just luck, spam SR every 10 seconds and you will catch one eventually (it's not like you need the rage elsewhere if your mob isn't being killed first).

EDIT: Yes, grounding works on the shocks. We didn't figure it out until later either <_<

 

Fellwraith

You can get 50% immunity uptime with spell reflect (5 second duration, 10 second cooldown). If you don't have rage issues, there are times that keeping it up as much as possible isn't a bad call. After you fight certain mobs for a long period of time, you can guesstimate when they're going to cast again (Kael trash and their HoJ's are a great example).

This is true of many mob instant cast spells that don't have a cast time. It doesn't cost you a global cooldown and it's another way to dump rage quickly and possibly build some threat.

 

• Bluefish

Does anyone know which is consumed first, the Grounding charge or the Spell Reflect?

Also, if there's any other non-trash uses for SR that aren't obvious? I definitely wouldn't have thought that Frost Shock would be reflectable.

 

Liar

Grounding Totem.

I don't know what you would consider obvious or not, but SR is really helpful when tanking Phoenixes at Kael since they don't even get consumed for a reflect (they stay up until the SR buff expires). Or to reflect Vashj's shock (which you can as well ground, so I guess if you can ground a spell, chances are you can SR it, too).