Graul

Buggy or not, they still usually show an increase in DPS when you input something that looks like it would be an upgrade. Most of the time you can simply eye what's going to be an upgrade because it has higher hit, crit, AP or whatever stat over whatever you are needing to replace. The majority of upgrades are common sense (at least to me) unless you really want to nitpick over 5-10 DPS and would rather wait for a future drop than take something in front of your face. These spreadsheets are meant to be used as a guide and a rough estimate of what you can expect. Are they perfect yet? No, they are still a work in progress, but they are still very useful as they are now. If you feel slighted because you spent DKP on something that gave you 2 less DPS than expected don't blame the authors of these sheets. They never forced you to use them in the first place.

And the reason people always reply with the same "look at the spreadsheet" answer is because the same question is constantly asked and each time the answer is going to be different depending on what someone is wearing.

 

Gruntle

Others have replied already just wanted to comment on this.

And just how do you think people are going to do this analysis that you ask for? By picking numbers at random from their brain based on what they "thought they might have seen" in the last raid they attended. "Someone" will just have to run through the spreadsheet, which are currently the best you can get in terms of estimating dps, guessing doesn't even come close (yes that was sarcasm). Why should someone else do the work when the spreadsheets are very easy to use (and will give you a more accurate answer since you can set it to use exactly the gear your wearing).

"ABSOLUTE best", that is just not how it works. Blizzard has not given us any exact formulae to work with, all spreadsheets and other guesstimates are theorycrafting, i.e. there are no guarantees that anything is correct. Granted, some theories are more secure than others but we just don't know enough to do better than what's in the spreadsheets.

 

Daronsk

In the previous discussion on the expertise cap I had suggested it was 6.5% as I had seen dodges with an expertise rating that gave me 6% dodge suppression. Well I present some proof of this claim:
Daronsk - WWS

This is the WWS from my guilds last BT run (Gorefiend->Illidan). The entire night I was dual wielding my S3 axes as an orc, with shard of contempt, and 2/2 weapon mastery. If the cap was truly 5.6% then with this set up I should never see a dodge as I'm well above the cap at 6%. It turned out that their were dodges(albait a very low %).

 

• landsoul

The following test was conducted to figure out if mobs use the Coefficient for 50% armor reduction per level: Z = 467.5 * [TarLvL] - 22167.5
70: 10557.5
71: 11025
72: 11492.5
73: 11960

... that players do since people have questioned that they may not.

I went into the plateau with ogres just outside of shatt and collected some data hitting bloodthirst against some 71 / 72 ogres of the same type (the ones with the maces).

With 2853 static AP (rampage5 and Bshout) and 980 ArP, I hit each 71 and 72 ogres with a bloodthirst and recorded the data. I then went back and hit each 71 and 72 ogre with a bloodthirst after sundering them 5 times.

Here's the data:

71 nosunder = 854
71 sunder5 = 1012

72 nosunder = 839
72 sunder5 = 991

Now since we are working with 2 unknowns per level of mob (Coefficient and armor value) I had to come up with some math to solve for both, so we 100% know how level effects the coefficient.
[%DamageReduced] = armor / (armor + Z)
[%DamageApplied] = 1 - [%DamageReduced]
[%DamageApplied] = Z / (armor + Z)
[%DamageApplied] = BT / .45[AP]
Z / (armor + Z) = BT / .45[AP]
.45[AP]Z = BTarmor + BTZ
Z(.45[AP] - BT) / BT = armor

First we need to solve for the armor value of the 71 and 72 respectively. Since we don't know Z, we had to come up with 2 sets of data, hence the 5 sunders, and then eliminate Z and solve for A.


71:
Z(.45[AP] - BT) / BT = armor-ArP-Sunder
Z = BT(armor-ArP-Sunder) / (.45[AP] - BT)

Z = 854(armor-980) / (.45[2853] - 854)
Z = 1012(armor-3580) / (.45[2853] - 1012)
854(armor-980) / (.45[2853] - 854) = 1012(armor-3580) / (.45[2853] - 1012)
armor = 6556

72 same method:
839(armor-980) / (.45[2853] - 839) = 991(armor-3580) / (.45[2853] - 991)
armor = 6854


Now we can re-plug these back in to find Z for both levels. And we can check it for all 4 levels of data.
Z = BT(armor-ArP-Sunder) / (.45[AP] - BT)
71:
Z = 854(6556-980) / (.45[2853] - 854) = 11078
Z = 1012(6556-3580) / (.45[2853] - 1012) = 11078
72:
Z = 839(6854-980) / (.45[2853] - 839) = 11078
Z = 991(6854-3580) / (.45[2853] - 991) = 11078


Hmm something definately strange going on here. I did re-double check everything and it worked out. There's no change in armor coefficient per level for this ogre but there's definately something weird about what it is initially. I will try out different types of mobs.

I did the same test only versus level 70 and 71 boars on Blades Edge Mountain. Same conditions apply and I made sure I didnt hit the shield.

70 nosunder = 828
70 sunder5 = 985

71 nosunder = 814
71 sunder5 = 963

70:
828(armor-980) / (.45[2853] - 828) = 985(armor-3580) / (.45[2853] - 985)
armor = 6772

71 same method:
814(armor-980) / (.45[2853] - 814) = 963(armor-3580) / (.45[2853] - 963)
armor = 7130

Now we can re-plug these back in to find Z for both levels. And we can check it for all 4 levels of data.
Z = BT(armor-ArP-Sunder) / (.45[AP] - BT)
70:
Z = 828(6772-980) / (.45[2853] - 828) = 10520
Z = 985(6772-3580) / (.45[2853] - 985) = 10520
71:
Z = 814(7130-980) / (.45[2853] - 814) = 10655
Z = 963(7130-3580) / (.45[2853] - 963) = 10655

Well It seems here that the numbers here are closer to the 70 level for both 70 and 71, quite a co-incidence. Could it be that Either A) blizzard didnt update the armor co-efficient with same type mob level variance, of B) this was intended.

It is conclusive though that the Coefficient did change with mob level and mob type in sync with the suggested forumla: Z = 467.5 * [TarLvL] - 22167.5

 

Gruntle

If I'm calculating things correctly I think you should have 16 expertise skill (including 5 from orc racial) + 2% antidodge, so 6% dodge removal in total?

Total swings from white+BT+HS+WW+Exec: 5166 (including trash)
Dodges for the same: 7

Total swings on trash: 2523
Dodges: 0

I.e. 2643 swings with 7 dodges (and assuming 6% antidodge)=>
The dodge chance is 95% likely to be within [6.13,6.54], mean is 6.26%.

We need 10000+ swings running with 6.25% to pinpoint the dodge chance to within ~0.1%.

 

Cavein

WWS for this week.

I had the same amount of expertise as last week on all bosses this time including Mother (Onslaught Belt, Shard of Contempt, 17 expertise plus -2% dodge from talents).

I was asked to tank on several packs in Black Temple where my expertise dropped to 5 plus -2% dodge from talents so best just to take the boss fight numbers.

As a point of interest, the character screen reports my expertise as a round figure of 17 (4.25% less dodge/parry). However, I have 70 expertise rating which would be 70 / 15.76 = 4.44% (6.44% less chance to dodge with talents).

I would expect my overall misses to dodge to be higher assuming the cap is 6.5% if the character screen expertise is used.

Any thoughts? Perhaps there just aren't enough hits to draw any conclusions.

 

Gruntle

You had the same amount of expertise this week so we can add the swings and dodges.

This week (including white, cleave, slam, MS, WW, Exec, HS, VR, hamstring)
all swings: 8572
all dodges: 6

trash swings: 5777
trash dodges: 5

Both weeks
Adding last week and this we get 1775+2795=4570 swings and 3 dodges =>

6.32% dodge with 95% confidence to be within [6.27%,6.44%]

Seems very likely that dodge chance is higher than 6% at least, but maybe someone else should look through the numbers too. Mistakes are easy to make.

 

Cavein

WWS - Dodge Events for Actor Cavein

I can verify that I was in tanking gear for the dodges against the Bonechewer mobs.

Next reset I will ensure that I retain the same amount of expertise in tanking gear so that the whole log can be used.

 

Mattam

I've searched the forums but haven't come up with an answer yet. Does using two weapons of the same speed still generate an extra charge of flurry?

 

• landsoul

Technically no, but what happens is as long as the two weapons are in sync it is impossible to expend the flurry before the 4th hit.

You don't get an extra charge of flurry, rather are able to swing four times before the flurry expires.

 

• Apate

Some info here http://elitistjerks.com/f40/t12720-m...s_flurry_work/

It's "outdated" but I would be surprised if flurry was a clean and tidy buff currently. Latency should still cause some issues.

See you, auntie.
"You don't need a machine to make a rainbow. For rainbows are made of happy thoughts, and dreams, and chocolate unicorns, and gumdrops, and licorice sunsets, and fuzzy gumdrop bears, and sugar-coated chocolate gumdrop land."

Originally Posted by DeeNogger I am coming for you Apate.

 

• landsoul

Now if we had some 74 mobs floating around, we could nail this problem right on the head with data from that. Unfortunately we don't. Does anyone have any data on this from pre-BC like in BWL or Naxx, etc? Maybe we can match it with that.

Like hit, it might make a non-linear step at +3 levels.
Say for instance, it might jump up from .2 per level to .4 per level making it 6.2% or .5 per level to 6.5%. I am absolutely positive its a per level related calc and not just a set number, it has to be either one or the other.

We need someone with -6.0% to dodge, with -6.25% to dodge, and -6.5% to dodge to send reports.

 

dr_AllCOM3

With different weapons you don't exactly get 3 "hasted swings". It's somewhat between 3 and 4, depending on stuff I haven't tried to figure out yet.

 

Voxx

I think it's safe to say now that 5.6% is not the dodge rate of a raid boss, in which case I'm going to have to change the info on the first post. However, I'm just going to leave it as it is until we get some definitive and undeniable answer as to what the doge rate really is.

 

• landsoul

A few comments on the Stat rankings:

Ap is almost always better until you reach a certain top-end point of AP with low crit. Crit is not always better than haste if you have top-end crit/AP levels. Penetration is better than everything except expertise when you have a decent amount, especially on lower armor bosses. You are right on though that expertise and hit are best until they are capped.

Comments on 17/44 rotation:

When starting, its always best to get battle shout and rampage up first (let the tank build threat) and then do BT / WW
After that, always BT then WW then rampage. If you cant refresh rampage in time, make sure to hit BT and WW first and use rampage on GCD. Do not hit rampage over BT or WW. When you need to refresh rampage soon, bank 75 rage while HS'ing so you do not miss the opportunity. While Rampage is not needed to be refreshed, bank 55.

Comments on ratings:
Where are you getting 22.1 rating per crit% from?
15.76 rating per % for both haste and hit.

Surviving:
There's no fight that I can think of that using commanding shout will save your life. If you use it while you are at half health, you are only gaining 50% of the buff. And there's no ability in the game that can one shot that you would have lived if you would have had CS up, except for maybe pulling aggro on a very hard hitting boss.

Also, you take 22.22% more base damage in zerker stance than you do defensive.

 

Vorkannis

I believe the point Voxx was trying to make was that it is not an absolute life saver with regards to being one shotted, but it can be used effectively at times. While running out of a flame patch/void zone/death and decay, during ground slam/shatter and moving out of a rain of fire come to mind. If you know you will be hit by expected and/or burst damage, surely it is wise to help negate this incase of further unexpected damage?

 

Voxx

This is an often-debated topic. The actual values of each stat almost ENTIRELY depends upon your current gear. I'm currently debating removing that section from the guide because of how many times someone comments on how they think the priority is wrong. Rather just state that each stat has it's place and it's time and that you should refer to the spreadsheets in order to make your own comparisons.

If you really feel the "rotation" I detailed is inadequate, I can fix that. I personally feel that just following the general guidelines in the "rotation" is enough. As to what abilities to use in the first few seconds of the fight, I would have thought no one in their right mind would start "nuking" immediately without already having Battle Shout and Rampage up? Perhaps I have too much faith in the dps warrior community?

I could use 22.08 rating per % if you want... but I figured 22.1 was close enough. If you have a source that states otherwise post it here and I'll take a look?

It's not so much being one-shotted in a single attack that Commanding can save you from, but for instance when you take a burst, or know you are going to take a burst in a short time and are otherwise at a high % of health, Commanding can increase your chances of surviving subsequent bursts. Ground Slam + Cave In comes to mind immediately, Commanding right before kiting phase of Supremus can help you survive geysers. There's a lot of cases where it won't directly increase your chances of surviving a one-hit kill but it WILL increase your chances of survival on the same basis that tanks stack stamina. It's not the one hit that will kill you but it gives your healers more time to react to multiple hits.

 

Moogul

Now you say that, but is it really that clear cut?

It's something I've been wondering about for a while - what should I spent my first 30 rage on? BT or Rampage? Battle shout seems fairly obvious as it's a higher amount of AP for a cheaper cost, but should I rampage and put BT back a global cooldown (or more, if it takes me longer to get rage. If I'm wielding 2 2.6 speed weapons, it's entirely possible to get a miss/dodge and maybe have to wait even longer for a bloodthirst), or should I just BT as soon as possible?

Threat is rarely an issue, on a large proportion of fights the tank goes in first anyway and/or gets a misdirect, so it's very rare that I have to say 'better not bloodthirst yet, I might pull aggro!'

I've always been of the general opinion that delaying a BT off cooldown is a bad idea, and most of the time I play like that for pulls too - putting BT on cooldown is my 2nd highest priority after battle shout (which is normally bloodraged up before the pull anyway).

I've not done the maths for this, so maybe I'm wrong. Is it really worth putting up rampage before BT?

 

Voxx

Well you delay the Bloodthirst for 1.5 seconds, possible more I suppose. However you gain 50 attack power off the bat, and more per swing. We'll say that delaying Bloodthirst cuts the value of the first Bloodthirst in half (3 seconds delay in getting it on cooldown is half the cooldown of Bloodthirst). Let's assume Bloodthirst hits for 1000 base without Rampage and that in 3 seconds you can get Rampage up to 3 stacks. (The initial stack, and 2 subsequent hits) This makes using Bloodthirst first worth 1000 damage. While using Rampage first makes your Bloodthirst worth 556.25 damage. This doesn't include the minor AP bonus from auto attacks with Rampage up for 3 seconds.

So it seems that Bloodthirst first is the best option via napkin maths. Certainly if your first few swings are non-crit and you don't have the option to use Rampage before Bloodthirst, then Bloodthirst away. Even without the crit-starvation initially it seems Bloodthirst is generally the best way to go.

 

• landsoul

Yes, you et more DPS out of putting BT and WW on CD right away at the start of the fight, but sometimes that can lead to chance of aggro gain and subsequent tank-pissoffery. The first few seconds of letting the tank build his ceiling is an optimal time to build a rage bank and get rampage up before you start your BT/WW cycle.

I'm not complaining about the whole commanding shout thing, just saying that its really not useful lets say versus proper movement and understanding fight mechanics. Only thing I could think of would be Fatal attraction and maybe taking the poison on council. You know when its coming, and are at full health, if you get the poison on you, immediately command shout and D-stance will give you an extra 10% or so health. Even then, you are still no where near dying unless you stand in a blizzard, flamestrike, consecrate, or get holy wrathed by the priest without a healthstone. In fact the chance is so small of that happenning if you are paying attention correctly, that wasting rage and 2GCD to switch shouts would end up being just a drop in DPS.

Yes could you please use 22.08%, 2 decimals is consistent with everything else. My bad on the 15.76, it is 15.77 after all, I was thinking about something else, apologies.

 

Grymm

On a lot of bosses, my sequence starts like this:

1. Battleshout
2. Demo shout a progression boss (we don't demo most farm bosses)
3. Sunder
4. Rampage
5. DPS cycle

This delays my first BT by probably 6 seconds (a full BT) on most bosses. What's important, though, is on bosses like Kalecgos and Brutallus, I'm often hitting the boss before the tank has shield slammed. On Brutallus, I'm attacking as he runs over me to the tank. I probably get a full cycle on the boss before the tank is even in melee range. So, did I delay my first BT 6 seconds? Or did I find 3 more seconds of DPS time? My DPS goes down, but my damage goes up. Also, each sunder on a boss is 3-4% more physical damage. A tank running SS/Rev/Devax2 will have:
Sunder 1: 3.0 seconds
Sunder 2: 4.5 seconds
Sunder 3: 9.0 seconds
Sunder 4: 10.5 Seconds
Sunder 5: 15.0 Seconds

By sundering yourself, you can have a 5 stack of sunder up by the 6 second mark. That's a gain of 16.5 sunder*seconds or however you want to quantify it. I am most prone to pulling aggro in the first 15 seconds when all the threat a tank might have is a MD, a PoM, and a bunch of misses/parries. I like to see a nice 10k aggro cushion before I start plowing into a boss. I'm sure this approach might hurt my pretty numbers some on certain bosses, but it's been a highly effective approach for me.