Shha

As for the leather gear sets. I made same thing you did Landsoul :

- Surv Hunter/Feral druid/totem twisting/ret paladin/ms warrior is very typical for my raid. I dont stack bm hunters in my group so no for that, but anyway.

ArP goes up much in SEP and plate items pull ahead.
Even without all those buffs, ArP will be ahead for 6200 armor bosses, and leather will be ONLY better with low raid buffs, and on high armor bosses. I dont see it as a huge problem.

 

• landsoul

Good point, I made the leather calcs on a 7700 boss with mostly lighter buffs. Can you post some numbers with your calcs?

if you do not have 2 weapon mastery, then you should enter that into the talents area. SEP for expertise should go to zero after you surpass the 5.6% cap.

 

• landsoul

New version available

 

Bersi

@landsoul, Talon of Azshara misses the 15agi in offhand

 

Dralmoo

No screenshot of numbers but I looted [Rod of the Sun King], in testing vs blasted lands mobs it had a similar proc rate. It was definitely low enough to make the proc useless and I banked it immediately. I used Recount to track the number of times I gained rage from the effect over 10 minutes.

 

Barmbul

Here is a wws of a full night of attemps on arch with Rod of the Sun King
Wow Web Stats

57 procs in 4 hr and 20 min

Armory profile

 

Shha

Also as far as I know haste procs dont change the proc % chance to occur.

PPM is constant. Based on that you discover "proc chance" , assuming unmodified weapon speed. So basically the proc chance is given by :

Proc Chance (PC) = Base Weapon SpeedxPPM/60

Haste increases the proc chances by giving you more hits. This is i believe the reason they put hidden cooldowns on items. They wanted to make items which would proc decently in normal environment and not become retardedly good in raids (like hand of justice - pretty bad trinket if you farmed with it, amazing in raids and sometimes pvp).

 

Machinator

I thought testing showed that the proc % of a swing changed based on current speed to keep the same PPM. Ill have to search for it again. I might test it myself if I can on the PTR.

"Information is ammunition."

 

Shha

Well it might be the case with flurry then. Its definitely not "global" truth though, as lets say dragonstrike mace is 1ppm (16% uptime) and thats what shows on white swing auto attack test. In raids its 35-40% uptime (which is well within what this spreadsheet shows) , and you cant explain it just by windfury/instants (2-2.5 times higher uptime for proc without internal cd, means probably around 3 times higher proc rate).

 

Grayson Carlyle

Weapons with procs have a straight up % chance to proc on hit. Enchants, and non-weapon procs are PPM and more recent testing suggests they adjust with your current weapon speed.

 

• landsoul

Breaking down Barmbul's Archi kill:

Power of the Sun King (Rage) 11

...procs out of 159 possible off hand counted attacks on your kill..

To find the number of off hand attacks I took your total heroic strike crits, hits, and total white crits glances and hits and added them up, and then divided them by 2 because you have matching speed weapons, then I took the number of whirlwind hits and divided them by 2.

So you had a 6.9% proc rate from off-hand... if your flurried offhand speed is 2.16, then you would get one proc every 31 seconds....

On total time to kill, 407 seconds, you gained 11 procs. yielding one proc every 37 seconds...

So you would get a proc every 31-37 seconds....

Looks to me like it could be somewhere between 1.5 or 2 PPM considering maybe having to move a little bit and/or flurry/haste. This is pure speculation though. Seems very weak. Still I would like to know more.

Next version will have the talon fix and a few other fixes in it.

Also, a arms/fury sheet is in the works but I have a few questions and speculations before I can really start cranking out this build:

Questions
1) Do people use Swing -> Slam -> MS, Swing -> Slam -> WW, Swing -> Slam.... or do they use Swing -> MS -> WW, Swing -> Slam, or in other words do people weave instants in between swings and slams, or do they weave slams in between swings and instants.
2) Does the difference in skill rotation preference greatly affect DPS? and by how so? Which yields the highest? Do these users have plenty of rage to close the pure swing -> slam rotation?
3) Based on varying base weapon speed and or passive haste, what are the different time tables people have found they used?

Speculations on pure swing -> Slam
1) White DPS will be lower than Slam DPS
2) MS and WW DPS will not be at full potential due to grossly overlapping cooldowns
3) Based on this and how timetables work out, pure Swing -> Slam may not be best for certain setups??
4) Based on base speed of weapon, pure Swing -> Slam may not outperform Swing -> Instant
5) Ideal rotation choice can be based on base speed of weapon

I own Torch of the Damnned and S3 Mace, but I dont have the time or energy to spec in and out of certain things outside of arena spec or 17/44, and I dont own any other passive haste gear to test any of this. Any help on this issue would be greatly appreciated. Thanks for all your support and suggestions so far!

 

WernerVonBraun

1) Swing, Slam, GCD (MS>WW), repeat.
2) See 1), it's the best way to work with Slam.
3) What? GCD+GCD+Lag < Slam+Swing.
4) Only in theorie, but they will be very close. Slam being a few percent under white due to movement and no rage.
5) No, since Slam is superior.
6) All good 2h weapons are slow enough, except during Bloodlust+Hasteproc/-pot.
7) You get Swing+Slam+Instant and pay 0,5s.
8) Should be answered by now.

Glance does 75% damage. Correct me if I'm wrong, you used 65%.

 

Shha

To put my 2c

1) I agree in general. Except i use WW>MS rotation for following reasons:
a) WW is more damage/rage and rage IS a limiting factor for 2h.
b) In case of rage starvation skipping one GCD lets you reuse WW.
Its a lot based on user preference, and I dont want to make it sound final.

2) Ill talk about rage later.

4) Slam will be more significantly lower. White damage = Swings (in theory equal number to slams), Wf Procs (20% more swings), SS procs - every 4 swing 4 Slam 2MS 1 WW and 0.8WF cycle you get 11.8x5% = ~0.6 white attack due to SS. So every 4 slams you get around 5.4 white swings because of Sword spec and WF. Additionally not always will you be able to use slams due to rage/movement. Usual white: slam ratio is more in 3:2 realm , approaching 4:3 in more perfect conditions (low armor boss, lots of raid buffs).

7) In limitless rage and hasted environment it COULD be true, but in reality it just doesnt happen.
8) Not really, its about the same.


Now the important part: 2h is WAY WAY more complicated to model then dual wield. Reasons:

a) Rage modelling. Rage is VERY streaky and it absolutely controls your dps.

a1) Typical dual wield rotation od 3 BT 2 WW 1RMP is 18 sec long and includes 90rage+50rage+20rage. Thats 160 rage/18 sec or 8.8 rage per second needed. Any extra is dumped into heroics which dont contribute AS much to your dps.

a2) 2h Rotation for a 2.7 flurried/hasted weapon lasts (2.7+0.5+lag)x4 = 12.8+4xlatency (lets assume 100 ms) = 13.2 long. Rage use is 60(MS)+25(WW)+60(Slam) = 145rage/13.2 sec. 11 rage per second with good accuracy.

So the rotation for 2h requires 25% more rage. The generation is lowered by around 0.6/3.3 = 18% due to weapon not swinging all the time. Generation is lowered by lower 2h generation in general. Due to this facts, its IMPOSSIBLE to have the rotation going constantly. You get 2 glancings in a row + no SS/crit/WF, and you are rage starved. Many spreadsheets dont calculate it. They just put average rage/sec, which IS enough to sustain rotation. But streakiness of 2h dps makes it much harder.

Thats why usually stacking crit is best for 2h despite what spreadsheets say. You want your SS chance+WF chance+Crit to be high enough. To keep rotation going a rule of thumb is, one of those needs to happen every 2 slams. So every 2 white swings, you NEED to either white crit, or WF attack, or Sword spec , or your rage will be depleted. That makes a good 2h model harder then execute modelling. It also shows why Sword spec is so superior. Given how your white swing is followed by instant/slam , you have 3 chances of sword spec occuring per white swing vs lets say 5% chance to crit for axe. Sword spec gives almost 15% rage generation increase apart from normal damage results. Axe spec gives around 3-4%. Most spreadsheets fail to model it.

b) Rotation is very lag dependant. Thats why I gave up on 2h, even though it appeals much more to me. Playing from europe on US server just doesnt work with 2h, period.

 

• landsoul

Yeah Shha, thats what I figured. I was afraid there were a lot of variables going on. I thought 2h could do just as much damage as DW, due to less headache of miss or mitigation streaks keeping up rotation, but its being said that 2H has even more problems when keeping up rotation. Movement and not enough rage generation would really break any accuracy of a sheet I could make.

I made a little table of rage per crit, glance, hit, windfury, etc with my buddy's Arms gear and buffs

glance = 13.81
hit = 19.36
WF glance = 14.90
WF hit = 21.04
WF crit = 43.14
crit = 39.67

Rage per glance
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[%DamApplied]*[Range]*[%Glance Applied]+3.5

Rage per Hit
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[%DamApplied]*[Range]+3.5

Rage per WF Glance
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[IfWindfury]*[%DamApplied]*([Range]+[BaseSpeed]*445*(1+0.3*[IfImprovedWF])/14)*[%Glance Applied]+3.5

Rage per WF Hit
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[IfWindfury]*[%DamApplied]*([Range]+[BaseSpeed]*445*(1+0.3*[IfImprovedWF])/14)+3.5

Rage per WF Crit
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[IfWindfury]*[%DamApplied]*([Range]+[BaseSpeed]*445*(1+0.3*[IfImprovedWF])/14)*(1+[CritBonus])+7

Rage per Crit
=0.01365*(1+0.01*[2hSpec])*(1+0.03*[FerociousInsp])*(1+0.02*[ImpSanc])*(1+0.04*[BloodFrenzy])*[%DamApplied]*[Range]*(1+[CritBonus])+7

Now keep in mind, each 3 possible windfury chances can come off the 3 possible hit chances, yielding 27 different combinations. One could list all those and figure out the rage per combination to determine if they can use slam and an instant, slam and rage for partial instant, slam only, or nothing at all. Then possibly compile the chance of those effects occuring, and make some kind of a decimal multiplier that is factored directly into the skill's damage. THink it would work??

 

• landsoul

On glancing blows:
Quote my Guide:

Glancing Blows

Each mob that is the same level or higher, has a chance to glance your attacks. The severity of the glance as well as the frequency goes up as the mobs' defense level increases beyond your attack skill. There is no item stat that can modify this mechanic, unfortunately.

Below is a table that demonstrates how exactly glancing blows tone down your DPS. When a glancing blow is rolled on the hit table, a random number is generated between a high and low value depending on the defense of the mob you are hitting. This factor is then multiplied to your normal damage to give you the final applied number. The following values are for melee classes only because casters have different glancing values.
Lvl Min Max Chance
70 .91 .99 10%
71 .91 .99 15%
72 .80 .90 20%
73 .55 .75 25%

For example, if you were attacking a 73 mob, 25% of your rolled white damage attacks will only deal 55% to 75% of its normal damage. These values come from blizzard's step formula for differing levels and classes. For our calculations, we will use the average of these factors because our calculations will span infinite time limits. This function we will call [glanceF]. For 70 and 71 mobs is .95, for 72 is .85, and for 73 is .65. For calculations we want to know the amount reduced by glancing blows, or [glanceR].

[glanceR] = 1 - [glanceF]
[glanceF] = 1 - [glanceR]

Numerically, glancing is a reduction that is always present on white attacks. Multiplying the chance to glance by the effect of the glance will allow us to use it in our hit table factor equation.

 

Shha

Yea it could probably work, but your rage figures are really confusing. Lets take hit and glance

hit = dmgpart+3.5, glance = dmgpartxglancemodifier+3.5 in short

From those calculations it seems you count glances at 0.65 damage of normal hit? Thats wrong it seems. I know that was the formula for pre 2.1 glancings, but it really doesnt seem to hold now. Take a quick glance at wws. Teron fights are easiest to spot. Glancing average for rogues vs average Hit values.

Using Vis Maior top Teron:

Rensy - avg hit 660, glancing 493. Glancing/hit =75%
Athaena 675 467 Glancing/Hit = 68%
Milu 681 495 = 73%
Enzy 680 444 = 65%

Then Abanana next one
Liet 684 544 = 79%
Mitski 742 531 = 71%.
Warriors (only MS warriors are important because of lack of HS)
Alabamaman 1777 1295 = 73%

Im not sure about 75% although its definitely closer to truth then 65%. Id say its safe to assume glancings do over 70% of damage, with POSSIBILITY they do 65%. Something to consider in your sheet for fury as well.

Edit to your post landsoul. The info you took about glancings was not updated since pre-TBC. The only updated part in wiki mechanics of glancings seems to be the chance. No changes were made to severity. Also glancings seem to be now more of a constant rather then a range.

 

Joeblack

I think there may be a glitch in using the drop-down menus for trinkets.

I was just playing around selecting different Trinkets and watching how the overall dps was adjusting and I noticed that identical setups were returning different overall DPS values.

Both times I saw it happen it was after I selected 2 of the same trinket and switched back to a configuration that was actually possible. I did it a couple times and watched the overall DPS fall about 10 each time.

Also, it appears that improved battle shout is being counted twice when 5/5 commanding presence is also selected. I mean... "full improved" battle shout & "commanding presence: 5" are the same thing aren't they? It's counting them double if you have both enabled (which is the default).

 

• landsoul

This info is updated after TBC afaik. They lessened the chance to glance by 15% from 40 to 25, and nowhere did they document the change in severity of the glance thereof.

I get avg hit/glance for one WWS Wow Web Stats
399/585 = .682 teron gorefiend
380/508 = .748 supremus
435/786 = .553 Akama
612/862 = .710 requiliary
432/581 = .744 Gurtogg

which is within the limits I had posted.

Rogues and sword warriors will be a higher percent because they get more main hand attacks than offhand, thusly throwing off the percentage towards main hand. Also with windfury, main hand will be more prominent which is valuing glances in the offhand as a lower percent. Also, fury warriors' WWS replaces the white hit with heroic strike, which also screws up the numbers. Also, each player is not always throwing out constant damage range in the fight, which can also skew numbers in favor of hitting harder on glances when those numbers are active versus not glancing when they are not... Glancing mechanics cat be accurately shown in a raid environment unless controlled.

Try just auto attacking on hydross, heck you could even go into MC or BWL to do this, without any +ap procs or windfury or any extra attacks or anything, and check the WWS there. A fury warrior, non-sword spec rogue or arms warrior will only work.

As it is currently working, most trinkets are only counted once no matter if you have two of the same, and when switching any item in a drop down menu, you must refresh the selection 2-3 times to converge on the correct value using (CTRL SHIFT Y) or selecting the same item again. This bug is only happening on drop down menus and I am trying to figure out why, but the numbers are still right.

 

Shha

Actually thats wrong. I thought kinda the same about wf and such throwing off the numbers but in reality they arent.

You are looking for average white hit damage. What does it matter if they get more MH or OH or whatever kind of attacks. ANY white attack whether procced by wf oh or mh can glance. So 100% of the white damage is susceptible to glances. Statistically it will be all the same. Just think about it. Sure raiding data is affected by RANDOM things, and thus we just need larger sample. The fact that sword rogues get more MH attacks doesnt change the glancing % at all.

Your data has serious flaws. Akama is the ONLY one far from my 70% estimate, and its too short fight to draw conclusions. 40 hits total is a bit lon the low side. Anyway first thing with multiple sets of data like this would be to kick the most extreme results. we are left with .682 .744 and .710. Thats more then your expectations.

 

Shha

Oh and while we can discuss our glancing blows, one thing for sure - you counted the crits as having impale (which doesnt work on white attacks).

 

• landsoul

huh?

 

issei

I could be wrong, but I believe Shha is referring to how Impale applies its 20% crit damage bonus to "abilities" instead of all damage.

 

• landsoul

Oh I see what I did wrong, ill edit.

 

Joeblack

no comment?

edit: OK there is definitely something wrong, I just watched the same thing happen while rearranging gems.

 

Joeblack

Is there a +hit cap factored into the spreadsheet? Testing around I was able to see that +hit gives more overall dps output than +strength at least as far as 333 hit (didn't test any further because that sounds ridiculous)