I don't think UB will double dip from anything - it'll simply do 30% of whatever your DC hits for. If you have things that modify your DC damage, such as Dark Death, then your DC hits harder, and thus your UB ticks for more.

So I believe he was correct in his 4.5% damage increase to UB from DD glyph. It's not double dipping - it's based solely off of DC damage.

Bear in mind BCB appears to be doing 31.25% per disease on the PTR (testing being done in the Blood DPS thread), so I'm not sure BCB's position is going to change.

Right now, DKs on the ptr have so many issues that it's nigh-impossible to do any meaningful testing. First, any Unholy DK transfering to the test realm will be missing a talent point (which is stuck in UB, but the it's the OLD UB), and they will have a permanent old-style Desecration V talent. That's on top of the fact that Blood Strike is set at 50% per disease (which, according to the patch notes, is supposed to be 25% now), which skews any testing of rotations.

Hopefully it'll all be fixed in the next PTR build.

Looks like Unholy Blight went down from 30% to 20%. Unholy Single target wasn't out of line by any means, so I assume this nerf was based on Blood DKs abusing it in the 50/0/21 spec.

At the moment, Desolation does not appear to be working for me on the PTR. Makes testing a bit of a pain =T

This seems to be a very buggy build. People are claiming that shaman nature's swiftness and priest inner focus buffs are lasting forever and pulling over 10k DPS on target dummies. it's kind of amusing, but hopefully we can get a fixed patch to do some real testing soon.

I can confirm that Dirge and Glyph of Scourge Strike don't work either.

Wow seriously? Nothing works, I was hoping to hop on soon and get some testing in!

An easier way to get the same result:
Denote by E[T] the expected number of SS in your cycle.
Now, consider the time you start your cycle.

If your first proc is on your first SS, then you're back "as if" at the beginning of the cycle, and your expected cycle duration after the cycle will still be E[T]. This means that the full cycle duration will be 1+E[T], with probability 0.25.
Similarly, if your first proc is on your second SS, you get a duration of 2+E[T], with probability 0.75*0.25.
Same goes for the first proc on the third or fourth SS...
If you don't proc at all, then your cycle duration is 4, and that happens with probability 0.75^4.

This leads to the equation:


Solving it (with variable E[T]) leads to the result E[T] = 8.641975309, unless I'm mistaken.

I don't know where the small difference comes from.

So, if I remember the numbers correctly from when this was first examined, UP w/ 2/2 IUP was still about 2-2.5% worse than running BP. Will the pet changes shift this at all? Ghouls, AotD and Gargoyles should be matching us at the hit/exp caps, pushing their contributions up and making higher gains from UP haste. Last I checked on the PTR this functionality wasn't working to do any testing with it.

Hi there, i recently played with some numbers to see if obliterate can be an interesting way of dps for unholy.
It seems that SS is > to OB in 3.1, main reasons are that SS proc desecration and an OB build cannot be a derivation of the 12/0/59 spec witch is a little > to 0/10/61 dps wise.
But with the nerf to SS in 3.2, will it still be the case ?

I believed that it was already calculated somewhere in this forum, but didn't find any precise information about it, so i did it myself.

Be careful as this post will be quite long, with a lot of data to check some game mechanics before doing the actual calculations.


1) Objectives

We want to compare an OB UH build to a SS UH build.
This comparison must be done for both 3.1 and 3.2 game version, with or without the 4pt8 bonus.
The 2 builds i will compare are : this one as OB build and this one as SS build


2) Preliminary comparison

We have to check if the 2 specs can be a base for comparison. An OB spec which would lose a straight 10% dps via talents is obviously useless to test, we don't want to loose our time.
The SS build is a standard one, so we just pick up the differences between the SS and the OB builds.

With the OB build we loose :
- Outbreak, which become an useless talent anyway if we don't use SS, we loose 30% damage on PS, a 0.5-0.7% dps decrease
- 2 points in Necrosis, which is a small 1-1.5% dps decrease in Unholy
- Scourge Strike, obviously
- the UB glyph in 3.2, currently a 1% dps decrease, or the SS glyph in 3.1, an unknown dps decrease, but we'll considere it to be a 1-2% dps decrease
- in 3.1, we loose 8-9s uptime of desecration, if not more in case of an encounter involving a lot of moves, so a 2% to 3% dps decrease.

But we gain :
- 3% crit on OB/BS/PS, which should represents a <1% dps increase
- Subversion, a +9% crit on OB and BS. Compared to Vicious strikes, it's a +3% crit for OB and +9% for BS, so if OB does the same damage as SS, we still are in the <1% dps increase range, but we have for sure a -25% threat

Conclusion : in 3.1, the OB spec must make up for a 4% to 5% dps decrease, which should explains why nobody use it. But in 3.2, the 2 specs seems quite comparable. The OB build just have to make up for less than a 2% dps decrease.


3) Checking in-game mechanics

In 3.1 :
- SS is 0.45*wpd + 357.2 with 11% more damage per disease
- OB is 0.80*wpd + 467.2 with 12.5% more damage per disease

In 3.2 :
- SS is 0.40*wpd + 317.5 with 10% more damage per disease
- OB is 0.80*wpd + 467.2 with 12.5% more damage per disease

We have to check if in-game numbers perfectly match the theory.

I will check the numbers on PTR, and considere that if we know how to calculate on it, there is no reason that the mechanics changed from live (for example the OB glyph is believed to behave the same way).
Setup on PTR : 3684 AP with a Large Club doing 4-6 damage (5 average damage or 1.6dps), 17.54% arp, the 4pt8 bonus, in UP, on a lvl 80 dummy target to avoid partial resists for SS, and on a lvl 83 fo OB
weapon damage = wpd = 3684 / 14 * 3.3 + 5 ~= 873,37
boss_armor = 10643 - 17.54/100 * (10643+15232.5/3) ~= 9130,15
adrm = armor damage reduction multiplier = 1 - 1 / (15232.5 / boss_armor + 1) ~= 0,62524

First of, while checking numbers in game, i found that the OB glyph behave in an additive way. It adds 20% damage to the 80%, so it's a 100/80 = 25% damage increase.
I will spare you the calculation and test to determine how it works exactly, so here are the results :
The OB glyph increase the base damage of OB without disease by 25%. The bonus doesn't apply any damage increase to the disease bonus.

no disease :
Theory :
SS = (0.4*wpd + 317.5) * 1.2{Outbreak} * 1.1{Black Ice} ~= 880,24
OB = (0.8*wpd + 467.2) * 1.25{glyph} * adrm ~= 911
In game : avg SS hit = 880 and avg OB hit = 911, it perfectly matches

ff and ep up :
Theory :
SS = (0.4*wpd + 317.5) * (1+0.1*1.2*2){2 diseases} * 1.2{Outbreak} * 1.1{Black Ice} * 1.13{Ebon Plaguebringer} ~= 1233,39
OB = ((0.8*wpd + 467.2) * (1.25{glyph} + 0.125*1.2*2{2 diseases}) * adrm ~= 1129,9
In game : avg SS hit = 1233 and avg OB hit = 1130, it perfectly matches again

ff, bb and ep up :
Theory :
SS = (0.4*wpd + 317.5) * (1+0.1*1.2*3){3 diseases} * 1.2{Outbreak} * 1.1{Rage of Rivendare} * 1.1{Black Ice} * 1.13{Ebon Plaguebringer} ~= 1488,03
OB = ((0.8*wpd + 467.2) * (1.25{glyph} + 0.125*1.2*3{3 diseases}) * 1.1{Rage of Rivendare} * adrm ~= 1363,2
In game : avg SS hit = 1488 and avg OB hit = 1363, it still perfectly matches


4) Calculating the threshold to have OB do more damage than SS

We will get rid of global multipliers that affect both SS and OB (Rage of Rivendare)
c = average crit % in raid
adrm = armor damage reduction multiplier (the threshold)
We have 3% more crit for OB from the Subversion talent compared to Vicious Strikes. We will factor it in
crit multiplier for SS = 1-c + c*2.3*1.03{meta} = 1 + c(2.3*1.03-1) = 1+c*1.369
crit multiplier for OB = 1-(c+0.03) + (c+0.03)*2*1.03{meta} = 1 + (c+0.03)(2*1.03-1) = 1,0318+c*1.06

In a 3.1 raid environment without 4pt8 :
SS = (0.45*wpd + 357.2) * 1.33{3 diseases} * 1.2{Outbreak} * 1.13{Ebon Plaguebringer} * 1.1{Black Ice} * 0.94{average partial resist damage decrease} * (1+c*1.369){crit mul}
OB = (0.8*wpd + 467.2) * (1.25{glyph} + 0.375{3 diseases}) * (1,0318+c*1.06){crit mul} * adrm
adrm = (0.45*wpd + 357.2) * 1.33 * 1.2 * 1.13 * 0.94 * (1+c*1.369) / (0.8*wpd + 467.2) * (1.25 + 0.375) * 1.1 * (1,0318+c*1.06)
adrm = (0,839159244*wpd + 666,105959904) / (1,3*wpd + 759,2) * (1+c*1.369)/(1,0318+c*1.06)

In a 3.1 raid environment with 4pt8 :
SS = (0.45*wpd + 357.2) * 1.396 * 1.2 * 1.13 * 1.1 * 0.94 * (1+c*1.369)
OB = (0.8*wpd + 467.2) * (1.25 + 0.45) * (1,0318+c*1.06) * adrm
adrm = (0,8808017328*wpd + 699,1608421248) / (1,36*wpd + 794,24) * (1+c*1.369)/(1,0318+c*1.06)

In a 3.2 raid environment without 4pt8 :
SS = (0.4*wpd + 317.5) * 1.30 * 1.2 * 1.13 * 1.1 * 0.94 * (1+c*1.369)
OB = (0.8*wpd + 467.2) * (1.25 + 0.375) * (1,0318+c*1.06) * adrm
adrm = (0,72909408*wpd + 578,718426) / (1,3*wpd + 759,2) * (1+c*1.369)/(1,0318+c*1.06)

In a 3.2 raid environment with 4pt8 :
SS = (0.4*wpd + 317.5) * 1.36 * 1.2 * 1.13 * 1.1 * 0.94 * (1+c*1.369)
OB = (0.8*wpd + 467.2) * (1.25 + 0.45) * (1,0318+c*1.06) * adrm
adrm = (0,762744576*wpd + 605,4285072) / (1,36*wpd + 794,24) * (1+c*1.369)/(1,0318+c*1.06)

We now have the equations. adrm is the threshold above which OB does more damage than SS. It depends on the weapon damage, and the crit rating.
Because OB scales a lot better with the weapon damage, the more we do damage per hit, the better OB is
Because the SS crit rating multiplier is > to the OB one, the more crit we have, the worse OB is.

A simple damage multiplier value isn't really easy to make comparison with for gear, so we'll always convert it into an arp % value.
Formula :
arpp = armor penetration percent (from gear)
boss_armor = 10643 * 0.75{sunder & faerie fire} - arpp * (10643*0.75 + 15232.5)/3
adrm = 1 - 1 / (15232.5 / boss_armor + 1)
adrm = 1 - 1 / (15232.5 / (10643 * 0.75 - arpp * (10643*0.75 + 15232.5)/3) + 1)
arpp = - (15232.5 / (1 / (1-adrm) - 1) - 7982.25) / 7738,25

So now we just have to put some arbitrary weapon damage and crit% numbers.

Example :
Say we have an average of 7054 AP in raid with a 232,6dps 3.6 speed weapon (ulduar 25 standard gear, not bis)
Average wpd would be 7054/14*3.3 + 232,6*3.6 = 2500
An average crit% in raid with every buff is ~42%, but we'll take 39 because bosses have an hidden crit depression. So the base average crit for our SS is 39+6 = 45%
In 3.1 : arpp = 0,334219691 so OB will hit harder than SS if we have more than 33.42% armor penetration
In 3.1 4pt8 : threshold = 34,30%
In 3.2 : threshold = -6.82% so OB will always hit harder than SS
In 3.2 4pt8 : threshold = -6.82% same here, and we can see that with the 3.2 changes, SS and OB scale perfectly linearly independently of weapon damage and crit

But we said that OB hitting as hard as SS wasn't interesting, because we have to makeup for a 4-5% global DPS loss in 3.1, and a little less than 2% in 3.2

if SS is 20% of global damage, then to make up for a 4% dps loss, OB should hit 20% harder than SS. It only should hit 10% harder for a 2% dps loss.
We will simply calculate 2 new threshold : for 10% more damage and for 20% more damage
For this, we just have to take the previous adrm and multiply it by 1.1 or 1.2.
Results :
In 3.1 : thresh = 33.42% / thresh 10% = 57.66% / thresh 20% = 77.85%
In 3.1 4pt8 : thresh = 34,30% / thresh 10% = 58.46% / thresh 20% = 78.59%
In 3.2 : thresh = -6.82% / thresh 10% = 21.07% / thresh 20% = 44.31%
In 3.2 4pt8 : thresh = -6.82% / thresh 10% = 21.07% / thresh 20% = 44.31%


5) Conclusion

So the final results are quite interesting.
In 3.1, because you loose uptime on desecration, to makeup for an at least 4% dps decrease, you'll need nearly 80% average armor penetration. So the spec is not viable at all.
But in 3.2, if things stay the way they are now on PTR, you'll only need 20% armor penetration to have an OB build becoming better than a SS build. With 44% arp, you'll get a 2% global dps increase, if not more because BS will have a much bigger proportion in our total damage and it scales very well with arp.
Also, don't forget that OB scales better than SS with pure str gear, so we don't have to give up on this stat at all.

Final words : We know that OB scales better than SS with gear upgrades, and in 3.2, having a 2800 average weapon damage will be quite possible, so, the OB spec will definitly be the spec to go for unholy 2H if SS and OB stay the way they are.


Bonus : A table which shows the arp% necessary to have OB do 20% more damage than SS in 3.2 with various crit% and weapon damage (in raid) :

crit \ wpd	2200	2300	2400	2500	2600	2700	2800	2900	3000	3100	3200
40	44,14%	43,52%	42,93%	42,38%	41,87%	41,38%	40,92%	40,49%	40,07%	39,68%	39,31%
41	44,54%	43,91%	43,33%	42,78%	42,27%	41,78%	41,32%	40,89%	40,48%	40,09%	39,72%
42	44,93%	44,30%	43,72%	43,18%	42,66%	42,18%	41,72%	41,28%	40,87%	40,48%	40,11%
43	45,31%	44,69%	44,11%	43,56%	43,05%	42,56%	42,11%	41,67%	41,26%	40,87%	40,50%
44	45,69%	45,07%	44,49%	43,94%	43,43%	42,94%	42,49%	42,05%	41,65%	41,26%	40,89%
45	46,06%	45,44%	44,86%	44,31%	43,80%	43,32%	42,86%	42,43%	42,02%	41,63%	41,27%
46	46,42%	45,80%	45,22%	44,68%	44,17%	43,69%	43,23%	42,80%	42,39%	42,00%	41,64%
47	46,78%	46,16%	45,58%	45,04%	44,53%	44,05%	43,59%	43,16%	42,76%	42,37%	42,00%
48	47,13%	46,51%	45,94%	45,40%	44,88%	44,40%	43,95%	43,52%	43,11%	42,73%	42,36%
49	47,48%	46,86%	46,29%	45,74%	45,23%	44,75%	44,30%	43,87%	43,46%	43,08%	42,71%
50	47,82%	47,20%	46,63%	46,09%	45,58%	45,10%	44,65%	44,22%	43,81%	43,43%	43,06%
51	48,15%	47,54%	46,96%	46,42%	45,92%	45,44%	44,98%	44,56%	44,15%	43,77%	43,40%
52	48,48%	47,87%	47,30%	46,76%	46,25%	45,77%	45,32%	44,89%	44,49%	44,10%	43,74%
53	48,81%	48,20%	47,62%	47,08%	46,58%	46,10%	45,65%	45,22%	44,82%	44,43%	44,07%
54	49,13%	48,52%	47,94%	47,41%	46,90%	46,42%	45,97%	45,54%	45,14%	44,76%	44,39%
55	49,44%	48,83%	48,26%	47,72%	47,22%	46,74%	46,29%	45,86%	45,46%	45,08%	44,71%

You only increase the weight of armor pen with an OB build, the str stat is more or less the same (a little > in fact), and the crit one is a little less important. Because of the new value of armor pen, you can virtually have a complete bis set with a total value > to a SS build one.
If you want to be sure of it, we should constitute a new bis item list with an OB build weighstat to be determinated, and compare it's theorical dps output with a SS build. The fact is that 20% armor pen is really easy to get, because you only need 12.31 rating per %.
To be honest, i don't think it worth the hassle, because things can change a lot before 3.2 goes live.

If you take the horde bis T8 for unholy from this topic, you can see that we already have ~12% arp. If you swap the feet for the melancholy sabatons, and the Comet's Trail trinket for the Mjolnir Runestone, you will loose nearly no dps, simply because you trade stats with a benefice (relatively to the items ilvl). But in the process, you have 2 less items to loot from algalon 25, and you stand at an average of 28% armor pen.

Don't forget that i compared a 3/13/55 OB build to a 0/10/61 SS build. I could have compared it to a 3/10/58 one, because threat could become an issue in 3.2, and i believe lots of dk will prefere having a -25% threat bonus via talents just to be peaceful minded.
3/10/58 is ~1% less dps than 0/10/61, lowering the armor pen cap to less than 10% to have an OB build > to an SS build.

Also, the conclusion from my previous post clearly states that OB > SS only on the current PTR, and the new items have a lot more amor pen than the ulduar ones. In fact it is greatly possible that the future bis t9 for unholy will already have nearly 20% armor pen.