I went to Silvermoon again just now to test it again.

Setup was my armory gear, new IT sigil and [Widow's Fury]/[Silent Crusader] with FC as MH with [Hailstorm] with Razorice as OH, no glyphs, no pets used. Dummy was over 35% HP all the time and everything worked, test lasted around 5min - yes I know too short.

S/F was ~2550 DPS.
F/F was ~2500 DPS.

Crit and miss was comparable, obviously either it doesn't matter or better testing is needed (how raid buffs will affect results?).

Did some very quick testing last night with exact same gear as on live (Widow's Fury/Hatestrike) and a slightly different spec (0 points in desecration vs 5 points on live, 1 point in Crypt Fever vs 3 on live).

My personal DPS seemed to be a bit higher than on live, using death runes for 2 HBs on second half of the rotation. Gargoyle DPS is definitely down though. It is way easier to maintain though since it only eats 160rp to maintain the full 30 seconds. Sustained DPS should not suffer too much by filling the extra 30 seconds we lose with normal RP abilities as has been suggested on previous pages. If nothing else at least we won't have such mega-inflated Patchwerk parses.

Definitely going to have to test more. I want to see how IT on death runes compares, with the new sigil/glyph. I'm also interested to try out a 44/27 with FS/Unholy Presence.

Does anyone know yet if Unholy Presence increases KM's PPM like haste rating would?

I would have thought that Unholy Presence would have helped out 32/39 more than it would 44/27. The rotation for 44/27 is really short compared to the pre-patch rotation, so I doubt Unholy Pres would help much if at all. 44/27 seem to be liking IT HB BS HB Dump. Although, I would assume that after the first rotation they'd settle more into a HB IT BS HB rotation, so as to get more KM HB.

44/27 likes unholy presence because of the change to the frost strike glyph. With freezing fog and 75 rp generated per rotation, it's common that you'll either have an extra HB or FS to cast, which will put you over the GCD limit if you're in blood presence.

You're correct that unholy presence would favor 32/39 as well, depending on what the person does with their death runes. However, if you convert them into HB, it's rare that the player will need that extra GCD.

Yes, a post earlier showed that haste effects including UP increased the PPM of KM. Link

Edit: The post says that this data came from another post on the Blizzard forums, so take it with a grain of salt. I am however inclined to believe that these results are fairly accurate.

I use Bone Shield for my 3rd major glyph in hopes of increasing uptime (increased survivability and damage) during charge eating encounters and to help whenever I off-tank. It is also getting buffed to add 2 charges rather than one.

Since announcement of the death knight class and some of its major abilities, like dual wielding, i've always been interested in making an awesome death tank. Since death knights cannot use shields, they are limited to using only two handed face smashers and dualing one handers. Now, I've been using a two handed face smasher since I was a wee little death knight, at level 55. Obviously using a two handed weapon has its advantages, but what can dualing two tank weapons do for me? I'm curious to see how well a death knight can tank an instance or a raid boss using two [Slayer of the Lifeless].

To my understanding, dual wield tanking is not the greatest of ideas, because of the parry haste that is given to the bosses. But, despite this major disadvantage, there must be something that can be done to make a dual wield tank viable, isn't there? Stacking higher amounts of hit and expertise is my theory, but at being only a 73 death knight, there isn't much I can do to test this.

So, I turn to you, ElitistJerks, in my quest to become a good dual wield death tank. I am not asking for handouts, nor am I asking for you to hold my hand as I try and solve this puzzle on my own, rather a gentle push in the right direction, the same way a mother bird does to her hatchlings upon their readiness to fly.

So with this recent announcement by GC

"Killing Machine is a "ppm" now or a proc per minute. That means it will proc 5 times a minute with 5 ranks, depending on your weapon speed. We generally use ppm when we do not want players to gravitate towards faster weapons with the sole purpose of improving proc rates.

Killing Machine now is based on autoattack hits, not crits. Overall the random element will decrease, but it won't greatly benefit faster weapons. It will still benefit dual wield, just not as much. "

are we looking at a slow / slow being best now for Death Knights with the changes to BCB mechanics as well.

I can't exactly help answer your question, but I'll point you in the right direction.

This thread is mostly DW DPS discussion, but go here for the DW Tanking discussion.

Um, right. Well, people frequently overstate the parry hasting effect - it scales with the number of physical attacks you land per second on average, not with your weapon speed, so a DW death knight tank really gets no more haste effects than does a typical prot warrior, if we use slow tanking weapons - most of our attacks are spells, and can't be parried. A pair of [Broken Promise]s would probably be best.

I doubt a DW dk in tanking gear can put out an appropriate amount of single-target threat, compared to a 2h tanking build, but I haven't investigated that yet. Certainly our aoe-tanking is more effective than any but another frost tank.

It's a common misconception that BCB does more dps with slow weapons, it's exactly the other way - maybe it should be added to the think tank post to help this confusion.

Values of haste and hit should increase while value of crit should decrease.

Yes, but still not very. I think even haste rating is worth more than hit after the spell hit cap..

I've been going over some of the current data, and something doesn't make sense to me. If you take a look at Ren's two WWS, he get 9.17 and 10.1 PPM, effectively(I'm guessing WWS...nor any mod? can accurately track buff overwrite). Now, obviously, from the post from the EU forums, with some easily gotten haste, that 10.1 number can be easily topped.

My confusion is if you try to reverse-engineer the PPM Ren should have gotten. He's currently 2H holy, so I'm not exactly sure about his weapon setup. But, I'm guessing he didn't have 2 hailstorms, so 1.6/1.5 should be safe to assume. According to his earlier WWS, he got 28% crit rate on his swings, and his current gear has 12% haste.

1.55/1.37*60*2*0.28*0.5 = 19 PPM

And that's not considering bloodlust. So, my question: is my math wrong(this is CURRENT killing machine, not PTR), or does buff overwrite account for that much PPM loss? And if the latter, how come the KM on PTR isn't suffering from that?

The numbers from the PTR used 2.6 MH, 1.6 OH.

2.6*5/60 = 21.67% on connected swing
1.6*5/60 = 13.33% on connected swing

I have no idea what that person's hit/expertise was, so I can't extrapolate that to just on swing. Obviously on current KM, chance on swing is going to be crit/2, which should fall about in between the percentages above.

If it is buff overwrite, would haste and hit(after spell hit cap) actually be quite undesirable stats? If you take a look at that PTR post I've been referencing, he gets 9.82 PPM with no haste, but 13.25 PPM with 46.68% haste. If haste contribution was linear, it should have been 14.4 PPM.

So, obviously some buff overwriting occurring on the PTR as well, but doesn't seem nearly as large as the current model(if buff overwriting is the source). I'm having trouble understanding what else could cause that deficiency.

That sounds a whole lot more reasonable.

Fast weapons scale slightly better with raid buffs though.

I think the new weapons choice method will shift from:
Speed > DPS > Stats
to
DPS>Stats>>Speed

Which i personally think isn't such a bad thing to happen to DW. In your example, assuming you're hitcapped the Widow's Fury might just be better due to having better stats.

Of course it might turn out that fast or slow is indeed quite some better.

So:

I've seen some claims that Frost Strike scales worse with AP than Death Coil, but these claims weren't backed up with any math. Let's break it down:

Frost Strike - 60% weapon damage + 150 as Frost
Assume a Silent Crusader, average weapon damage (274-509)/2 = 391.5
Assume 40% melee crit, 40% spell crit.
(391.5)(.6) + 150 = ~385 (This is the base damage, unaffected by AP, but included for the sake of being thorough)
Talents affecting Frost Strike damage:

Glacier Rot (10% on diseased targets)
Black Ice (30%)
Guile of Gorefiend (245% crit damage)

So Frost Strike damage is:

(350 + AP*2.4*1/14)(1.1)(1.3)(.6 + .4*2.45)
So the AP contribution is:

AP*2.4*1/14*1.1*1.3*1.58 = AP * 0.387

Now, for Death Coil:

Death Coil - 443 base, AP coefficient 0.15
Talents affecting Death Coil:

Morbidity (15% damage)
Impurity (1.25*coefficient)

So Death Coil damage is:
(443 + AP*0.15*1.25)(1.15)(.6 + .4*2)
So the AP contribution is:

AP*0.15*1.25*1.15*1.4 = AP * .301

Excluding GoG, assuming 0% crit for both:

Frost Strike = AP*2.4*1/14*1.1*1.3 = AP * .245
Death Coil = AP*0.15*1.25*1.15 = AP * .215

Including Bone Armor for Death Coil, you get AP * .22

So if you stack the deck against Frost Strike, including no crit, Frost Strike still scales better with AP. If you consider the fact that raid buffed crit is extremely high (for me 40% or more) and the new Killing Machine will necessarily be used on Frost Strikes, it seems to me that Frost Strike simply crushes Death Coil in general.

If you combine this with GoG affecting HB crit damage and Unbreakable Armor coming off the GCD (easily paired with Blood Tap for a psuedo-trinket giving ~300 AP once a minute), it seems to me that 0/44/27 will be strictly better than deeper Unholy variants.

Thoughts?

Looks good Althir, I agree with your math. You might have to consider FS stealing KM procs from other abilities though. Also, you can't get FS and UA in a 44/27 build. At least, not with out sacrificing other, more important talents.

Althir your math looks good. But in this case I think you have to step back from looking purely at scaling and see the whole picture.
KM procs will partially be used for FS. Now this isn't terrible, but my IT and HB (iirc) do more damage, and thus benefit more from critting.

Also you can't conclude purely by FS > DC that a 44/27 spec is better. Factors like the new NotD will play a big and nearly incalculable role in 3.0.8.

gotta agree here - DC doesn't steal KM procs (which on a ppm system means you need to watch it even more post patch).

It would be hard to calculate whether or not you gained or lost anything. I would like to see a sim comparing the two based on the PTR (as right now it's pretty much a dead horse beaten with a tree live). If anyone feels like taking this on I'd love you forever.

Does anyone have numbers on the DPS increase from FS, UA, and garg?
Is it possible that with a MH like Last Laugh or Silent Crusader that 0/45/26 (taking both UA and FS instead of garg) is better than 0/44/27?

This is why I was doing all my calculations with 'consumed ppm' instead of actual ppm - yes, the buff overwrites really do amount to a lot of the procs, especially with gear like his. I posted code a few pages back to calculate an estimated consumed ppm given stats, if you want to compare that to the WWS - I'm not at work right now, so I don't have it in front of my to run.

Haste is not an undesireable stat. Though it increases buff overwriting, it increases proc number by the same degree - overall effect on consumed ppm (cppm) is positive. Buff overwriting on the ptr will be exactly like the current model, except that the cppm (and the overwritten ppm) won't change with stats.

I tested both 32/39 and 44/27 a bit more today and they seem about equal near as I can tell. Admittedly however I was unable to find any IT or FS glyphs to do these tests with so take from that what you will.

I found that for 32/39 the standard rotation was best:
IT > PS > HB > BS > BS > RP dump
IT > PS > HB > IT > IT >RP dump

for 44/27 however it seemed to do more damage to use:
IT > PS > HB > BS > BS > FS/DC
IT > PS > HB > HB > FS/DC

where DC was used if KM was active and FS if not. I will have to try the HB > IT > BS > HB rotation I have been reading about though.

One thing I have noticed is that with KM not being tied to crits, procs seem to be a lot more evenly spread out, to the point that they are almost predictable. After a while I started tweaking the rotation on the fly a bit and it seemed I could reliably place HB infront of IT > PS and have KM up for it almost every single time, and if I opted to use anothre HB afterwards, it usually crit too.