Also according to Kahories DK simulator 15/0/56 > 3/13/55 by roughly 300dps on single target fights.

Your point about being hit capped and not having the 3% from nerves really doesn't make much sense to me, most of us are running between 8-11% hit anyway just due to gear is my guess. As long as you are over 8% hit then nerves is only really helping you cap white hits, Its extra white damage (which means more necrosis/bcb) I'm not exactly sure how you are saying its an advantage not to have it, all its doing is giving 3/13/55 3% more white hits, which is a dps gain.

If 3/13/55 beats 15/0/56 on 5 target, it would beat it on 2-3 target as well, there is no difference in the aoe rotation between 3 mob and 5 mob situations. What exactly do you think would change? Perhaps you can clarify?

If the simulator is showing a 14 dps difference, and you are getting a 200 dps difference, can you post some logs with both specs? Perhaps its a raid buff difference or a play style difference? We all know that everyone plays their favorite spec better usually, maye its just that you are more used to and as a result better with 15/0/56?

Edit-

Wait, you are refering us to a post where the simulator results show a 14 dps difference, now you say its a 300 dps difference on the sim? Can you post some evidence of this magical jump?

I think you misunderstood, i wasnt saying not having the Nerves of Cold Steel was a benefit, just saying its not worth as much if u already are at over 11% hit without it. Atm i have no logs to confirm what i have written, but you can check urself with kahories DK simulator or try it out yourself, and also the extra move speed from Imp UH presence can help alot, and yes i might be more used to 15/0/56 or something.

Edit:
I cant speak to the accuracy of Kahories DK simulator as its results generally comes out abit higher than my own actual dps. But i've done several simulations on it with both long and short fights and 15/0/56 beats out 3/13/55 with a few hundred dps every time.

Edit:
Also i might have aproached this the wrong way, what i really want is for someone to do some extensive tests with this build as i think i am not qualified for it, dont know the programs well enough and dont really have the time for it. I just think this specc warrants some more looking into as from my experience it beats out 3/13/55

My guess is notody is going to have the time to do extensive tests, I sure don't. But if anyone wants to work on it, or if you want to do it yourself and post results I'll add it to the OP. Mostly I don't think anyone wants to put effort into this since all known 3.3 changes point to unholy going to a two handed only style. Chances are this thread is going to roll into frost only once 3.3 goes live. Consider will likely remove all DW builds from his post as well.

What kind of extensive tests are you looking for? Quick tests do indeed confirm what you have already said: 15/0/56 seems to be better than 3/13/55 for unholy DW right now. If you want, you could try out my simulator to do some further testing -- it has a simple UI that might enable you to do the kinds of tests that you'd like without sinking too much time into it: Team Robot WoW Simulator.

Simulators will always come out a little bit higher than your actual results -- don't worry too much about that. The more important thing is that the simulator scales correctly with gear, talents, etc. Kahorie's is pretty solid, and my own simulator gives very similar results.

Can someone post these "quick tests?" The post he linked to confirmed the exact oposite, that 3/13/55 was the better spec. If we want to say something is better than something else, lets post info. Don't just post dps totals also, let people know what imput you are giving the simulator so they can reproduce your results. I'd love to throw this through my sim, but I'd need to use the same stat set and priority others are. So far I've seen zero data or math to show why this new spec is "better," but we've seen a very detailed post from a regular contributor showing it is 14 dps higher on single target, but 100 dps lower on multi.

Can someone please explain how suddenly the results we discussed earlier are invalid? IE how is the simulator now showing a 300 dps gain where it was giving only a 14 dps gain on the last page? I'm all for new builds and data, but lets keep it in the realm of theorycraft and math. Lets see some numbers.

What is that? If that simulator is new, the results will be far from being correct. Especially if you're developing it alone. That's just my experience after a lot of spreadsheets and the current sim.

Yes, it is a new simulator. It has been tested quite extensively by several players over the past couple weeks, and the results are looking pretty accurate. If you feel that something is incorrect with the simulation, please let me know specifics -- simply asserting that it is incorrect without trying it is not very useful.

It is particularly suited for "quick tests" like this -- you can easily change gear, talents, rotations, raid buffs, etc. with a simple visual interface, and load characters directly from the armory.

I have done extensive comparisons of my formulas to those used by kahorie, and done independent research on all of them as well. In the near future I will make all of those formulas available for public scrutiny in an easy-to-read format -- for now you can ask for whatever information you need, and I will provide it.

I havnt tried the Team Robot WoW simulator yet, but the results seem to be similar. With Kahories DK simulator, latest version, i used my exact gear and specc showed in my armory profile, i was using FC on both weapons, i used blood presence, i was counting with all the raidbuffs, was doing 100hour simulation with many 350second fights, with the rotation:
it-ps-ss-bs-bb-dc-dc
it-ps-ss-bb-bb-dc-dc

Hope this clears things up.

EDIT:
did a quick test again only 10 hour simulation now but results was

15/0/56 = 8388DPS

3/13/55 = 8112DPS

So, roughly a 270dps difference

It's newly public. It's been in development for a long while now. Yellow and a couple of his closest cohorts develop it, and several of us guildies have been testing/breaking it. I've personally been using along side Rawr for nearly a month now and I can say from experience it outputs pretty much exactly what I would expect it too. There is a thing or two I'm not convinced is correct... but I've thus far been unsuccessful in proving anything is sufficiently out of whack.

In my opinion, the rotation (priority!) editor is far superior to anything else available. It is quite complicated at first glance, though. But the default (Unholy) rotations aren't 100% accurate (they don't intentionally clip diseases). I can PM my rotation xml to anyone interested, I'll be tweaking it in the mean time to make sure it's still accurate.

I had a bit of a play with it the other day trying to get a nice GoD DW unholy build working. You can do things like start-up sequences pretty easily with it. It does seem to be more of a priority editor than a rotation editor though, but the logic allows you to do strict rotations if you want. (and sub priorities etc).

Doc's right though, it does seem to get a few mechanics wrong. In particular, when I tried it at least, the diseases at least were only critting at 150% damage. Still it's another tool in evaluating what to run with. In any case you have to take the results of simulators with a grain of salt, and be smart in iterpretting what out put they produce.

Yeah it didn't take me long to figure out that it was actually a very configurable priority system. I too prefer them and also used to play a feral. Like feral, I find it actually quite hard to dps as a DK optimally, they both have rather a lot of things to try and pay attention to, I think a priority is a better way of trying to handle the complexities of what to do while actually raiding. The objection is purely cosmetic, in that it's called a rotation throughout the interface.

Excellent news with respect to the disease crits. I think most of the theorycrafting for 3.2 is probably been done by now, so I think the best thing to add to it would be a 3.3 mode.

If you want to get more feedback and expose it to a larger audience, I suggest creating a whole new thread, which if you've posted 10 times you should now be able to do.

I found a very nice Addon for DKs.

Bleeding Hearts

It shows me in a bar how much AP I have at the moment and at which AP-Scale my deseases are ticking. Also very good to quickly notice when to pop gargoyle during heroism, if you are playing Unholy. DK really has A LOT of procs flying around with the current content, so this really helps me. And since frost dw with GoD kind of focuses on keeping deseases up at large AP-scale this addon is really helpful.

Maybe give it a quick notices in the starting post?

It doesn't say on the actual mod post nor is there a tab for it under the interface in addons, what is the command to be able to pop up the actual mod? This is a very nice mod indeed for GoD Frost DPS if it indeed tracks as well as they make it sound.

Stats used:

Everything tested with 4PCT9:

Look at 15-0-56



Look at 3-13-55



Looking at this numbers, Nyhmsi is right.

BUT there is a big BUT

Look at 3-8-55



There is something really odd with Black Ice.

Looking at the numbers:

Average Damage	w/o Black Ice (3-8-55 simmed)	with Black Ice (3-13-55 simmed)	with Black Ice as it should be (adding 10%)
Blood Plague	1929.7 dmg	2010.2 dmg	2119.7 dmg
Frost Fever	1924 dmg	2006.1 dmg	2116.4 dmg
Scourge Strike	4645.9 dmg	4853.6 dmg	5110.49 dmg
Icy Touch	3182.5 dmg	2913 dmg	3500.75
Death Coil	5763.8 dmg	6001.2 dmg	6339.3 dmg

The simulator seems to have a problem when switching spec actually and does not calculate properly.

Resetting the sim and starting another sim:

Looking at 3-13-55 again



These numbers now really match the calculated numbers above.

Now let´s look at the difference with 4PCT9:

15-0-56 = 8396 dps

3-13-55 = 8320 dps

Without 4PCT9 15-0-56 is still winning but not winning by several hundred dps. I don´t actually know what has happened to the sim. The 2 Specs were closer before the Bone-Shield-Nerf, 15-0-56 only winning with 4PCT9, loosing without, as you can see in my sims done a few weeks ago.

But looking at the actual numbers 15-0-55 + 1 pulls ahead of 3-13-55 on single Target.

With 4PCT9 15-0-55 +1 does more damage on 3 Targets, too. Without it the difference is minimal in favour of 3-13-55.


With the actual simulator i have to get to another conclusion... 15-0-55 +1 is the actual winner. Coming 3.3 this spec will be even better for single-Target DPS as Black ice won´t effect the whole SS anymore.

The miss rate of white hit in your first 3-15-55 is too high It look like the sim didn't read the template correctly or the template was wrong somehow. But as the code is done, the template file is read at every new simulation and if the file would be incorrect the sim would crash with a nice error message. It's very strange.

I have the English game client installed, however it is not in my Interface-->Addons tab. I don't know why it wouldn't be but it isn't.

I ran into the same problem - turns out it's because Bleeding Hearts depends on LibSharedMedia-3.0 but neglected to include it in the zip. Download it from that location and drop it in the Interface\AddOns\BleedingHearts\libs folder and it should work.

I noticed that the stat weights in the OP for 0/53/18 have changed recently, but unfortunately I can't tell when the change happened. Most notably, spellhit (hit rating after the melee cap but before the spell cap) has fallen a lot (to 0.99 as of this post) and arpen has fallen somewhat. I can understand the arpen changes given 3.2.2.

Going back as far as page 58 of this thread, there are no new stat weights posted since, so I'm very curious about the source of these new weights. On page 58, discussion was still pre-3.2.2 and spellhit still had a high value. I would appreciate if somoene pointed out where the new numbers are from and how they were calculated.

I must have forgotten to edit the changelog of the OP, I ran new statweights at the request of Zerack for a new push of his gear optimizer tool. Statweights are run with an average TOC25N gearset, without 4 piece tier 9. As for the actual calculations I ran 3 separate 2000 hour EP simulations using kahories sim (non 3.3 version) and averaged the results. I find that doing very long sims multiple times and taking the average to be the best way to go about statweights. The numbers in the OP are now exactly what you will find in the default gear optimizer from Zerack. This applies to both frost specs in the OP.

As always, statweights are an approximation, I suggest entering your own stats and running a set for yourself. Chances are each of us have a slightly different statweight based on our current gear.

Which is why I as curious about what run you did for the numbers in the OP. Thanks for the detailed info. One more question: which version of Kahorie's did you use, and what would you say is the reason for the dramatic drop in spellhit's value?

Edit- Removed a part regarding a question for afubar on the simulator's spellhit functions.

As for the drop in spellhit, it confused me as well. I've run the tests many times, and feel they are accurate. Now, I also strongly feel there are several intangibles regarding spellhit that a simulator can't catch. The simulator is hours of endless fighting, without player skill involved. Several situations impact spell hits value, this is anecdotal at best but I think you get the gist:

-Using Pestilence at opportune times to spread diseases from one mob to the next, in order to save time on the 2nd target. Juraxis comes to mind here, use pestilence on an infernal that is about to die so when you turn back to the boss you can simply move straight on to obliterate. Missing here denies you the benefit of skipping disease application on the 2nd target.

-Applying diseases just before you need to break contact with a boss for any reason. Missing here denies you dot damage the entire time you are running.

-Dps while moving, all of our abilities while at range are based on spell hit. The simulator isn't counting any missed damage you may have in this situation due to low spell hit.

-Player skill. Lets face it, sometimes getting a miss on an icy touch causes your rotation to get set back. Knowing what the optimal ability to drop for that extra GCD on the fly isn't easy, we all make mistakes. The simulator does not make these mistakes.

-I'm going to check with Kahorie to see if the simulator is correctly checking a KM Howling blast. Spells use a two roll system, so a KM HB can miss. Not positive if the simulator takes this into account.

This ended up being long-winded, but I think I've made my point. Spellhit is most likely a personal decision that can't be correctly modeled on the simulator for reasons listed above, and probably numerous reasons I can't think of.