It surely can be optimized, but the simulator try to refresh with pestilence when disease are at 1.5s to fade if there is a blood rune available. I'm open for suggestion to enhance it.

I don't mind you modify the simulator, but if you have some code that you want me add into the "official" simulator, let me know.

Are you still using 200ms for you delay? That seems high from my personal experience, and will bias toward the unholy DW spec. Also have you been using SoA and prioritizing OB over FS? In 3.2 its blood strikes that you want to drop in order to get more FS's in, as OB looks like the most damaging ability.

Anyways, by my paper napkin calculations, switching the delay to 100ms should net around a 100 to 200 DPS increase for Unholy in unholy presence, but a 400-500 dps increase for Frost in blood presence.

Yes I used 200ms delay, due the fact that the usual delay plus human natural reaction adds up to that amount (probably more). You can't expect someone to hit the buttons at the split moment all the time (even if you faceroll xD or keep pushing 1 button continually). So it's reasonable to do so. I'd thought that 400ms - 500ms was actually more accurate, but i'll leave it to 200ms to evaluate people with really fast hands. (This was suggested to me by Orlgin, and he was suggested to by someone in shaman forum).

As you can see, the more buttons you have to push, the more the spec suffers from delay & human reactions. These will be taken into account when I choose my personal spec, and people should take a look at that as well when deciding which spec thay wanna play.

If anyone has better luck with tweaking Deep Frost to maximize DPS i'd really like to hear about it. However these are the best results I got so far:
Best spec of Deep Frost: 10/52/7 (Glyphed with OB, FS, BS, Awareness RI/FC BP) pulls out the best numbers with priority Diseases > BS > OB > FS > Rime. However it's still lower than 0/17/54 by a good 100 - 150 dps. However Necrosis is bugged as of present sims, so 0/17/54 would top that by a bit more.
Comparing 10/52/7 with its shorter diseases cousin 15/55/1 or something like that, 10/52/7 gets a little more dps due to longer diseases time ensuring the best use of Blood Strikes and Obliterate.

When Necrosis is fixed, 3/53/15 will probably be able to compete. Necrosis is a good dps increase per point for DW while Bladed Armor doesn't scale (although it's very good to start out with), some extra strengths from Ravenous Dead is probably better than 2RP every 5s. It's not worth it to get to BCB since slow weapons are terrible as offhand for that. 3 points in Subversion will help nicely because most of your damage comes from Obliterate and Blood Strikes (approx. 35% from BS + OB, Subversion then results in approx. 1.05% dps increase per point).

Yes, BSx2 is always better on average than OBx2 for Death Runes. At least in sims that's the way. In practice, if you can soak FS, don't bother use BSx2 for DR because OB + extra FS from soaking AMS is probably better.

In 500ms ping, OB is better use than BSx2 (200 haste). Also in 500ms ping, the difference between 0/17/54 and 10/54/7 are shortened but 0/17/54 remains higher.

If someone finds out other results I'd really like to hear it.

Two problems with that analysis. One, you arent reacting to the global cooldowns, you are to a certain extent anticipating them (or button mashing).So human reaction time is moot, since your not actually reacting to anything. Second, I believe the client/server interface has a little bit of compensation built into it.

Personally in dummy testing, running Unholy Presence with a 6 IT+PS rotation, I could consistently get 9 abilities in before the first rune refreshed, with a little bit to spare. Its going to vary between individuals though.

Edit: There really isnt any reason you have to guestimate this. Any sufficiently long dummy test should be able to let you infer your average number of ability activations per second.

Yes, during dummy testing, you can simply smash the buttons. During the real moving fights however, it's not the case. AFAIK, yes the game has a mechanic to reimburse you with the loss in minor lags, which is the 1.5s buffer rune cooldowns. However this is not going to help the builds which have to use up 9 GCD in 10 seconds in UP, or 6 - 7 GCD in 10 seconds in BP.

You are reacting to the GCD cooldown & rune cooldown, also the RP changes. Even if you're smashing button, in order for the delay to always be within 100ms windows, you'll have to smash your button at least 10 times per second. Because if you don't, chances are you'll push the button 1ms before the cd finishes (it won't let you input it), and the next one you hit will be out of that 100ms window (Ofcourse this is to talk about with no ping lag at all). I don't know about you but my keyboard is not gonna be very happy for me to press 10 times per second for 1 button!

9 abilities will go like this: First ability starts off, rune starts counting 10s, GCD counts to 1s then you can fire 2nd ability. The rest 8 abilities will take up 1s each of GCD, which still leaves you with 1s to spare for the button-smashing delay, which is the maximum of 125ms delay/reaction. That's pretty impressive considering you have to smash your key at least 8 times per second! You may used reacting system rather than smashing, so that's pretty impressive too!

200ms is already not realistic for button-smashers (5 times per second), but it's for good gamers who can react fast enough and time their hands themselves. 400ms - 500ms sounds more reasonable to me (2 - 2.5 times per second).

First, the compensation will make a difference because it is factoring out a certain amount of latency on the network.

Rune cooldowns and global cooldowns are predictable events. You arent really "reacting" to them at all, because you know they will occur, and when, before they actually occur. With practice, you can more accurately predict when those events will occur. Practice and skill will allow you to time your abilities right.

And if you are mashing 5 times per second it will yield an average delay of 100ms.

Assuming your button pressing is asynchronous, you will have a random distribution between >0ms to <200ms before the ability comes off of cooldown. The average of that distribution will be 100ms. You would actually have to try to get up to 200ms of delay if your pressing the key 5 times per second.

Button mashing isnt the way to do things mind you, but even that should yeild better then an average delay of 200ms.

Edit: Its kind of a silly thing to debate in abstract though, what your average number of activations per 10 seconds really should be something players should keep track of. Not only to help select the right build/rotation, but also as an area in which you can work to improve, as it will directly improve your DPS.

It works on the boss DUMMY, for sure. I don't think it works on actual raid bosses, though. When I solo ZG I have to switch to my pvp frost spec for Hakkar, chillblains pops up "immune" and does not apply the debuff to him. I don't know if he is a special case or not because I haven't done any other bosses as that spec.

@Edimasta: BS Glyph works with Chillblains but there are tons of other abilities your raid will bring to help you with that. Glyph of IT is only a bit of DPS loss (not good in sims) but it's more versatile. I'm only using BS glyph to make the sim number closer to the full potential, to create a 'guidelines' kind of thing, but in practice, it really depends on personal taste.
@Odii: I really calculate for the worst. It's prudent to do so. The rune buffer mechanic is not going to help if you have to dish out 9 GCDs in 10 secs constantly because, if 1 of your rotation takes 11s, in order to catch up, your next rotation has to be 9s otherwise you'll waste a bit of time. If it's constantly 11s then it just means you dished out less than 9 GCD in 10s. It helps in several situation though, like when you use 7 GCD in the first rotation in BP, but only 5 GCD in the second rotation. In the total of 20 seconds rune cooldown you can still dish out 12 GCDs without wasting any extra seconds.
As far as calculating for the worst goes, you can't just average out 100ms delay by the time from 0ms to 200ms. If you have 20s of 150ms delay, then you're wasting a few seconds. Rune buffer won't help you there. On the other hand, if you are good enough to have 20s of 50ms, you'll end up waiting for rune cd to finish, and won't receive any benefit. So it's really neutral and minus (either you lose something or you don't), not plus and minus (gains make up for losses) situation to have an 'average'. The sims always assume you push the button at least within 100ms window though - so the real results will be very different, making the sims even more unrealistic.
It's predictable event, but gamers aren't machines, and neither am I. I'm leaving 200ms to my confident instead of 400ms - 500ms for a guarantee.
P.S. I tried with lower ms values, and Frost can't be better by 400 - 500 dps like you said. Perhaps it's because of your rotation which can't fit into a high ms threshold? Could you please tell me your used priority so I can optimize the deep frost? I got the same difference results for Deep Frost vs. Unholy even with lower ms, and Deep Unholy always pull forward by 100 dps or so. I'll do some numbers in 0.9.2 too and that's gonna take a while.

No we've established chillbains does not work on raid bosses. Someone said that previously it did, but it's been tested on sarth and other raid bosses to confirm that raid bosses are immune and not getting the debuff applied to be worth putting a point in it. The only way to get the glyph of Blood Strike to work is with the druid infected wounds or mage frostfire bolt debuff.

I think the my point would be to use actual real values for activations per second. I understand why you are using the 200ms, but you have to acknowledge that arbitrarily picking a point will effect the relative balance of the different specs/rotations. And to arbitrarily decide its going to be one spot or another isnt maximizing your DPS. And while I think I overestimated the impact of the latency changes with my napkin math, I dont think its wise to pick a point without checking if that point is grounded in fact.

As for why I got that increase, I figured going from a rotation of 22.1 second going to 20.8 seconds is roughly a 6.5% dps increase, and did some math in my head. That estimate is high though as far as practical simulation is concerned.

As for the best results, I got them using a rotation of OB>OB>FS>BS>BS>FS:OB>OB>FS>IT>PS>FS. Reasons being 1, I like having rotations that use a 1:1 death rune usage (ie, BS creates 1 death rune, IT uses 1 death rune), and it naturally is going to maximize killing machine procs (IT always comes after a FS, and Frost Strikes are naturally threaded throughout the rotation without having to interrupt any rune cooldowns. Using a rotation also lets you get a rhythm going, meaning you can more quickly use your abilities.

Edit: Ill add, glyphs Obliterate, Frost Strike, Blood Strike

My newest "discovery" on the PTR thanks to Miks Scrolling Battle Text:

When you hit with a strike, there is a seperated chance for MH and OH to hit critical.
This occured several times, f.a. when hitting with OB or FS.

One time there was only 2 Hits => 1 Crit and the other time 2 Hits => 2 Crits.
Pretty interesting for me

If this was posted before, sorry for that... was just curious about that.

Got myself 2x Arena Axe and played about with 3/54/14
Necrosis is about 4% in Damage, pretty acceptable for 4 points in it.

OB crits are about nearly 7k, FS crits only about 5k. Really depending on if you crit with both weapons at the same time
or not.

Just one thing I've noticed that I'm not too sure I care much for.

I'm really enjoying 15/53/3 as DW on the PTR. It's so nice not having to rely on a pet for a part of my DPS, especially when they have a tendency to die on any kind of AOE fight.

With this spec, I used BP, and this was my Rotation

IT > PS > OB > BS > BS > FS > FS
OB > OB > OB ***** FS > FS

Where the * are above is the point in which the diseases from the IT and PS drop. With latency, it's before the third OB in the second rotation, but the 2 FS in the second rotation don't get the bonus damage from Frost Fever from talents.

I'm having a hard time deciding if the third OB would be worth replacing with 2 IT instead...

Further testing with CG/RI it seems CG really does only apply to MH with its boost.

highest offhand crit with cinder was around 1600+~.
Highest with OH crit with RI was around 1800+~
That right there is about a 10% difference.

I was using a normal training dummy(80) instead of the heroic one since it had ebon plague on it. Interestingly enough using a crappy rotation of PS>IT>OB>2xBS Then 3 OB. with FS weaved in there I was getting actually 100 dps more with CG. (Could be due to the crappy gear I am using and stuff like FF not scaling enough from my AP. Only using ilvl 200 weapons)

It seems thus far that the specs people are trying are deep frost with a split on subspecs, with at least 10 in blood.

For the first 10 points in blood you gain:

* 120rp over 5m
* ~333 AP (figured with 12k armor)
* 9% crit on BS and OB

Reading the thread I see initial discussions concerning Necrosis and BCB additions, but the conversation seems to halt and shift to the dual subspecs.

With a setup as: http://talent.mmo-champion.com/?deat...&version=10072 you lose

* ~333 AP - 84 AP = 249 AP (Though the 84 AP in the tree is str, which is further modified by self/raid buffs)
* The other items listed above

The question is simply, which pulls out ahead - the passive damage due to DW from BCB/Nec or the 9% crit on heavily used abilities coupled with 24 RP/m? I know that when DW was in its heyday at launch, some changes went into effect from Necrosis and BCB. Is this the limiting factor for the usefulness of the build?

From the simulations I've run, the hybrid subspec seems to pull ahead slightly (60-100 dps) every time. This was tested both using a gearset optimised for a hybrid subspec (the EP/stat values I posted on the previous page, which means ArP gets a high value but haste a quite low one), and my own gearset aimed for a 2handed unholy build (but using two onehanders obviously) meaning more haste and no ArP at all, which in theory should favor a necrosis/BCB build. To get a more fair comparison I'll run some EP calcs for the unholy subspec overnight and see if another gearset would improve the results.

Edit: Nothing really interesting to report, the EH values were slightly different, but still ended up with the same gearset regardless. Something to note is that currently the simulator adds FS/RoR damage bonuses to autoattacks as well. I removed that, but it didn't change the relative dps values between the specs much at all (it did lower the listed dps of both specs by a few hundred though).

I have read all the preceding posts, and I did not see a mention of this. There is unique interaction between ToT and Rime. Is there an unspoken concensus not to talk about it or has no one discovered it yet?

Decided to go on to the PTR to check this interaction of Rime with ToT, and Rime can in fact proc from both attacks, as seen in this combat log extract (Freezing Fog refreshing twice in ~0.8 seconds):
The first thing I noticed though is that the offhand no longer hits/misses separately from the mainhand. Tested this with an axe (my axe skill is still 270/400) in one hand and a mace (400/400 skill) in the other, and the results seemed consistent. I obviously got a lot of misses with the axe in mainhand, and it didn't even attempt an offhand attack. Putting the axe in offhand I never noticed an offhand miss with Obliterate (only tested ~100 Oblits or so, but with the easily 30%+ miss/dodge rate when using the axe in mainhand I think that's proof enough).

Are the procrates on PTR still bugged? Unholy Blight? These are things that need to be checked before anything can really be accomplished. Although if KM is only affecting the MH that would seem to be a bug that we need to make sure gets attention, most likely a product of MH/OH critting independantly.

Procrates didn't seem to behave weirdly from some random dpsing on the dummy, not sure what abilities it affected though? Fallen Crusader was proccing from offhand hits at least, I didn't do any extensive testing on other procs though.
Unholy Blight is a straight +20(28)% damage bonus to Death Coil (in theory I suppose it could be based on noncrit damage and deal slightly less damage, but considering the mechanics it's based on that's kinda doubtful), easy to simulate/calculate even if it's still bugged on PTR (didn't test).

I've got a question that I haven't seen brought up yet. Since KM is only procing the MH and not the OH for crits, if this isn't a bug and remains this way, would that push rime proc HB as the priority for KMs when available or are the 'half crit' FS still doing more damage?

30% Glyphed. Unless they've completely changed the intended design which I haven't seen any news of, its supposed to 'roll' like Ignite, which last I checked does not function and each DC simply overwrites the last. Until someone can actually test it and find out how it stacks, when it stacks, and what modifiers it may or may not get there really can't be any solid conclusion drawn. Just because it's supposed to work like Ignite or Deep Wounds doesn't mean it will.

Assuming mmo-champion/wowhead are up to date it's 28% glyphed. Besides, I think it makes more sense to assume it won't go live with odd bugs for mechanics that work properly on live. If it does go live without a working rolling mechanic, that just pushes an unholy spec even further back where as far as I can see frost is already superior.