Apps

It seems like it's time to start testing the edges of raiding deep subtlety, and we might as well break this madness off into its own thread. The modeling for this and testing are pretty rudimentary at this point, but at the least those of us inclined to try it out can start testing builds and showing parses.

Fortunately for me, I raid with another rogue I generally basically tie with while we're both assassination (Spleen) so it gives some grounds for comparison. Unfortunately, I only have about a third of a naxx clear and a Sarth+0 kill, and no patchwerk until next reset.

Wow Web Stats
Talent build for this parse

Party members were the hunters and DKs. I'm in a mix of 25, 10, and level 70 pieces in this parse.

Play experience and lessons:

Unfortunately, my complete lack of a non-dagger mh forced me into a backstab varient; additionally, not working imp snd into the build was a huge mistake, and there were more than a few play mistakes - dipping below the energy to use a finisher should be avoided at most costs, as it can throw away combo points like nothing else. My basic priority list for finishers was SND, rupture, evis.

Obviously, the gain of combo points is unpredictable, bursty, and - worst of all - not recorded in the combat log to allow for any kind of analysis other than firsthand testimony. My intention is to do heavier testing, in particular as to whether pet ability crit proc the talent and if so whether they share the 1 second ICD with the hunter, however, basics I was able to establish within the raid and solo:
1. As established by the blue post a while back, the ICD appears to be on a per party member basis.
2. White damage crits, as indicated by the talent itself, do not count.
3. More interestingly, neither do poison crits. Blade Flurry attacks off crits also do not count (and in fact aren't crits themselves anyway)
4. You do get credit for your own combo abilities and finishers.

Open questions for testing (which I plan to take care of myself, but if someone's already done the work, by all means):
Channeled spell interaction (Mind Flay, Arcane Missiles)
Pets

Edit:

No, the build in this post is not optimal. Read the thread for more information; in particular, Mavanas' beta spreadsheet is useful, though again I would recommend reading the entire thread.

Zyla: "What's the functional difference between my computer breaking and getting hit by a car?"

 

TraumaSK

I tried out a HAT spec today as well. We did a 10 man Naxx with a druid tank and no other melee buffs aside from the 5% crit, so damage potential was actually alot lower then it would have been if there were to be a warrior sundering or any form of melee haste. I stacked my group with an ele shammy, hunter, and 2 mages. The combo point generation was ridiculous after about 5 or 10 seconds into the fight. Very sporadic at points like the OP mentioned, but I believe this may be "somewhat" leveled out in a 25 man raid with more potential buffs. My overall dps was pretty low due to not fully understanding the build or having any set rotation, but I believe that with practice and knowing what to look for and when is going to be key in this build, and frankly I did not know my head from my ass walking into 10 man naxx with this build. Build during raid: Talent Calculator - World of Warcraft

From my findings and experiences(limited as they may be), the potential dps of this build is absolutely insane. It does require alot of attention towards your energy bar..probably more so then usual for combat/assassination rogues. I found myself out of energy with 5 combo points many times throughout the night, and I would just sit there auto attacking waiting for energy to go back up. Not having the increased energy regen from vitality or focused attacks really can put a hampering on dps if you don't spend your energy wisely. So as Apps said it's very important to not get ahead of yourself and end up without any energy left to spend on finishers.

As for weapon choices I was running [Calamity's Grasp] [Omen of Ruin]. Wound/Deadly. Although, with the amount of finishers I was throwing out, and the general lack of using Hemo, it may be more advisable to do more of a mutilate setup with two fast daggers and Wound/Wound.

For those who haven't seen the potential dps that this build can throw out, check out this wws. Wow Web Stats

The other talents and abilities the HAT spec entails are things like Vanish for master of subtlety, and shadowstep for the increased damage on next attack. As for how those are used, my guess is that it's Vanish>ShS>Garrote. Or at least that's how it jumped out at me initially. With prep that combo is/could be used a few times throughout the fight.(Could be wrong on how these tools are used though.)

My foresight tells me this is one of those builds thats going to improve leaps and bounds as gearing improves overall and things like crit, hit, and group composition become more consistent. I'll try this out again next week for 25 man raids and try to report on any findings or changes that I see in 25 man raiding. I'm sure it'll only get better.

 

Nerevarine

I tried it last night, just for fun It was pretty interesting play but i must say i completely destroyed the power of this Build, tired from work + laggy server. The main point i realized is, as you stated always have enough energy call a finisher, sometimes i had to wait for about 4-5s because of no energy + kick was needed :/ And also it is not the Build to play with more then 500ms =)

Wow Web Stats anyways but not a very talented play :/

 

Kumar

The problem with studying HaT builds right now specifically is that some rogue players are reporting that with two HaT rogues in the group, you get two CPs per crit instead of just one, which would be a bug and a big one at that.

 

Apps

Right, but that doesn't mean you can't study parses and builds without having two HAT rogues in the group.

Still, if people could at least have a parse, their group makeup, and their build of the time when posting about the spec, it would be far more helpful than a rough description or a standalone parse.

Zyla: "What's the functional difference between my computer breaking and getting hit by a car?"

 

Nerevarine

which would explain the huge amount of finishers pimpy does. Under 4s for Evi alone not counting SnD and Rupture Finishers

 

mostlikely

I've been playing with this too.. having to use abilities reactively sure is refreshing after years of combat.

I was wondering since combo points can come in such bursts.. how viable is it to do 4pt or even 2-3pt finishers?
Sitting on 2 combo points waiting for more with the chance of capping and losing points in a short burst feels like a worse option then wasting some energy.

Also it often felt like combo point generation stagnated until I was forced to do a *hemorrhage* (edit from: mutilate) after which I got a burst of combo points.
Often it felt like Hat only worked after I had 1 combo point, but it could have been the laggy server.. anyone else experience this?

 

Apps

So, out of curiosity, how DID you get the 84 talent points to have mutilate and honour among thieves?

Zyla: "What's the functional difference between my computer breaking and getting hit by a car?"

 

Tinbum

That's a build I'd like to try...

 

koaschten

He obviously is lvl90 already. But this is offtopic here.

On the topic of when to use a finisher, i dare saying, pool energy and

- if cp < 3 and energy > 90 -> hemo
- if cp >=4 and energy > 80 -> finisher
- if cp < 4 and energy < 40 -> pool energy

you get the idea. Dont waste CP, but neither energy.

 

mostlikely

Sorry.. .. I keep confusing hemorrhage and mutilate.. I meant the former naturally...
My talent build is 6/21/44

Sorry for the confusion.

 

Faylinn

As far as we know HaT is "working as intended".


Wow Web Stats

As you can see it's more or less an Eviscerate every global cooldown. So don't forget to refresh SnD while smashing that one button.


/edit: Ok, now i understand the possible bug Kumar was talking about.

 

Sarah

Having seen Pimpy's report, I find this one a bit more imppressive:

Wow Web Stats

Excuse the speculations, but could it be that with 3 HAT rogues in a group, everyone gets 3 combopoints on a crit?

 

Tinwhisker

Apparently so.

"Buggy" Rogue Mechanics?

You're a lot of DPS, you know that? You wanna' earn 14 achievement points the hard way?

 

doozer667

I'm praying it's not a bug and actually working as intended. IT seems incredibly OP at face value but it is called Honor Among Thieves. Rogues = Thieves.

 

TraumaSK

Well, if HAT is currently giving an extra combo point per rogue with HAT in a given group, it's safe to assume one of the following:

A) It's working as intended, but has a tooltip thats already misleading and missing information.

OR

B) It's currently bugged and probably going to be fixed in an upcoming patch, or hotfixed rather soon.

I'm going to unfortunately go with option B as of right now, because it's hard to believe that the tooltip on the ability is that wrong. So, in this case, during the next raid reset I'm going to gather a combat parse with one HAT spec'd rogue(myself) and 4 other fast attacking high crit ability classes. (Fury war, hunters, another rogue not spec'd HAT, Enhancement Shammy to name a few.) Thinking about it, if the ability is bugged, and the extremely high dps results we've seen on these wws reports are because of that fact and no other, I'm not sure if the build can keep up with combat or mut. But, the only thing that will prove this is more testing.

Also, I really like and agree with what koaschten posted. Setting rules for certain situations and scenarios looks like a must for this build. Finding the right approach to do what and when based on current energy pool and CP seems like it's going to really help in maximizing dps.

As far as energy maintenance goes, due to the fact that a HAT build has little to no energy compensation outside of Relentless Strikes, making sure you have higher combo points to save finisher cost looks like it's going to be very crucial. If a 5 point eviscerate effectively only costs 10 energy, but you still have to maintain the initial 35 energy to use it, then it's safe to assume that one should never drop below the 35 energy threshold if at all possible. Which makes me think an effective method of using evisc would be to only use it at 3+ combo points to maintain necessary energy to continue to eviscerate when applicable.(SnD and Rupture are both currently up.) Of course, this is all based on whether or not you get the combo points fast enough to allow one to have relentless strikes proc and leave you with a 10 point decline in total energy per eviscerate.

 

Apps

The first post contains a parse and discussion assuming only one HAT rogue. It still seems to be competitive; I gained from it, though in a very small sample size, but also with great room for improvement.

Zyla: "What's the functional difference between my computer breaking and getting hit by a car?"

 

PessimiStick

For what it's worth, I saw good results both at 70 and 80 as the only rogue running HAT in my guild. We just all respecced last night because we knew about the bug. The spec itself is definitely competitive with Combat and Mut standing on its own, it's just... a bit better when you exploit it. Once the bug gets fixed I will probably stay HAT for a bit and I'm assuming Injek will go back to combat and Deathbiscuit will go back to Mut, so I can hopefully contribute some un-bugged parses then.

Edit: As for what the bug actually does, it seems that crits from non-rogues work as intended, but a crit by any rogue in the party gives a CP for every HAT rogue in the group. When it was just the 3 of us outside under Naxx before we started, a Hemo crit on a fresh target without the other 2 rogues attacking would give them 3 CP and me 4CP. Also, from playing it on bosses, it seemed like the CP gen goes berserk once you start spamming evisc. The "warmup" period where we are getting SnD and Rupture up does not have the same infinite CP feel to it. If someone else wants to specifically test the mechanic, that would be a good starting theory, in my opinion.

 

Nerevarine

I basically can confirm this one, i was on par with the combat rogue, with Rupture Gylph from his old server :P, playing it the first time and Im curious to see if it will scale better than combat concerning group itemization.

 

ekval

You should go two fast high dps daggers instead of using Calamity's Grasp or variable slow fist weapon in mainhand. You should not need to use hemorrhage except in rare occasions, and fast weapons give you better poison damage.

Build like this is probably the best. I don't think ShS brings much on the table compared to 15% Rupture damage. Glyph choices are pretty self explanatory, you basicly want Rupture and SnD glyph so you can Eviscerate more, and of course Glyph of Eviscerate itself. Glyph of Preparation resets Blade Flurry cooldown but I don't see that big investment compared to previously stated ones.

Gem for AGI & Expertise is my bet.

 

Arogue

Right now the most important point is to get at least 2 HaT Rogues in a group.

Tested with a Group of 2 Players (Rogue/Rogue)
If you have a nonHaT rogue and a HaT rogue, the HaT rogue will get 1 ComboPoint/Critt
If you habe 2 HaT rogues, both will get 2 ComboPoints.

(Glyphs were Evi/SS/Rupture and SnD/SS/Rupture)

 

Nerevarine

if you bring the poison argument which is of course a good point i would go for AP instead Agi

 

ekval

Agility scales with both Deadliness and Sinister Calling, AP only scales with Deadliness. AGI also gives crit which is beneficial for Eviscerate.

 

Amerilina

I hacked a hat cycle into Vulajin's Spread sheet. It was very rough and I made a few assumptions, but Agi seemed to be king. You also need to figure agi and AP are almost on par for rogues now. The fact sinister calling gives 15% more agi compared to deadliness 10% would logically make agi more valuable. Also when looking at stats, Assuming your finishers can't miss armor penetrations relative worth would go. As this spec is relatively unknown or modeled most things here are speculation.

 

Ashere

As for the WWS's, the thing we're mostly interested in is the amount of abilities that can crit that each class/spec puts on the table during a given timeframe. White DPS can easily be worked out with existing models, the avarage damage of an Eviscerate can easily be worked out with the existing models. It's the CP generation that's hard to model since it depends on the members of the group the HaT rogue is in.

So, if you'd like to model HaT for a given rogue, look into the other classes in his group, get their amount of crittable abilities/second, which may or may not depend on gear, multiply that with their crit chance, which does depend on gear, and you'll get the avarage amount of CP generation through HaT for each groupmember.

Now for how to model this more effectively, to get something more generic to use.
We'll want:
-The base amount of critables per second for each spec while not geared/buffed
-Input fields for each group members' class, spec, hasterating when buffed and crit chance when buffed
-We'll want multipliers for each spec to calculate the effect of gear and buffs for Haste to the amount of crittables per second.
-Calculate the amount of crittable abilities per second for each groupmember, using the above information
-Multiply the amount of crittable abilities per second for each groupmember by his chance to crit.
-Cap the results at 1
-Add all these values up, and you'll have the avarage amount of CP's generated through HaT per second.

From this point, you can start designing avarage rotations, work out whether you'll need to weave in your own combo builders, work out how to manage this with energy generation.

And you'll still end up with less reliable cycle design than Mutilate, which also occasionally has bumpy CP generation.


I know, long post, but maybe this explains for a big part why Valujin and others are so hesitant to include HaT into spreadsheets. It is a LOT of work to figure out all these things, a LOT of work to implement, and even then, it'd have so many more uncertain factors (each party member being a seperate collection of factors, skill not being the least) that the result would be unreliable at best. Basically, the best guess I have for implementing this into a spreadsheet is by adding an input field for "avarage amount of critting abilities of other groupmembers per second" or "estimated CP/s from other members (0.00-4.00)" and adjust CP generation and cycles to that. Still highly unreliable, generally severely overestimated by those trying it, but it's both the fastest and most convenient method of implementing HaT.