One thing that has become clear based on the analysis of WotLK mechanics so far is that Envenom is going to play a much larger role in our DPS rotations than it did in TBC, for a couple of reasons.

First, in TBC, the only reasonable raiding builds were Combat Sinister Strike builds - Swords, Fists, or a combination of the two. For these builds, it was ultimately concluded that SnD/Rupture cycles were going to be the way to go, so there was no room for Envenom at all.

Second, to the extent that it was used, it was generally felt to be inferior to Eviscerate; while the initial damage might be competitive, the loss of Deadly Poison damage was sufficient that most estimates put it behind Eviscerate. So even if one was going to use a finisher that was neither SnD nor Rupture, it was probably going to be Eviscerate.

However, with the advent of 3-finisher cycles for Combat builds, the rise of Mutilate/CttC builds, the vastly improved scaling of Envenom relative to Eviscerate, and the higher relative armor value of bosses, Envenom has most likely passed Eviscerate as the instant finisher of choice, and there's now room for it in the cycles of the best known raiding builds. As such, if we are going to have accurate models of Rogue DPS, be it in a spreadsheet, RogueCalc, Rawr, or something else entirely, we need a better model of how Envenom actually works.

In particular, the damage of Envenom has 3 components that all must be computed in order to estimate the overall damage delivered by it:

1) Initial Envenom damage
2) Bonus Instant Poison procs
3) Loss of Deadly Poison damage.

Lets go through these in order.

Initial Envenom Damage
This one is pretty easy. We know the base damage of the ability, we can work out crit rates, resist rates, raid buffs and the like, so there's nothing mathematically hard here. We will, of course need to look into spell resist mechanics more than we have previously, but perusing the various caster threads should turn this up easily enough. And there are some questions in terms of how buffs stack, but that's probably better done in another thread. In terms of the actual modeling, this part is easy.

Bonus IP procs
Again, easy. A n-point eviscerate gives n+1 seconds of increase IP proc rate. So if you know the number of attacks that are being made per second with the IP-ed weapon, it's a simple matter of multiplying together the number of attacks per second, the duration of the buff, and the increase in proc rate (15%) to get the number of extra IP procs from the buff. And as calculating the damage of an IP proc is no harder than calculating the damage of Envenom, this part is again easy.

Bonus DP procs
So, here is where the challenge lies. Deadly Poison has always been tricky to model, and Envenom makes it even harder. There's really two issues here.

First, the question of "how much damage do you lose from a single envenom" is actually sort of a tricky one in it's own right. If you assume that you have DP stacked up, perform a single Envenom, and then let things stabilize again, you can put together some estimates - but they're going to depend on how many stacks you had up at the time, and how long you have left on your stack. So even that proves to be a fairly sophisticated calculation, and one what we've never really solved.

Second, the real situation is even worse than that, because you're going to be envenoming again before the situation has a chance to stabilize. The exact timing of the ticks becomes highly relevant, as you're going to want to do another envenom again in less than 10 seconds in many cases. So having IP tick at 3, 6, and 9 seconds is really fundamentally a lot different than having it tick at 2, 5, and 8 seconds, because it can easily be the difference between 2 ticks of DP and 3 - and with a tick doing as much as 1000 damage in good gear, that can matter a lot.

It seems to me that we might almost need to evaluate Envenom only in the context of a cycle, rather than trying to analyze the damage of the move by itself - as the timing and spacing of Envenoms is going to matter quite a bit in terms of ones total damage output. Of course, even if that's the only way to do the calculation totally right, it's worth having some standalone estimates in order to make quick and dirty estimate as to whether a cycle is at all viable.

So: anyone have any ideas about how to go about modeling this?

Without being able to Envenom, your Instant Poison damage will go down.

Weapon switching to keep the poison requirement, seems like it will also be a DPS loss because you are losing GCDs and I also believe it resets your swing timer. Not to mention it might feel very clunky having to switch weapons on top of watching 3 timers already.

So, the point of this discussion is to answer exactly that sort of question. I *have* looked at using Eviscerate instead of Envenom, and I'm *pretty* sure it's not as good, but our model of Envenom just isn't good enough to answer that question with any degree of certainty. But lets be clear: the point of this discussion is not that Mut is definitely better - it's that we need a good model of it before we know whether it is or not.

I'd also say that unless you're a lot more patient than I expect you to be, you're not going to find anything of interest about the relative merit of the options based on target dumbies - it would take hours if not days to come up with anything remotely conclusive, and even then you'd need to have all raid buffs active. Testing is all well and good for figuring out mechanics, but once it comes time to work out cycles, it's just not very useful.

Do you have the correct formula's for rank 2, 3 and 4? Right now at least the tooltips for these three ranks are the same:

Something else to keep in mind that may or may not be relevant is also for those of us who are mutliate and use envenom to keep slice and dice going. I know for myself I always do cheap shot, snd, then mut to stack up the cp, but once my snd gets close to falling out I use envenom to bump it back up to 20s, but I also use envenom for dps as we are discussing too, so it might be a little more of figuring out which rotations would be best depending on what you are doing.

I would agree that it only makes sense to evaluate envenom in terms of a cycle.

If we assume the rogue is paying attention to DP ticks, I would expect that we can assume he lets the poison tick right before performing an envenom. Since envenom resets the DP stack, we need to calculate how long it is before the first DP proc, and then any envenom is a multiple of 3 seconds after that. Things get a little messy if we model energy pooling in addition to watching DP ticks - modeling will have to account for which should take priority.

As for stacking DP back up, I would expect that just using the same logic as we would on IP would work (how many DP procs do we get during some period of time?) and then modifying the average to adjust for the fact that there is negligible benefit after 5 procs.


Finally, evaluating finishers in a vacuum has always been inexact as compared to their value in a cycle, so a first order approximation is really all that's needed anyway. In this case, I would think that envenom modeling as has been done in the past would be sufficient, just including extra IP procs and faster DP stacking.

I thought about his as well, but do you pool until just after the next DP tick or do you let DP tick until you're no longer able to safely pool without capping energy? Is getting that extra tick (or maybe two) better than moving your cycle through faster? I.e. over the course of a fight is it better to have X DP ticks and Y+3 Envenoms or is it better to have X+6 DP ticks and Y Envenoms?

My own thought is that having more Envenoms and thus more time under the increased poison proc effect would be better, which is slightly counter to the idea of pooling, isn't it?

The estimate time to reach a 5 stack of deadly doesn't just depend on weapon speed, it depends on how many special moves you perform as well.

Because of this, we would have to model energy pooling in consideration as well, since pooling affects the number of mutilates and finishers we use. Getting two mutilates and a rupture off under the +poison proc buff is clearly better than just one mutilate. The extra mutilate would have a chance to proc IP as well.

Here are some relevant questions I have:

Envenom shares mechanics of both spells and melee attacks (it isn't affected by armor, but it can be parried/dodged/blocked), but what mechanics does it share with each?

Can envenom be fully resisted? (not missed)
Partially resisted? (based on mob resist as well as innate 6% partial resist chance for bosses?)
Can it glance?
Is it affected by buffs/debuffs that affect spell dmg/crit/hit?

Can an envenom hit itself proc poisons?

It seems to me a couple things to further consider when modeling envenom in a cycle are as follows (note: Ihaven't worked on ANY of the math for this, just jotting my ideas down in case someone isn't as busy at work as I am currently)

1) It seems iin regards to DP pooling there will be a threshhold where it is acceptable to allow your energy to cap out to wait for an envenom (As long as SnD isn't close to expiring) to allow for another tick of DP. If this threshold is negligable (ie less than a second, possibly more to account for human error and the like) then the models can safely ignore this threshold and still confidently state that if given the choice to wait for an DP tick or energy capping, envenom early. On the other hand, it may be a fairly significant threshhold (2-3 seconds, I would say) it might alter the (so far) fundamental idea that energy capping is an absolute "no-no". This is especially true, in the cases of letting energy cap, if the DP will proc before the next tick of energy regen (save for bonus energy from talents) as if the energy tick (Etick) occurs After the next DP tick would occur, you really lose nothing as your energy will drop to a regenable value before the tick goes off. This is a fairly simple mathematical problem in concept, essentially: DP_dmg_per_tick / Sec_left_to_tick = DP_DPS_at_time_t if this value is > the potential damage that the tick of energy could provide then it is more beneficial to let your energy cap out and wait for the tick.

2) It also seems that at some point (depending on gear) double loading DP on your weapons would be more beneficial than IP and DP (namely in a "lower-level" case, that is fewer gear choices and an unoptimized raid). I know for the highest content this doesn't affect play, but for MODELLING purposes, it is relevant.

3) By the same token, it would be theoretically possible to achieve a 5 stack of DP in the time *between* normal DP ticks, in which case, at this level of gear and buffs/debuffs, there is no loss in dps due to lack of DP damage from using envenom. So, essentially, there is an upper bound ("cap") which could significantly alter cycles, depending on the rest of the discoveries made during this process of hammering out envenom.

I'll ponder some math.

Yes special moves. Forgot about that. As for the rest, I cannot tell you.

I strongly suggest starting to model this off combat builds, it seems more stable and less prone to variations in when an evenom is done. 5SnD/5Rupt/5Evenom is very stable and, aside from SS Glyph, you rarely need less then 13 SS to complete it (well, I rarely need less then 13 SS with my gear and current spec).

Also, without pooling, there may be increased down time on deadly poison, which may be an issue when providing the master poisoner debuff for a raid.

This depends on whether envenom can proc poisons, and I assume it can, and also if it procs the poison before or after removing deadly poison. I'm not sure if this has been considered in previous modeling.

At a proc rate of 45%, there will be an expected proc after 2.22 hits, so 3 hits is a safe expectation to proc deadly poison. Worst case that would take 3.6 seconds to get 3 hits (without pooling), 2.4 sec with. Best case there would be 1 second of downtime (gcd on mutilate), or less if envenom can proc deadly poison after it removes the stack and you make a white hit during the gcd.


What sort of raid dps loss would be reasonable for 3% less crit/sec? Does it affect the dps of envenom enough to consider in the discussion of using envenom v. eviscerate?