Rogue Simple Questions Thread: Cataclysm Edition - Page 62

Tumblebeer · 11/20/11, 2:27 PM

It could be worth adding the post that started the discussion:

The question posed here for discussion is whether to use tricks the second before you pull, ie. together with your prepot, or if you should use it a few seconds in to the fight, when your trinkets etc have procced.

Hypothesis

Case 1: You use tricks 1 second before you pull.
Theoretical pros - You do not waste a global cooldown to use the tricks in the fight, since you can expect yourself to be gcd capped (and thus in essence energy capped) until a few seconds after your first adrenaline rush has ended (which should happen about 4+2.5+1+20=27.5 seconds in to the fight assuming no lag and no SS glyph proccs).
Theoretical cons - You waste 1 second of uptime. You will not get the bonus from having devour and matrix restabilzed procced should you get a haste procc. Your recipient will get the bonus damage during the first 6 seconds, at a point where he will not be doing maximum dps.

Case 2: You use tricks a few seconds in to the fight, when your trinkets have procced.
Theoretical pros - Your buff has a 33% chance of being bigger. Your recipient of the bonus damage buff will get more use of the buff.
Theoretical cons - You will use a global cooldown which could have been used for a sinister strike. You will loose uptime of the buff equal to the time that your trinkets take to procc.

Math

Case 1:

Pros:

"You do not waste a global cooldown to use the tricks in the fight."
Now, how much is this global cooldown worth? All of the numbers here are averages taken from yesterdays Ragnaros fight.
A sinister strike hits for on average 13.4k.
This strike generates 1.3 combo points, which means 26% of an average Eviscerate, which is 35.7k*0.26=9,3k. (I am ignoring the energy costs, since I assumed from the start that we are energy capped during the relevant timeframe)
This strike also has a 37.14% chance of proccing instant poison, which is 4.1k*0.3714=1.5k.
This strike also has a 29% chance of proccing a main gauge attack, which is 8.0k*0.29=2.3k.
This strike will also procc burning wounds, which is an additional 13.4k*0.06=800 damage.
This means that an average Sinister Strike will hit for in total 27.3k.
That means that this wasted gcd costs 27.3k personal damage.

Cons:

"You waste 1 second of uptime."
The EP of the 4p is 439, and since 1 EP is a bit more than 1 dps, let's assume it's worth 450 EP. Thus, 1 second of lost uptime is 450 damage lost.

"You will not get the bonus from having devour and matrix restabilzed procced should you get a haste procc."
We can first of all completely discard matrix here, since the buff lasts for 30 seconds and will always be up for one tricks, no matter what.
Devour gives 1730 haste, and thus the buff would give you 432.5 extra haste, and since none of use are pretricking to ensure a 50% procchance, we can divide this number by 3 to get the average haste gained during the first tricks. That is 144.2 haste=245 EP = 250 DPS = 7.5k damage during the tricks duration.

"Your recipient will get the bonus damage during the first 6 seconds, at a point where he will not be doing maximum dps."
At this point I take the log from Neilpro, that I know got no tricks on rag yesterday, since I was giving Zamos.
During the first 6 seconds of the fight, he averaged 23k dps. Getting tricks here would give him an additional 13.8k damage.
Assuming one would build 2 combo points and then SnD, Neilpro would instead be 3 seconds in to the fight before getting tricks. At those 6 seconds, he averaged 28k dps. Getting tricks here would give him an additional 16.8k damage.
The damage loss for Neilpro from getting early tricks is then 3k damage.

Case 2:

Pros:

"Your buff has a 33% chance of being bigger."
We have already calculated that, the dps gain is 7.5k damage over the course of 30 seconds.

"Your recipient of the bonus damage buff will get more use of the buff."
We have already calculated that as well. The damage gain for our recipient is 3k damage.

Cons:

"You will use a global cooldown which could have been used for a sinister strike."
We have calculated this as well already. The damage loss is 27.3k damage.

"You will loose uptime of the buff equal to the time that your trinkets take to procc."
Let's assume that your trinket has procced when you get your first SnD up. It should have.
That means you delay your first tricks by 3 seconds. By using our earlier calculations, those 3 seconds of lost uptime is worth 1350 damage.

Conclusion

Case 1:
You gain 27.3k damage from the GCD.
You loose 450 damage from the uptime loss.
You loose 7.5k damage from the bad tricks.
You loose 3k of raid damage from the recipients bad tricks.
Net gain: 16.35k

Case 2:
You gain 7.5k damage from the better tricks.
You gain 3k raid dps from the recipients better tricks.
You loose 27.3k from the wasted GCD.
You loose 1350 damage from the lower uptime.
Net gain: -18,15k

All in all, case 1 wins over case 2 with a total of 34.5k raid damage.

I would like to point out that I have not in these calculations included the absolute damage you get from the tricks buff, nor the absolute damage your recipient gets. Adding those factors of course pushes the negative value of case 2 to positive, so that neither of the strategies are an absolute dps loss.

The thing about energy is that you are in essence energycapped even if your energy is not at 100.
As you should know being energycapped is not in itself a dps loss. What you loose dps from is the missed opportunity, the missed chance of actually performing a strike. And when you are in a streak of using every GCD that will eventually end in having 100 energy, you are in essence performing your rotation in a way that is by no means dependent on the actual energycost of a move. That means you could just as well have 10 energy after each strike as you could have 100. The energy wont matter until you at some point will have to wait until you can perform your next strike.
In this particual case, you will be using every single GCD until a few seconds after your AR ends. And since you will at some point during that AR hit 100 energy, the energy you had during the first seconds of the fight are irrelevant, since you are performing strikes that are in no way connected to the amount of energy they consume.

TL;DR: If you are using every CGD and eventually hit 100 energy, you will in essence have been energycapped during that entire time, even if your actual energy has not been at 100.

nextormento · 11/20/11, 8:07 PM

Minimal gear differences will have hugely different answers. Furthermore, using tricks on another higher damage (and lower ramp-up) non-rogue would *very* possibly yield a significantly higher output than any of the optimizations you're running. Also, micromanagement over an already micromanaged time-frame is, well, not very profitable.

When computing the damage output we do in a fight we have it easier, because each and every combination of procs/cps/energy happens enough times that we can average them out and come up with answers. There is no way to average an oppening sequence for it happens only once per fight, so the only thing we can do is average it with every fight. In fact I'd question the merits of such analysis: if I had to hazzard a guess, I'd say an average high-end raider won't pull the same boss (under roughly the same circumstances, gear-wise and raid composition wise) more than five hundreth times; 500 samples is enough to draw some conclusions, but certainly not big enough to be statisticaly significant as to figure anything game-changing (for comparison, take a default SimulationCraft output: it runs 10k simulations and yet it renders some uncertainty).

The granularity of the question begs for no asumptions whatsoever; thus, omiting SS procs and asuming energy cap is very risky; using average stat weights introduces very big errors in the computation: the EP given are for the whole fight, but its value fluctuates dramaticaly all along the combat: for this particular instance, the value of any positive proc/buff is significantly higher (since it's used along with very powerful cooldowns).

Now, in abstract you can follow two different ways to study this matter: one being the highest damage, and the other being lowest opportunity cost (a combination of the two is the only way to get to the correct answer); in-game they're usually mutually exclusive: if you study highest damage, you won't know how much energy could be lost, and if you maximize opportunity cost (energy/cp expenditure), you won't know how much more damage could have been achieved (since some of it would be ovekill thanks to, again, resource capping). That's why you need to study highest probable damage instead of highest damage, and/or lowest probable opportunity cost. However, to do so you need to know the whole sample space and asign probabilities to each event; given the buffs involved (weapon procs, gear sets, combat potency, etc) we may very well be talking about several hudreaths of millions of combinations so (/dramatization, but not too much), it certainly lies within the realm of what computers are there for. Now, the kicker is, as of now, we only have SimulationCraft for this kind of task, and so far I've never seen it decide for a clear winner in any oppener sequence argument.

Tumblebeer · 11/21/11, 7:33 AM

I think you may be overcomplicating a really simple issue.
First of all, we are working under the assumption that in the opening sequence a rogue will be global cooldown capped up until the point where your first Adrenaline Rush ends. Also, it is assumed that you will at some point during this timeframe hit 100 energy.

If this is true (which it should be, even assuming worst possible RNG) you are basically replacing tricks of the trade with a sinister strike if you choose to use said tricks the second before the fight starts.
What the original post is trying to establish is whether this sinister strike is worth more damage than the larger benefit the raid gains from having the tricks of the trade used 4-5 seconds in to the fight.

The argument held amongst our rogues is from one point met with above napkin math (with several further explanations that were deemed unneccecary for this community) and from the other point met with the logical assumption that the raid gains more damage from giving tricks to another rogue while he is in green insight and KS.
We have yet to produce any mathematical evidence for this point of the argument, yet the above math is not accepted as enough evidence for this side to accept it.

Edit: I just realised that Xerofour described both sides of the argument very clearly in his first post, and that this post might have been an unneccecary repetition of what he said. Yet I still believe that a decent estimation could be produced without interating this tens of thousands of times.

Kroyfel · 11/21/11, 1:52 PM

Shouldn't the first tricks be on a hunter anyway? Rogues don't seem to have anywhere close to the burst of some other classes at the moment.

Naihan · 11/21/11, 3:14 PM

Originally Posted by Kroyfel

Shouldn't the first tricks be on a hunter anyway? Rogues don't seem to have anywhere close to the burst of some other classes at the moment.

You bring up a pretty good point: if people are going to split hairs like this to gain a non-guaranteed, arguably trivial increase in raid damage, then you probably shouldn't be "trading" tricks with another rogue to begin with. If your rogues are constantly out DPSing your Dragonwrath wielders, then there's a bigger problem in your raid than when to properly time tricks.

Cynz · 11/22/11, 1:26 AM

Hi,

What is the best time to use the racial troll? (Combat spec)
I use it for the adrenaline rush with the 30% BG. But I do not know if this is the most optimized.

Naihan · 11/22/11, 3:50 AM

Originally Posted by Cynz

Hi,

What is the best time to use the racial troll? (Combat spec)
I use it for the adrenaline rush with the 30% BG. But I do not know if this is the most optimized.

You should absolutely avoid using this during AR; your energy-capping will be through the roof, at least that's what I experience. If you're energy-capping during KS, don't use it during KS. Every time I use it during KS, I energy-cap badly. I find that using it during bloodlust/heroism/timewarp doesn't lead to energy-capping for me, but this may be different for you. Other than that, you can use it on cooldown, or time it with buffs.

Shinryu · 11/23/11, 7:51 PM

Hopefully a simple enough question, are there any good ovale scripts for rogue besides the rather icky default ones? I noticed feral druids have a surprisingly good looking one, is there anything like that for (preferably) mut or combat rogues?

PikaPika1006 · 11/24/11, 1:49 AM

I played around with writing a rogue script in the past for Ovale. Only spec I spent a decent amount of time working on was combat, but the other two are functional enough to help someone hit the ground running on an alt or similar.
None of them have cooldown use suggestions, as I find those get in the way more than help with these kinds of addons.

Ovale.zip

Himmel · 11/26/11, 2:04 AM

I'm wondering is there a more convenient way of switching from Mutilate to Backstab once boss is on low HP than using a different button?

saedo · 11/26/11, 3:19 AM

Originally Posted by Himmel

I'm wondering is there a more convenient way of switching from Mutilate to Backstab once boss is on low HP than using a different button?

I suppose you can use a modifier macro. Something like...

/cast [mod:ctrl] Backstab; Mutilate

Furtim · 11/26/11, 3:26 AM

If by more convenient you mean have it switch automatically to backstab when you hit 35% while spamming the same button, then no, there is no legitimate way to do it.

Tumblebeer · 11/26/11, 8:04 AM

Well, technically you can. Have your backstab button and your mutilate button on separate barsets, using for example Bartender, then bind the paging between these two bars to a keybind. When the boss hits 35%, you press your keybind to page the bars, and you can keep spamming the same button for the desired effect.
I would not however recommend this, since you will not be able to easily swap between using mutilate and backstab should you end up standing in front of the boss.

bigredpaul · 11/27/11, 8:15 PM

My toon's professions are mining and skinning, but it seems like everyone thinks LW and JC are the best professions. Should I really switch over?

Velgore · 11/27/11, 8:53 PM

Short answer: Yes!

JC gives you the three epic Gems that can grant you a substancial ammount of Agility as is the same with LW with the wrist enchant. Stacking these two professions is a wise decision.

Skinning gives you a Crit increase... one of the worst combat ratings and Mining gives you Health... which does nothing for your damage output!

However, only switch over if you are serious about getting that extra DPS out for raids.... otherwise keep them two professions for making gold.

Hint: if you do decide to switch your professions, farm all the materials you will need with your current gatering professions.... it will save you money... but maybe not time!

/Vel