See I'm not sure I agree with the 1dps for 10 stam thing. For me, it was incredibly easy to get to 10k hp+ buffed. I can't see a real reason to have much more than that, and I am already approaching 11k as I acquire more T6. Sure, if you are under 10k hp, I could see a reason for skipping out on some dps and going with more health; but for any gear level above t4 health seems to be pretty easy to come by.

For me, extra stam at this point isn't even factored into the equation. As long as you have enough to survive any quick combination of abilities(IE najentus's globe explosion + geyser) you don't really need more. I mean, bosses hit for 15k or more on me, so its not really realistic to try and survive that type of attack. From my experience 10k seems to be a decent minimum for hp, and anything above that is just icing.

I think we are also having two different values for variables in our arguments here. I am arguing from the prospective of a late t5(ie someone with all ssc/tk gear) and t6 raiding rogue; others seem to be approaching this as a karazhan/gruul geared player. As a kara geared player with 8-9k health, stam is going to be more important. However, as you quickly close in on having "enough" health it becomes the case that sacrificing 5stam for 10atk power becomes very viable.

This is the thing: it really doesn't have to do with survivability being "meaningless." It has to do with the fact that Cheat Death's contribution is negligible. In most cases of things that can one-shot you (e.g. Cleaves), the damage is completely trivial to avoid and any rogue speccing for the express purpose of surviving those cases where he fails to avoid it is doing a disservice to his raid simply by being there. In most other cases of rogue death, it's because of a lot of incremental damage that adds up and that your healers should have plenty of opportunity to heal. If you are routinely dying to damage like that, you should more likely be looking at your healers rather than gutting your own DPS just to stay alive.

Additionally, I disagree on the very premise that Cheating Death is in and of itself a useful thing. I would argue that a large number of the times where a rogue dies in combat occur during wipes. Cheat Death is not going to help the raid at all in that case, since it's already a wipe. If the rogue dies during something that's not a wipe -- would Cheat Death really be terribly useful there either? There are two possibilities:

- Only the rogue dies, and the attempt proceeds to the end with 24 people still alive. Is an attempt like this really going to fail just because of the one rogue being dead for maybe half of the fight? Obviously on something like Archimonde, yes, the attempt will likely fail. Maybe on Vashj or Kael'thas, too. But the DPS requirements on these fights are also pretty high. You can control whether you die or not, but you can't make a spec do more than its theoretical DPS (allowing for variance due to random number generation).

- The rogue dies together with, before, or after one or more other people. Is an attempt like this really going to succeed just because of the rogue being alive rather than dead for that remaining duration of the fight? On most fights, probably not, particularly if the people who are dying were doing something other than DPS. Again, should you be addressing this by looking at your raid's strategy or your other raid members, or by gutting your own DPS so you can stay alive?

1) The specific numbers are to some extent irrelevant; if you don't like 1 DPS for 10 stamina, make it 1 DPS for 10000 stamina. The point being: there is some amount of stamina that it's worth trading 1 DPS for. And, analagously, there's some amount of damage that's worth sacrificing for Cheat Death.

2) While 10k buffed is certainly a nice start, a little extra padding never hurt. There are plenty of fights where you take unavoidable damage in sufficient quantity that healing is often needed to survive, for instance: Illidan Phase 2, Illidari Coucil, Reliquary phases 2 and 3, and Teron in Black Temple, and Azgalor and Rage in Hyjal. Is 10k enough to surive? Generally. Is 11k better? Yes. It provides that much more cushion for your healers, and allows you to survive even if your healers are having an off day. Would I dump 100 DPS for than extra 1k HP? Certainly not. But 10 DPS for 1000 hit points? That one seems pretty appealing. And 1 DPS for 1000 hit points? I'd take that in an instant.

3) Even if you assert that at some point incremental stamina becomes useless (which I don't agree with, but I can see why some would argue it), it doesn't change the fact that all stamina up until that point is still useful. If you had the option to swap out every piece of your gear for equipment that had twice as much DPS stats but no stamina, would you trade out every single piece of gear? No; you'd still keep some minimum amount of stamina. It's sort of like the hit cap. Just because you're at the hit cap doesn't mean hit is useless; the first 363 points of hit are all still quite valuable. It's just that further hit doesn't continue to benefit you. It becomes a question of optimization - "what's the most damage I can do and stay at or above the stamina threshhold I want to have" - which is correctly done by assigning a value to stamina and seeing which DPS gains cost you the least stamina. So even in this case, stamina does have nonzero value for purposes if optimization.

Well, my counterargument would be this: how many of you can honestly say that you haven't died in a raid except during a wipe (that you didn't cause) in the past month? There may be a few of you, but I think on the whole, it's safe to say that most of us randomly die at least every once in a while. And my point is: with Cheat Death, you're significantly less likely to suffer one of these random deaths. If the heal if half a second late in arriving - without cheat death, you die; with it, you live.

Now, there's the question of "how much does it actually matter if you, by yourself, die?" I mean, a lot of these fights are 24-manable, and many guilds can even survive healing through a Soul Charge or whatever. So are you going to wipe the raid by dying? Probably not, no. But that's not really the right question, I would argue; the right question is "how much am I contributing to the raid?". If you're a corpse on the ground, the answer is "you're not contributing anything". So the question I would ask is: is it better to contribute 100% in 95% of fights and 0% in the other 5%, or 95% to every fight? Well, I can certainly see arguments either way. But I would at least say that there is value in considering the second. Now, in practice, we're not in that case - it's a choice of 100% in 98% of fights vs 95% in 99.5% of fights (or whatever) - which is significantly less worth it. But the theory is valid.

The cases where I can see Cheat Death being very useful are on fights like Vashj where adds can potentially spawn on you and one-shot you instantly. That's been one of the main causes of our Vashj wipes. A 15k melee hit that your healers have no chance of healing and your tanks have no chance of intervening on simply cannot be planned for.

Cheat death is not a viable means to increase your dps, and it is obvious because of the way the ability works. Lets look at a scenario where 2 rogues are dpsing 1 has cheat death 1 doesnt have cheat death.

Illidari council, two rogues dpsing the paladin (Gathios) concecrate + blizzard on the melee.

Rogue 1 (combat) promptly gets out of the aoe for 5 seconds and takes 2000 damage, while doing no dps but rupture ticks that are still on gathios.

Rogue 2 (cheat death) Stays in through the aoe to get 5 more seconds of dps in, gets his HP reduced to 1000~

Now, which is better.. well rogue 1 can bandage to full from the minimal damage he took. Whereas rogue 2 puts so much strain on the healers that shouldn't normally be there.

More strain on healers = people die. Isn't this the opposite of what cheat death is supposed to accomplish? And this is assuming that he didn't die from the aoe. 11k buffed, aoes tick for around 2k and hes in 2 of them. that means he would be at 1000hp in 3 seconds before cheat death procced. That is still dangerously low, and in a fight where it is healing intensive cheat death might not even save you if you are sitting in aoes.

Sorry if I am rambling.. I just don't see how you can justify using cheat death to cause more strain on healers.

Not really true.

Nightbane has the same exact number of transitions regardless of the dps. Has the same number of spawns.

If you take it slow -- when learning -- and don't try to play "max the dps" you have fewer chances to pull aggro, fewer chances to have people roast in the charred earth and drop death for no reason. Yeah, it's a bit more taxing on the healers, but it's a lot better to control it than to race through it.

I'm sure other examples exist. In fact, Magtheridon after the channelers is arguably one of them. Hydross once you learned it is hardly a dps race. Vashj Phase 2 is. Phase 1 of course is not. I mean it's a semantic point. But dps races are the ones where you go competely all and slam against the aggro cap whenever possible (e.g. Leotheras demon phase). Other fights are much more about keeping steady dps up and not mucking things up too badly in the process.

This make zero sense. Yes, Nightbane has the same number of transitions, but more DPS means those transitions happen faster, enough DPS means you get no fears, more DPS means the skeletons die faster and there's less risk of them gibbing a healer (although this is very unlikely anymore considering how badly the fight has been nerfed).

On Magtheridon, more DPS means less banishings, which means less cube clicking, which means less chance for anyone to screw it up (although this is probably less prone to fuckups since the 2.3 nerf).

Hydross is a DPS race; you have an enrage timer. More DPS means less transitions, which means less chances for people to screw up and pull him back across the line, more DPS means the adds are up for a shorter period of time, lessening the chances of one of them breaking loose and gibbing DPS or healers, and also allowing more DPS time on Hydross himself.

Even Vashj phase 1 is sort of a DPS race, as the faster you get her to 70%, the less charges she can place on people, which is less healing healers have to do.

Basically, every fight in the game is made easier the higher your DPS is. Every fight is a balls to the wall max-DPS fight, even when it's a control/execution fight or when healing is heavily emphasized.

If the attitude is that it's OK for the rogue to die and the other 24 players to pick up the slack, because it doesn't really make a difference I don't really see how you can also say that a spec that does 10% less dps but has greater survivability sucks.

I think I generally agree with the conclusion that cheat death is a sub-optimal raid spec, but I find many of the arguments on this board to be specious and ignorant of reality.

It is all well and good to say that cheat death is useless because (almost) any time it would save the rogue, the rogue could have saved himself through perfect play, but in the real world mistakes happen and more importantly most rogues can't play perfectly all the time. If a tactically poor player player can make a strategic decision to compensate for their weakness, it will help them overall.

In the meantime, they should be working on their situational awareness so they can be more effective, but that does not change the fact that for that player it makes sense to spec into cheat death.

That said, a player that is raiding black temple and mount hyjal should probably have the situational awareness (and necessary raid support) to not need cheat death, otherwise they will be a burden to their raid group.

The point I'm driving at is that cheat death is a talent whose worth is more or less inversely proportional to situational awareness (skill). In a perfectly ideal world, cheat death is a useless talent. In the real world, a rogue needs to honestly appraise their skills and the fights that are progression, and decide whether or not cheat death is a worthwhile crutch to help them while they improve their game, or waste of talent points.

Few points - I agree with Aldriana in theory, Cheat Death can (and will, if the situation arises) be accounted for with DPS - it's just that there is not a significantly large number of these fights to consider the talent to be anything other than a 'trick'. Vulajin's right in his statement that most of the cases in the T4->T6 raiding that I can think of, if Cheat Death saves you from something, you're VERY likely to die very soon regardless, limiting the usefulness of the talent.

Consider a theoretical Magtheridon/Prince like boss that once a minute did a blast nova that did 25k+ damage to everyone in 30 yards, in which the entire raid had to run out of range or get instagibbed. (Actually Al'ar's little explosion on P1 death is ... sorta like this, just lesser in scale) - in this case we could very well conclude that Cheat Death had a very specific DPS number for the fight - 60 yards of running once a minute, etc, etc. (Well, except of course that in that specific example we'd just cloak it off, BUT YEAH .. if it was "Cleave Nova" or something.)

I listed half a dozen fights in T6 content where randomly targeted abilities, AoE, and general raid damage are sufficient to kill you without healing, such that stamina is useful. Cheat death is equally useful in that scenario. Clearly it's better to get healed than need Cheat Death - but that doesn't always happen. And given that that's almost half the fights in T6, I think it goes above "situationally useful". The fact that it doesn't go above it by enough to be worth the tradeoff notwithstanding, there's still value in it.

Obviously, you want to avoid mucking anything up along the way. But assuming you can increase your DPS without mucking things up, that's pretty much always better. Nightbane? Sure, the adds come by life rather than by time. But you still want to burn down the adds quickly, and having fewer fears to survive between transitions certainly makes the fight easier. Hydross? Pretty much every wipe I've ever seen is a messup on transitions - if you kill him in 5 transitions instead of 10, that's half as many chances to botch a transition. Even one of the ultimate examples of a control fight - C'thun phase 2 - if you could kill him in two weakens instead of four, you can absorb a lot more attrition on the way. The first time my guild killed C'thun, he died about half a second before the end of the second weakened, and we would not have survived a third. Had our DPS been even 1% lower, we would not have gotten the kill on that attempt.

So, yes, there are certainly "control fights" where the majority of the challenge is not letting things get out of control. But even in these fights, once you've learned to control them, more DPS still does make them easier. And in particular, as a rogue you don't always have a lot to do with keeping the fight under control; for instance, there's nothing you can do to help the tanks make the transitions more easily on Hydross. As long as you hold fire for a second or two around the transition to not mess it up yourself, your only other job to do is kill stuff as fast as humanly possible - and by doing so, you make everyone else's life easier. So while your first mandate is "don't screw up", your second is "burn as hard as you can" - and that applies to just about every fight there is. So there really is a fair amount of truth in saying "every fight is a DPS race" - because from a rogue's perspective, it's largely true.

Of course more dps makes fights easier. If you fail to get the difference between Hydross and Leotheras with regard to dps races, I cannot help you. Once you kill Hydross, you will find minutes left on the enrage timer. We now have about 5 minutes left on it. On Leotheras, uh no.

With regard to Nightbane of course the skeletons are a dps race. And of course the ground phases are not. I cannot help people who don't get the difference.

My guild kills both Leo and Hydross in under 6 minutes. So I'm not sure I catch your point here.

They are DPS races of different degrees. They are still both DPS races. From the perspective of a rogue, on 75% (or more) of fights, you only have one job: Kill stuff as quickly as possible. There is always also the job of "don't break control and wipe the raid", and the relative importance of the two components certainly does vary. But there are very few fights where both factors don't play a role. Even on raw DPS fights, you can screw up and wipe the raid. Even on primarily control fights, you need good DPS. I'm not saying I don't see a difference between Nightbane ground and Nightbane air, or Hydross and Leotharas; I'm just saying that the difference is one of balance between the two main jobs of a rogue, and not a situation wherein something is required for one and totally absent in the other.

I really enjoy reading this thread, and considering the ideas and numbers people provide. Thank you everyone for contributing to this thread.

Call me a black sheep, but I don't think this is a path we will want to travel. We are the only class left with one , and only one role to play in raids. Everyone else can fill many spots in one form or another. We are the only ones left bringing plain vanilla dps to raids.

I've had my days in the sun as #1 on the charts being boring swords. Why would we want to keep this race? Is it really worth staying #1 at the cost of being forced into 1 or at best 2 specs that bring very few tricks, and use the same attack pattern day in and day out (the only change being the target your after?).

While I can understand the basic, concept of, ' you rogue, you bring dps', fighting techniques such as cheat death and shadowstep give us some flare, and change the fight, fighting against these specs only hinder our ability to ever get past the combat tree for much of anything.

A few posts back I saw a guild leader say something to the effect of 'you spec shadowstep you lose raid and get guildkicked'. Take a step back, and relax. It's just a game. What happened to having fun?

I'm lucky enough to have found a raid that doesn't care what I spec, and I have some room to move around in, I hope many of us can get to that point someday, instead of having to be combat to be considered a 'good' rogue.

When a fight comes along that requires me to poof around or take a 200k hit every 61 seconds then sure wtf Ill go sstep. As it stands now, sub offers no really definitive advantage and does less dps. There is no good reason to use it. I'm all for expanding the role of a rogue but "Not quite so much dps" isn't really helping anyone.

speak (or rather, type) for yourself.

i've browsed very carefully through the last 20 or so pages & have been unable to find anything that did not reference a dps sheet w/ regards to the comparative effectiveness of trispec vs cs & vice-versa post 2.32. i'm here to look for additional information so i am looking to change my mind.

i do note, however, that your previous reply to my post consisted of:

* an off-hand disparaging comment about capitalization (which is obviously relevant to the discussion)
* a non-sequitur about trispec being considered viable due to people anticipating hemo charges fix (if that is so, then why the hell go trispec in 2.3 as opposed to when the bugfix rolls through?)
* an accusation that my math has been "poorly thought out" w/o any details as to why, exactly, that is

i've asked for thoughts regarding trispec viability not contingent on a dps spreadsheet. if you or others are either unable or unwilling to provide that, that is fine--but you should acknowledge that fact rather than brush it aside.

p.s. i also note that a forum mod has notified me of an "infraction" referring to my lack of capitalization. lollerskates, i suppose i won't be posting anymore after this XD

So, here's the thing. The spreadsheets represent the most detailed, most accurate calculations that the combined theorycrafting community has been able to come up with, based on the most detailed analysis and most up-to-date theories on rogue mechanics. Disdain spreadsheets all you like, but they really are by far the most accurate theorycrafting tool we have. No amount of casual theorycrafting will match the hours of effort that have been invested in them.

Really, if you disdain the spreadsheets, your only real hope of doing better is a systematic study of hundreds of fights with identical gear. I'm not talking anecdotal "I did 1100 DPS last week and 1150 this week" - I'm talking hundreds if not thousands of fights. And since no one has done that yet, if you don't like the spreadsheets, I think you're sort of out of luck in terms of quantitative theorycrafting, leaving only the realms of speculation - which isn't exactly the most useful topic of conversation.

So, fundamentally, I guess my question to you would be: if you don't want information based on the most accurate theorycraft available, what are you looking for instead?

not sure why you're assuming that i actively 'disdain' spreadsheets--i don't, nor did i state otherwise. i've asked for thoughts on a very specific issue (2.3 ==> 2.32 hemo changes) that do not refer to any spreadsheets for purposes of independent verification. if "the community" finds that somehow beyond the pale, so much worse for "the community."

there's the assertion that 'casual' theorycrafting, which exists apart & aside spreadsheets, is necessarily inferior/worthless. that's nonsense of the highest grade, & it surprises me to encounter it on ej. when doing relatively basic math, either:

#1. the assumptions one is working with are reasonable, or they are not, &
#2. the equations one is using are correct, or they are not

the whole thing is binary, regardless of "the hours of effort that have been invested." you can have beautiful math that has been worked out in 10 seconds that is correct, and complete rubbish that has been hatched over a year--and vice-versa.

i've posted a concise but explicit summary of why basic math, using what i consider to be extremely reasonable assumptions, points to a fact which popular wisdom disagrees with (i believe that in terms of net raid dps, 2.32 will be a buff to trispec compared to 2.3.) if you find that either my assumptions or my math are wrong, please state so--or, alternatively, say, "this is not worth my time, gtfo." but acknowledge the fact that you are either unwilling or unable to critique what otherwise amounts to extremely simple math.

which is of course quite ok with me, it just means i'm likely right =D

Well, if you say "I'm unhappy with the last 20 pages of analysis since they're all spreadsheet-based" implies that you think that there's something wrong with the spreadsheets. If not, so be it, and I apologize for reading into your statement.

Now, the issue with the "simple math" thing is that rogue DPS is anything but simple. There are tons of subtleties and details that are necessarily omitted by simple analysis of the problem; to get truly accurate answers you need to include all these, and the typical way to do so is via the spreadsheets.

So in terms of debunking your math and assumptions specifically:

1) 2k AP is probably not a valid estimate for any rogue in 25-man raid content.
2) Even if 2k AP and 30% crit was valid for some rogues, it wouldn't be valid for all rogues; measuring the validity of a spec based on it's value for a rogue with one particular set of stats is somewhat shortsighted.
3) There's an awful lot of rounding and estimation involved in that 2% personal DPS loss figure; I question the actual precision of it. It wouldn't surprise me if the "actual" answer (subject to the other assumptions) is anywhere between 1.5 and 2.5%.
4) I don't recall anyone ever claiming that Hemo was viable with 1/4 of the charges wasted; I think most assessments of it's viability were based on full charge consumption. So, you're right, Hemo might not actually be viable in 2.3, either
5) So, if we compare the theoretical viability of how Hemo is supposed to work in 2.3 (which is viable) to the viability in 2.3.2, we find a loss of between 1.5 and 2.5% of the rogue's DPS... which you estimate as 1500, but in reality could easily be anywhere from 1200 to 2000. So we're losing somewhere between 18 and 50 DPS, and gaining 6 damage per charge on 10 charges per 3.5 seconds (modified by armor, crit chance, etc. etc. and thus not necessarily anwhere near the 60 damage per 3.5 seconds one might naively expect) - but we'll count it as the 17 DPS that that would imply. Now, that would mean that our net DPS is down by some intedeterminate amount between 1 and 30 DPS, depending on the accuracy of the other assumptions, meaning that triSpec hemo is, in fact, somewhat less viable than the theoretically viable 2.3 build that people have been endorsing.

...I'm not sure how useful that conclusion actually is. So allow me to prevent my alternative assessment of the situation:

According to the 0.9.1 version of the Gear Spreadsheet, my theoretical DPS as combat swords with typical raid buffs is 1804. If I adjust my build to 11/28/22 tri-spec Hemo and make the gear adjustments to accomodate the new spec (i.e., use items that become better with the new spec that I have on hand), the highest DPS I achieve is 1694, a loss of roughly 110 DPS. Hence, in order for tri-spec Hemo to be worthwhile for me, I must generate at least 110 DPS from the debuff. With a calculated average of one Hemo every 3.60 seconds, this would mean that at least 396 damage out of the 420 possible must be realized. Now, with the spec in question, my actual expected damage (factoring in crits and armor) is about .993 times the raw damage; if we assume this is typical, this means that the actual damage potential of my debuffs is only 417, rather than 420. Hence, so long as at least 95% of my Hemo debuffs are actually utilized, I will experience a (slight) damage increase.

Now, on single-target boss fights, I will generally achieve this 95% number, and hence might benefit slightly from speccing Hemo; however, on trash or fights with multiple DPS targets, there is a reasonable chance I may not. Hence, while it might just barely make sense from a damage consideration, I do not consider it worth the risk for myself. Based on my own personal analysis, I conclude that for T6 rogues such as myself, tri-spec hemo is probably only marginally worthwhile in the best of cases, and that it's certainly the case that it's no longer viable to have more than one Hemo rogue in the raid.

To me, this second assessment - one that every rogue can easily do for themselves with the help of the spreadsheet - is rather more useful than the first. But feel free to refer to whichever is more comfortable for you.

I doubt that you were killing him in under 6 minutes before you farmed BT/Hyjal to get your 5/5 Slayer gear...

--edit--

To make sure there's some solid numbers in the post, I just thought I'd point out that you also say you need to ensure 396 out of 420 possible damage is used by the Hemo debuff. In actual fact you're looking more at something like 396 out of 525, since the extra damage can crit and the average critrate for physical damage classes is likely to be at least 25% and probably even higher if you're in BT/Hyjal gear.

But then again it is also affected by armor, which will sit be at most 4,500 or 3,000 (for bosses with base armor of 7700 or 6200) with just Sunder Armor and Faerie Fire up. At most this will give you a reduction of at most 27.3% or 20.1%. That takes the added bonus from the Hemo debuff to 381 or 420 per application. In fact Curse of Recklessness is likely to be up, giving another 800 armor reduction, taking the Hemo debuff damage to 401 or 443 per application before passive armor pen or procs.

In short, the Hemo debuff is often worth even less than you thought, but is still clearly enough to compensate for your personal dps loss, given that all the charges are being used. To be honest though, I do think that there are enough fights where you're not just sitting on a stationary target (Hydross, Vashj, Lurker to name a few that I'm personally encountering) where it's really not worth taking Hemo because you'll be wasting charges left right and centre. Hemo at Vashj is particularly wasteful, especially if you're on elemental duty. You may as well do as much overall damage yourself, to ensure that your job gets done quicker in those situations.

I used to be a big fan of trispec and Hemo in general, but I'm slowly coming round to the idea that it really needs a bit of a buff to the hemo debuff, either in number of charges or simply in the extra damage it deals (preferably through the deep Subtlety talents) for it to be worth universally losing personal dps.

Because people didn't know about the bug. It appeared in the 2.3.2 patchnotes and everyone said "Ah, so that's why the buff's been disappearing so fast!" It's impossible to actually measure how much damage the debuff contributes, so all we have is the mathematics, which of course doesn't account for bugs. What this means is that since 2.3 came out, many people have been using tri-spec even though it's an actual DPS loss for them relative to combat (because of the bug). When the bug was discovered, a lot of us specced back out of tri-spec. And when the changes to Hemo were revealed, we stayed out.

When 2.3.2 comes out, the bugfix will in fact increase tri-spec damage somewhat, however the best available information is that it will still not be significantly better than combat for high-end raiding. In low-end raiding (~Karazhan gear level), tri-spec will be the best spec. However, be aware that the superiority of tri-spec depends critically on using all 10 charges every time. This is likely in a 25-man raid, but not in a 10-man raid. So in fact the niche where tri-spec is superior is restricted to 25-man raids with Karazhan-level gear - i.e. groups just starting Gruul/Magtheridon/Lurker/VR.

If that's you, by all means go tri-spec.

I guess I'm kind of done trying to explain this. It's so completely obvious to me what constitutes a dps race and what constitutes the need for stready dps that I'm apparently having a hard time explaining it. Or perhaps you are confusing my comments with a suggestion they modify the task of playing a rogue. They don't. But that really isn't why I made the comments, as they came into this already derailed thread to discuss dps races, per so, not rogue-ing per se. The fact is I do the same thing on Hydross and Nightbane. It's just a matter of how many seconds it takes to get it going and how often I have to stop it.

But rogues are perhaps the poorest judges of dps-race fights vs. non-dps-race fights. From the mob's rear end, every fight is dps, don't pull aggro. They all seem to be about dps maximization. On some fights, the whole raid is joining you in that, on other fights it isn't. If >>raid dps<< falls in Vashj phase 2, it's pretty uniikely Vashj is killed. If >>raid dps<< falls in Magtheridon post-channelers, you can probably sustain that fight indefinitely. One is a dps race, one is not.

Disagree, agree, tell me I know nothing about WoW, I can't really be bothered to make the point anymore.