Scattered Shots: Hunter spec and class balance

Last week, Ghostcrawler, WoW's lead systems designer, stepped in with a post to let the players know what Blizzard was looking at in terms of class balance. Hunters had the privilege of being the only class that had all three specs being mentioned in the post. He acknowledged that BM and MM DPS was too low and were in need of buffs. He also pointed out that SV single-target damage looked too high (along with DKs) and that AoE damage might be too high, as well.
Responses from the hunter community have been pretty divided. A lot of hunters are saying no, SV is not too strong, while a lot of others agree that they are a bit too high. Still others argue that hey, someone has to be on top -- why not us? And then there's the crowd that says it's too soon to tell. But everyone agrees that BM and MM need the buffing.
Today, we're going to take a look a hunter balance. We'll delve into raid parse data to see if we can find out just what Blizzard is seeing, we'll talk about hunter spec balance, and finally, we'll talk a bit about how current bugs in the game are affecting our DPS. Join me after the cut for an in-depth rundown of hunter class balance at the beginning of Cataclysm. Oh, and even if you're not a hunter, you're gonna want to see this.
What Ghostcrawler said
Before we get into specifics, let's have a quick review of exactly what Ghostcrawler said, pulling out specifically the parts that pertain to hunters:
We're happy with damage overall. We have very few traditional tank and spank fights (even Argaloth likes to parry melee) so it's hard to get consistent numbers without very large data sets. Still, we see Survival hunters and Unholy DKs on top of a lot of single target fights. Arcane, Marksman, and Beastmaster damage is too low. ... On fights where there is a lot of area damage, Demonology warlocks, Frost DKs and possibly Survival hunters are all too high. ...
As part of the Marks and Beastmaster buffs, we're buffing Aimed Shot, Kill Shot, Chimera Shot, and Kill Command. ...
"GC, is this the final list of changes? Does this mean I can expect no changes for my class? Does this mean you don't care about me?"
No. This is some stuff we are looking at so that you'll have some context if you see changes on a future PTR. The final list of class patch notes for the next patch will doubtless be much longer.
As part of the Marks and Beastmaster buffs, we're buffing Aimed Shot, Kill Shot, Chimera Shot, and Kill Command. ...
"GC, is this the final list of changes? Does this mean I can expect no changes for my class? Does this mean you don't care about me?"
No. This is some stuff we are looking at so that you'll have some context if you see changes on a future PTR. The final list of class patch notes for the next patch will doubtless be much longer.
Let's start off by diving into some numbers ourselves, to see if what we can see matches at all with what Blizzard is saying. Of course, Blizzard is seeing far, far more information than we are. For us to get a feel for how specs are performing, our best resource is World of Logs, where there are already hundreds of raid parses, including all of the currently available bosses.
All of the data I'm about to present is pulled from World of Logs on 25-man normal fights, compiled as I write this on Dec. 30. I chose 25-man because we're more apt to have every raid buff there, whereas in 10-man, you could get skewed numbers right now because certain teams are missing buffs that are great for certain specs. However, since we're likely to see far, far more 10-man teams in Cataclysm, that's something we're going to want to keep an eye on in the future when we're into the tens of thousands of parses (and thus, balancing out the buff advantages).
I also specifically removed the Halfus Wyrmbreaker fight from this data, since it's a DPS gimmick fight and not a good representation (with 80k DPS common, as an example).
Not all specs have a good enough representation to provide reliable data; typically, the lower-performing specs rarely raid. So while there might be a good 300 SV hunters having done a certain boss, we might see only 20 BM hunters and 7 MM hunters. Thus, I have not included any specs whose representation was too low to be statistically significant; it doesn't necessarily mean that their DPS is very low, just that too few people are raiding with them to tell.
I've been monitoring all of this information and updating my spreadsheets weekly. While it's certainly early in the expansion yet, with only the best raiders completing all of this content regularly, it's interesting to note that the data has been very consistent from week to week. The numbers are climbing upward as people gear up, but the relative positions are remaining pretty fixed. Finally, I should not that hunters are performing much better in 10-mans than in 25-mans. Whether this is because we're less buff-dependent or we just are less dependent on as wide a variety of buffs, I'm not sure.
Okay, let's get to it. We're going to start by looking at top DPS representation.
For this, we're looking at the top 40 DPS benchmarks for each boss and seeing how many of those spots each class holds. For this chart, we're just looking at classes, not specs. For most classes, one spec was on top -- SV was the only hunter spec with any top benchmarks -- but other classes had two specs representing (and warlocks have three, if you count demonology totally owning every AoE fight).

What's happening here is that hunters, death knights, and warlocks are all doing very well on single-target fights, with hunters probably the best. On AoE fights, demonology warlocks completely destroy the top benchmarks, going so far as to claim 39 of the top 40 slots on one fight. Interestingly, SV hunters continue to do well on the AoE fights, though not nearly so well as the warlocks.
Of course, this is just showing us who is getting the very top results on various bosses. It's not telling us what the class representation is -- for example, maybe there are three times as many hunters as other classes, and that's contributing to our high DPS (it would make a difference, just not a huge one, as long as everyone has a statistically significant representation -- for example, you could have 10 times as many MM hunters as anything else, but they still wouldn't own any top spots).
The big thing this chart is not showing us is actual numbers. Who cares if hunters or warlocks are No. 1 if they're No. 1 by 10 DPS? After all, as many have pointed out, someone has to be on top -- why not us? It's only a problem if we're ahead of the pack by a large margin.
So let's look at some numbers.
The following chart takes the median of the top 20 benchmarks for each spec for every boss (except Wyrmbringer) and averages them to get an average top DPS. We're taking the median of the top 20 to get rid of some of the crazy outlier results, where the top result is 10k DPS higher than No. 2, but while still trying to represent the best the spec is capable of in the hands of the best-geared and best-skilled raiders currently out there. Specs with poor representation are not included.

Clearly SV is on top, and we're also on top by a lot.
To put this in some perspective, we can take a look at what was going on in ICC, during the 10% buff period (when I did a huge statistical analysis of hundreds of thousands of lines of raid parses). In ICC, every class (except the poor shadow priests) had a spec that was within 10% of the top spec. Right now, not even the median is within 10% of SV hunters. We're a bit over 13% ahead of the median (and the mean).
So yes, there's always going to be a spread, and some specs will be ahead and some behind -- but when the differences get this large, and if they continue this way as we get more results, then I agree that it's a problem and that SV needs a gentle nudge downward. After all, if I were a raid leader, I'd stack hunters, DKs, and locks as my only DPSers if I could. Well, and maybe bring a shadow priest for replenishment.
This leaves us with the "it's too soon to tell" argument. It's certainly possible that these numbers will radically change in the next few weeks and hunters will just be good, and not too good. But frankly, that's improbable. SV hunters in particular scale very well with gear, and over the past three weeks, though these numbers have increased, the positions have remained consistent.
Of course, if it is too soon to tell, then it's also too soon to say that SV and BM need buffs, right?
Hunter spec balance
While we don't have enough raid data to get a good feel for the BM and MM hunter specs, we're fortunate in that our class has very accurate qualitative tools for measuring these things -- at least for measuring them in a standstill fight. We can pop over to Femaledwarf.com and make some good comparisons between specs to get a feel for the difference.
BM and MM are very, very close to each other in theoretical standstill DPS, and both specs are about 15% behind SV.
But we also need to consider in general how much various fight mechanics will hurt the DPS of the specs to get a sense of how they'd perform in a real-world environment. Let's take a look at some pros and cons of the specs:
- Beast mastery BM hunters have excellent focus regeneration, and since they get a slightly larger percentage of their DPS from their pet's various auto-attacks, they have a nice advantage to their DPS in movement-heavy fights. On the other hand, they are particularly poor in target-switching fights, when compared to the other hunter specs. They have to wait for their pet to travel from target to target. Furthermore, fights with an air phase hurt them, as they lose their signature shot. BM also has the worst AoE DPS of all the specs -- they need a suboptimal DPS pet to enable their AoE, which lowers the damage of their pet and their signature shot considerably.
- Marksman MM hunters have the most difficult rotation to maintain optimally, with the constantly changing priority involved with maintaining the Improved Steady Shot buff combined with random Aimed Shot procs. The result is that complicated fight or movement mechanics can make it easy for that delicate rotation to fall apart, and losing the Improved Steady Shot buff hurts MM a lot. MM also has a fairly poor AoE.
- Survival SV hunters have it pretty good. SV doesn't have a lot of problems with target-switching fights -- less, in fact, than MM -- and in general has a far more forgiving rotation. If you screw up the SV rotation and delay an Explosive Shot or mess up your Lock & Load, the hit to your DPS isn't that large and your rotation doesn't crumble around you; you just pick up where you left off. In addition, SV has by leaps and bounds the best AoE of any hunter class -- better AoE than a MM and BM hunter combined.
Let me be clear here, before I get a swarm of angry SV hunters accusing me of hating SV. I like all the hunter specs, and I play whatever gives me the highest DPS. Furthermore, I think that if one spec has to be on top, it should be SV. They've earned it, after all -- MM had all of vanilla and the latter half of Wrath. BM had all of BC and the very beginning of Wrath. SV only had a small window of months being the top spec. They're due.
But I still want as much balance as we can get. I wan BM's target dummy DPS to be slightly higher than SV's -- then, in actual raids they'll be the same, or SV will be slightly higher. I think SV is due to be the top spec, but I also want hunters to be able to raid as any spec they want, rather than a Wrath situation in which BM underperformed so severely that you hurt your raid team -- or in BC, when we had the same situation with SV and MM.
The impact of hunter bugs
Something that I'm sure Blizzard is taking into account as it considers any possible hunter buffs or nerfs is some of the bugs currently in the game. We have a handful of known bugs remaining. A bunch of these are pesky pet-related bugs that are relatively small, and I can see how they aren't on the top of the bug to-do lists. But a couple of them impact our DPS and will increase our DPS when fixed. And they will boost SV's DPS more than any other spec.
The first is that our Serpent Sting and Black Arrow are still using the spell critical multiplier (150%) rather than the physical crit multiplier that all the rest of our shots use (200%). We are, by the way, pretty certain that this is actually a bug and not intentional, though it's certainly within the realm of possibility that Blizzard will choose to leave it be while SV is already so high on the charts.
Fixing this will slightly boost all hunter DPS, but SV's more than the others. SV does more damage with Serpent Sting because of its mastery. It's the only spec with Black Arrow, and it gets more benefit from those DoTs anyway with the Toxicology talent. Also, it's worth noting as an aside that SV gets more benefit from both of our tier 10 set bonuses than any other spec.
The second bug has to do with the 10% attack speed buff provided by the windfury totem or the SV hunter's Hunting Party. This buff is actually reducing our focus regen by 10%, rather than increasing it by 10% (the way that every other ranged attack speed buff does). Now this amounts to only about 0.4 focus per second -- perhaps an extra Arcane Shot every minute -- but that could really help make SV's rotation more fluid, removing those annoying choke points where there isn't quite enough focus to fire both Black Arrow and Explosive Shot on cooldown.
So this is where I'm seeing the hunter class balance at the moment and the factors that will contribute to it. What do you guys think? Too soon to tell? Do you have a different way of gathering World of Logs data that shows a significantly different result? Can you think of other factors that will affect our DPS that I haven't considered? How would you go about adjusting the DPS of the various hunter specs?
As a special favor to me, let's try to avoid saying anything like, "In my raid, the rogue always does more damage than me, so SV is clearly not too strong." That's anecdotal evidence and is neither compelling nor useful in this discussion.
Filed under: Hunter, (Hunter) Scattered Shots






Reader Comments (Page 1 of 6)
KrusherX Jan 3rd 2011 9:15AM
Do you ever rest? Do you have interns or something?
The amount of articles and data you produce weekly is impressive and we all appreciate your work.
Thank you, you crazy dwarf
Know Jan 4th 2011 4:21AM
Gotta agree with this comment. Us Warlocks wish that you had an evil twin, maybe then we'd get something that vaguely resembles a regular column!!
benbettis Jan 7th 2011 1:17PM
I know I'm a few days late on reading this, but I could not agree more. I'm seriously not a brown noser, but Frost, your articles are the best on this site hands down. I consider my hunter my second main and it kills me on the inside when I read some of this excellent material when the material for my main is all worthless fluff (all of your spec 101 articles were delightfully thorough, and this is coming from someone who scours Elitist Jerks before falling asleep).
I hope the people you work with don't hate you because we love you more... but we do. Just so you know.
Amenman Jan 3rd 2011 9:18AM
Honestly, I can feel the all encompassing nerf bat floating over our collective heads.
I have enjoyed Survival the most since creating my hunter. I used it during ICC while I was still picking up ArP items(and shortly later, they went bye bye) and I would hate to see it overnerfed like BM of old. One could only hope that if blizzard is going to be right for once on pure class balance, they get it right with the great hunters.
kingoomieiii Jan 3rd 2011 9:38AM
Well, they have more tools to balance with now- they can adjust one spec's Mastery bonuses without buffing up any other specs (though this does have the effect of changing Mastery Rating's worth for that spec).
Ronin Jan 3rd 2011 11:54AM
Hopefully they've learned their lesson and won't simply nerf SV to the ground. And as kingoomieiii said, they have more tools now for balancing out the specs. And, more importantly IMO, a different approach-- they seem to want to keep everyone viable.
I can understand your concern though-- Blizz has shown an incredible willingness to nerf Hunters at almost every step, in the past.
Clevins Jan 3rd 2011 3:05PM
The two issues I wonder about are
1) is SV damage where more DPS should be by design? That is, are the encounters designed for most DPS classes to be closer to what SV is doing or are they designed for a lower average dps and SV is significantly above the design goal and
2) is it really a good idea to base design changes on how top players play specs? You can say that top players will be expressing the potential of the spec best, but the question to me is what happens when a spec is played just well and not perfectly? For example, on my shadow priest in Wrath, dot clipping and uptime had a HUGE impact on DPS. People who were really quite good at not clipping dots while keeping a high uptime did very well but if you clipped even the last tick of some dots, your DPS suffered quite a bit... even though we're talking about differences in the tenths of a second range. I'd hate to see a spec have to be played close to perfectly in order to do well with anything less than that seeing disproportionate dropoff.
AS for 'too soon', well, I suppose that the argument is that with more gear lower dps specs/classes will improve more than the top classes... I'd hope Blizzard would have a handle on that, but then I'd have thought they would be able to simulate DPS closely in the first place and we have the above spread.
JT Jan 3rd 2011 9:21AM
I'd disagree with your statement that anecdotal evidence is neither compelling nor useful. After all, the information you're presenting is merely an aggregate of anecdotes.
If an individual hunter in a particular raid is being outdps'd by a rogue (as your chart suggests ought not to be happening) then at the very least his information could be used to explain situations in which the exception proves the rule.
Simply because you have MORE numbers does not mean your numbers are somehow better, and statistics are simply a means to an end, not an end in and of themselves.
Also note: seeing a chart that shows hunters at the top for a WoL of the top 25 man guilds (a population excluding the majority of this site's readers) does absolutely nothing for hunters just moving from normals to heroics, or from heroics to 10 mans (a population including the majority of this site's readers.)
At the very least, a thorough statistical analysis ought to have included 10 man parses, especially given how many guilds have switched from 25 man to 10 man.
All that to say: I'm not disagreeing that hunters aren't a tad strong at the moment; I'm simply noting you can't just run around yelling "My numbers are better than your numbers HAHA!" and expect anyone to take that seriously.
nasents Jan 3rd 2011 9:48AM
you should look up anecdote.
Scott Clark Jan 3rd 2011 9:48AM
If some guy says to you, "Hunters have been top performers in my guild's raids," that is anectodal evidence.
If several people say to you, "Hunters are OP in 25-mans," that is an aggregate of anecdotal evidence.
If you scrape logs for data, that's data. That's why they give it that name!
Feel free to disagree with the conclusions reached. Feel free to search for errors in logic or patterns missed. Don't impugn the principle of using data and numbers to back up an argument.
Eric Jan 3rd 2011 10:05AM
The reason the aggregate numbers are more valid than individual stories is that they reduce the impact of gear, individual play, etc. on the numbers. Looking at the top numbers also has the same effect, since they're likely to all be in about as good of gear as you can get, and likely to be good players. Removing the influence of gear and ability as much as possible allows you to analyze the spec itself.
JT Jan 3rd 2011 10:19AM
nasents,
I basically defined anecdote in my comment. Learn to read?
Scott Clark,
I did both. He completely ignored 10 mans, and he only included top parses (the majority of WoW players do not qualify as "top".) That's a poor analysis, and if he had included 10 mans as well as a range of parses (top, bottom, middle) in his sample size, the resulting data might well have painted a different picture.
Furthermore, I am not impugning the concept of using numbers and data to back up an argument, I am impugning the concept of *selectively* using numbers and data to back up an argument while pre-emptively discrediting disagreements by calling into question the foundation upon which that disagreement would be based, merely because it is the only data point that one particular individual has access to.
Numbers and statistics are numbers an statistics; they do no good without a concrete rationale for which are included and which are excluded. That's clearly lacking here.
I'll try to put it more simply: quoting 25 man WoL parses from the top guilds DOES ABSOLUTELY NO GOOD to the SV hunter currently being out dps'd by some other class/spec in a 10 man.
The only conclusion drawn here is that SV is doing more DPS in 25s being run by the top guilds. So what?
Most WoW players are simply not in those guilds, and are not raiding at that size, so those data are irrelevant to them. A thorough parse would've included 10 mans, and would've included all data, not merely those from the top.
JT Jan 3rd 2011 10:28AM
Eric,
Analyzing the spec "in a vacuum" doesn't do us much good since WoW isn't played in a vacuum and gear, skill, contributing raid buffs, fight mechanics, latency, and a host of other factors all contribute to the state of *overall* dps.
Frostheim Jan 3rd 2011 11:18AM
If someone says "In MY raid, blah happens" that's an anecdote and is not at all relevant. We don't care if he outdpsed the rogue or the dk outdpsed him, because we have no way of estimating skill and gear and buff levels, or even whether he's exaggerating. If a lot of people say that -- an aggregate of anecdotes, then we get interested and want to look into it. When we plow through thousands of WoL benchmarks, that's data. In science we use data.
My numbers are better than someone reporting what happened in their raid, because of quantity, because they've been sorted in a way to let us assume equal skill, gear, and buff levels (thus looking at the top result, ie: what the class is capable of), and because they are verifiable.
If you'll re-read the article, I did address the 25-man vs 10-man. Right now the number of parses is on the low end and thus the 10-man results are going to be a bit less reliable since we could have buff differences in the top parses. That said, hunters are across the board performing even better in 10-mans, which I also mentioned in the article.
And finally, you touched a pet peeve of mine... and exception DISPROVES a rule, it doesn't prove it.
Stilhelm Jan 3rd 2011 12:11PM
JT, the reason for using the top parses is because that allows you to compare top players, where the variance in skill is not so great. That allows you to explore and track the capability of the individual specs, which is what we are researching here.
JT quote: "I'll try to put it more simply: quoting 25 man WoL parses from the top guilds DOES ABSOLUTELY NO GOOD to the SV hunter currently being out dps'd by some other class/spec in a 10 man. "
So, because a poor SV hunter, improperly specced and geared, got outperformed by a sub rogue in his guild's 10-man, we should use *that* data to determine that SV needs buffed and sub needs nerfed? That is the logical conclusion of your statement.
I'll argue the opposite point. If any survival hunter is being beaten in damage by another comparably geared spec in almost any fight, they can look at that chart to see that the problem is *the player, not the class*. They can then use that information to do the proper research and learn to play/spec/gear to the best of their class's abilities.
Ace Jan 3rd 2011 12:24PM
"Analyzing the spec "in a vacuum" doesn't do us much good since WoW isn't played in a vacuum and gear, skill, contributing raid buffs, fight mechanics, latency, and a host of other factors all contribute to the state of *overall* dps."
Please explain how we are to factor in latency? Are pallies playing on older computers and slower connections? Are hunters more affluent and able to afford better PCs?
Sure some raids may have missed a buff or two, but for the top guilds doing 25 man content, gear is pretty much the same. To negate the factor of differing buffs you go for quantity of data. It flushes out these abnormalities and gives you better average numbers.
As for the use of anecdotal evidence, it is completely NOT useful. Here's why: To use Frost's example "In my raid, the rogue always does more damage than me, so SV is clearly not too strong." Ok... by how much? What kind of rogue? Do you have the proper gear choices? Is he really good and you just suck? That is why an anecdotal comment like that does absolutely nothing to help the objective approach to SV (or any) dps analysis.
Now, you're right that looking at the top SV hunters in WoW does nothing for the SV hunter that's in 5 man gear wiping in heroics being out-dps'ed by a warrior or something. HOWEVER, the purpose of this article is not "How to play Survival". Frost already wrote that article. Its called Survival 101. The purpose of this article is to analyze how SV is overly strong potentially, and what Blizz can/will do to balance things out and why.
JT Jan 3rd 2011 1:17PM
Frostheim,
First off, it's from Cicero: exceptio probat regulam in casibus non exceptis, and it's a principle from Latin jurisprudence, so your pet peeve there is with him, not me, but I assure you that the quotation is correct. It means, essentially, that exceptions prove that the rule is confirmed in all cases except those excepted. It's still taught today in legal schools.
Look, I'm not disagreeing with you that SV hunters are doing more damages than other specs/classes. I saw that you addressed 10 man, however it was a dismissal more than a discussion of it. If hunters are indeed performing better in 10 mans as well, is that not also relevant? Or are you proposing 10 man isn't "real raiding"? Wouldn't it be helpful to include that parse in your aggregate?
Here's the meat of my problem with your final statement though: you can't just say an individual's single experience is irrelevant because all you're doing is gathering up groups of single experiences and looking at them in a larger picture.
But that single experience is still relevant and valid. That's like saying "Well, sucks that you died from a jet engine dropping on your house, but that doesn't happen to most people."
The issue here is one of tone. You've basically put out a claim and then said "You can't disagree with me based on your personal experience" when realistically, that's all most of your audience has.
Stilhelm,
I understand the concept. I disagree with the concept: most players aren't top players. Specs are simply tools in the hands of players, and you can't compare them adequately without comparing also all players. For example, let's say that SV has an easier rotation than MM. Top players may be able to perform both adequately, but the average player might not. Why would it make sense to compare top players with top players when most players (by definition of the word average) are average?
We should use the poorly geared player's data because his data is still data. We should use all data, not just the data we like the most. That's...the whole point of data gathering. Try publishing a study where you've ignored data because it didn't fit your preferred model. You'd be laughed out of every academic publishing house in the country.
Ace,
Again, data is still data. You can't ignore it because you don't like it. It is unhelpful to say "But that rogue has more dps than me" because the data is incomplete. It is not unhelpful to say "But that XX/XX/XX specc'd rogue of average ilvl XXX gear is our dpsing my XX/XX/XX specc'd hunter of ilvl XXX. We've both been playing for a few years, and we were competitive in ICC."
Data still exists whether you choose to acknowledge it or not, but failing to acknowledge it weakens your overall argument.
messiahxi Jan 3rd 2011 1:55PM
I feel compelled to mention that "data" is the plural of "anecdote."
Nimfe Jan 3rd 2011 2:04PM
JT,
First of all, why would Cicero be more right than Frostheim? Cicero lived 2000 years ago and he may have been a philosopher, -actually just a lecturer but whatever- that doesn't mean what he says is right, even so Popper -a philosopher not quite as old as cicero- says a rule is a rule untill proven wrong (exception disproves a rule) so there are many different opinions on that one. And personally, i think that if there is a rule stating: "All crows are black". And i find a white bird i can correctly identify as a crow, then that exception did just disprove the rule. How dare they teach that in your legal schools?!
Moving on, you say: "you can't just say an individual's single experience is irrelevant because all you're doing is gathering up groups of single experiences and looking at them in a larger picture."
Actually you can, all positive sciences work with greater numbers of data. If the data is a personal experience (subjective data), then it will not give an accurate measure of the situation.
Example: if a person writes in his/her diary about an earthquake, and writes it was the most horrible experience they have ever had, would you think of that as true? Maybe, but maybe not... see, this person could be exaggerating the situation (because of the emotional shock), or maybe if a person with a mental disease would write it (thats a reference to the sub-optimal players you want to include in the evidence), that would make the story a lot less plausible.
Now if you would instead have 500 scientists and 500 other villagers write about what they felt and saw during the earthquake, that would probably be a lot more truthful if you wanted to find out what happened during the earthquake.
Now to get back to the data, that he excludes sub-optimal players and 10-man raiders, is not because frostheim hates them (he will probably make a full coverage of 10 man hunter stuff things, right? right!). It is because it makes the data less accurate and that will prevent him from reaching his goal.
OT: Frostheim, you are a great writer, i always read your essays and i find them highly entertaining (more Grandpappy Frostheim plz!). I never really felt the need to post before but i could not let this discussion slip ^^
Cheers
messiahxi Jan 3rd 2011 2:06PM
...and yes, data exists whether you acknowledge it or not. That doesn't mean it is statistically significant. As many have already pointed out, using only data from top 25-man raids you minimize the number of variables that act on the data (gear, skill, buffs). Class/spec is the only variable we care about here.
also, i don't believe Frost ever claimed that this was a peer reviewed study, so just let it go.