Frost answered the call. He surprised me on the last podcast by telling me that he took his whole raid to Ironforge to pound on a dummy for two 5 minute tests, ostensibly to see what the stand still DPS of his raiding team was without any gimmicks. He did the first one as MM and got 9701 DPS in recount, and the second one as BM, hitting 8255 DPS in recount.
His complete methodology is on his post. Notable:
- They had heroism both times
- Neither test was using kill shot, which would have favored MM
- He actually blew several hundred gold regemming for BM stats (and then regemming back for his raid)
- He doesn’t mention this in the post, but on the show, he told me he had to build his survival spec while kiting blood beasts in ICC :P
The only thing that would make this a better data point would be a combat log or world of raids report or something. As it is, recount damage done is a pretty good yardstick for performance, and now I have a good answer for the question “On stand and shoot fights, how much less DPS is BM than MM?” About 15%.
Also to note- Frost was in the top 5 with BM. If that’s good enough for your raid, then BM is indeed raid viable. If it’s not good enough, then it’s not.
Euripides
jurgwena
lemli
eidotrope
pherra
femaledwarf
Shadowstreak
Yep, I’d say that’s a perfectly reasonable set of data, especially with TOC25/ICC gear. (In Naxx/Ulduar/early-TOC, I think the gap was less obvious, though that may just be because the 15% now is significantly larger in terms of raw DPS numbers.)
Our guild threw together an alt/pug TotC(25) run last night. I decided it would be fun to try a couple of fights as BM (I’ll e-mail you the log link).
We had four hunters in the group (2 SV, 1 MM and 1 BM). Unfortunately, it was a very sloppy run. The MM hunter (who was pretty well geared) died on the worms.
We all lived on Jarraxus and that probably has the best data. For faction champs I switched back to MM, b/c the pug hunter was a jerk and I knew I could out dps him. We wiped and the pugs were kicked. I stayed MM the rest of the night. In short it’s not the best data, but it’s something.
For me, my BM dps is probably above the threshold of what’s needed for our raid team, but it’s below what I’m capable of doing as a hunter.
I might do a couple of ICC fights as BM this week. We’ll see.
This just has me extra-enthused to collect some comparison data! Way to go Frost!
I think examining this further should be interesting – the data points that are showing a 30% disparity are likely looking at unskilled BM vs. skilled MM hunter numbers. Should be neat to see how it looks with skill on both specs doing their best.
For my own MM vs BM test, the results show….
…I need moar skillz to do MM justice. Heh. Doing a test dummy fight as BM and then switching to MM showed an improvement in DPS, but not nearly the marked improvement that one would expect. The difference was so small, in fact, that it could have largely been attributed to the general variance of crit streaks and trinket procs. This is likely largely due to skill – I understand the MM rotation, but my muscle-memory for MM whack-a-mole is not so refined. There would be simple mistakes – I would sometimes miss a Chimera Shot when I’d accidentally hit Aimed Shot when they were both off cooldown, etc.
I also shied away from completely regemming (but I did re-glyph) when it occurred to me that there were BoE epics I could still be gunning for on the AH. I’ll probably stick with MM for a while to get some better skills at the rotation, and then revisit the comparison in a week or so.
The difference in DPS makes perfect sense:
If your raid needs Ferocious Inspiration, then BM > MM/Surv, because the +3% damage you are bringing to the rest of the raid is greater than the ~15% hit to your personal DPS.
If your raid doesn’t need Ferocious Inspiration, usually because there’s a LOLRETADIN in the raid, then BM < MM/Surv because you are taking the ~15% hit for no gain.
@gahad I do not like being referred to as a Huntard and I am sure you don’t either. Was it necessary to mock Ret Pallies in your reply? In our raid group we have had two excellent Ret’s (currently & in TBC), nothing lolret about them.
The 10% to AP that MM provides is nothing to sneeze at either, especially if your team has no Blood DK or Enhancement Shammy.
I find “lolret” to be more of a term of endearment than anything else.
I think “Lolret” is far more distinguished and much less prejorative than “huntard”.
In classic WoW it was a terrible spec for end-game raiding or PVP, almost completely useless. For Paladins, it was Holy or gtfo. “Lolret” was a slight on the comparative weakness of the spec, and the spec itself, not an insult directed at the player.
During Burning Crusde, for a brief period, the Ret paladin was a virtually unkillable stun-locking killing machine in PVP. “Lolret” was a slight on the comparative overpoweredness of the spec, and the e.z.mode that it represented, not an insult directed at the player.
Then they got nerfed “into the ground, baby”, and the balance swung back. And it’s been swinging ever since.
Anyway, I don’t accept that “lolret” is even an insult, let alone one on a par with “huntard”.
If you dont like being called a huntard… dont be one.
Would it be possible to get some more information on on what Frostheims testing? I recently respecced to what EJ has as the 3.3 BM build, grabbed my Devilsaur (respecced it too!), and jumped into a TOC 25. I was very surprised at what i saw in DPS. i usually run MM and hold my own in DPS and i was about the same….just found that BM was so much easier and my Devilsaur owned the Faction Champs. Not saying im ready to go BM in ICC tonight….but i just might!
I’m glad Frost finally got experimental on this issue, and I’m looking forward to reading all about it. I’ll just say that for me, this result illustrates two things:
1) BM is so “viable.” :-P
2) The BM scaling issue is real, and large, with current end-game gear.
Is not flasking ‘viable’? Not having food buffs? Gemming for mp5 instead of agility?
Assuming that you’re in a 25-man raid, where the buffs brought by both specs are covered (so you’re not bringing a corresponding raid buff), all of these things are much much smaller DPS losses than raiding as a BM hunter. As a raid leader, I would feel uncomfortable with a hunter gemming MP5 over Agility; the same way, on any progression content, I’d feel uncomfortable with a hunter going BM instead of MM.
You’re describing one type of guild. There are indeed guilds where people aren’t expected to min/max, and in that environment, BM is viable.
Gemming for MP/5 is an indication that the hunter doesn’t know what they’re doing. That’s not what we’re talking about here.
I’d feel uncomfortable (or charitable, take your pick) carrying a hunter gemmed for MP/5. I’d have no problem with a BM hunter placing in the top five (or any class or spec doing 8K+ DPS), ’cause they clearly know what they’re doing.
Gnorf: Okay, change it to “not gemming at all.” You’re right that gemming for mp5 indicates cluelessness — but not gemming at all indicates that you are willing to give up DPS in order to have more money in your bags, at the cost of DPS loss.
Essentially, raiding as BM — which, I’ll note, I do regularly — you’re making a choice to intentionally give up DPS in order to choose a different playstyle. Whether the playstyle you’re choosing is “running as MM, but not bothering to spend money on enchants, because I’d rather collect pets/mounts”, or your playstyle is “running as BM, because I prefer the way that plays”, you’re making a choice about what you do which affects the people you play with.
If you’re not raiding ICC, BM is completely viable for raiding. If you’re on a non-DPS check fight, raiding as BM is completely viable. In both cases, you can meet the minimum requirements of the fight.
If you’re wiping on Festergut because you’re bumping into enrage, running as a BM hunter isn’t viable — in the same way that if you were ungemmed and running into the enrage, that would no longer be a viable playstyle.
I play BM all the time. I use BM on fights where it makes more sense to me, fights where it’s more useful, or fights where my DPS doesn’t matter. Do I still beat other better-geared Survial and Marksman hunters? Absolutely. Does that mean that I go to BM when DPS matters? No.
Gemming and enchanting your gear is a playstyle. Running as a BM hunter is a playstyle. Each are reasonable things to do in certain situations; neither implies a lack of knowledge.
In my opinion, having a BM spec that you use when it’s appropriate is perfectly reasonable — but when DPS really counts, you’re likely going to need to have something else you know how to play well to fill the gap that BM leaves.
Not gemming at all sends a different signal. Even in a not-so-hardcore raid guild (like mine), the guild supplies gems, enchants, glyphs, and plenty of expertise in choosing appropriately.
It sounds like we actually agree on this. I’m all for measuring BM according to actual results, against the requirements of the raid. If we’re not getting the job done due to lack of DPS–opposed to failing to learn the fight mechanics (in which case, I’d much rather have a clueful BM hunter than replace him)–then I don’t mind asking a BM hunter to respec. But I’d use that standard for every DPS in the raid, and not try to pin failure on a single DPS. And before I looked at the BM hunter, I’d look at my own damn self, and try to fix my own failures.
Yeah, I think we’re in violent agreement here :)
I think that choosing to run BM is giving up DPS. If you give up DPS and still top the charts in your guild/raid group, then it just doesn’t matter (except that kills take a couple extra seconds; even a 20-minute fight is only 30 seconds longer with a BM hunter in place of a Marks hunter in a 10-man raid). But everyone should know what tools they have available to them — and any BM hunter should, imho, have the tool of a second higher DPS spec on hand for cases where they need it.
Last night I killed Rotface with no one doing over 5k DPS. What was far more important than our DPS was our coordination; by the end of the fight, we had up to 2 disesases + 2 small slimes up at a time. We downed the fight without a wipe.
There’s no question that I would have loved more DPS — but only with coordination + skill were we able to keep going in that situation.
When we went to Putricide, we had to call it off on that boss — and went to Velithera for a while — because the DPS to get down the slimes wasn’t there, and we were nowhere near the amount we would have needed for P3. In that case, I’d have loved to know that our hunter could switch to another spec and pick up 15% DPS. (Sadly, he was already Survival, so no such easy fix existed.)
was this on the test server with the new buff to steady shot and arcane shot if not the gap is closing
did they have a ret pally covering the Ferocious Inspiration buff and a blood dk or shaman covering the 10% attack power buff
also mangle from the druid would also help the MM hunter more than the BM hunter unles mangle buffs a BM pets attack
We had all raid buffs (it was a 25-man, after all, and we know that getting the right buffs contributes more to our dps than several tiers of gear :) )
This was on a live server. Again, with no kill shots in the mix MM showed a bit lower than it normally would, but I figure the coming BM steady shot buff should just about balance that out. So that 15% is more accurate post-buff. Currently on live it’s probably closer to 16-17%.
Frost, what was your BM spec, pet and shot priority in BM during the test ? your article doesn’t go into this detail.
While for MM there is a “certain” bets priority, spec and pet known to everybody for BM this is not the case and some MM hunter respeccing BM often use suboptimal spec and suboptimal shot rotations.
Your results of 15% difference are still a little bit far from the truth but much closer than in the infamous article where you advertised 30% difference.
I have reasons to say that today the difference is at no more than 9%-10% (including Kill Shot) and after PTR 3.3.3 this gap will almost close leaving MM ahead by just 4-5%.
Don’t forget, Frostheim raided as BM until they nerfed it, and I am willing to assume that he did as well with his BM rotation/cooldown timing as anyone could do.
The aimed/arcane shot buff is going to be about 1% of an increase for BM, btw.
I don’t doubt Frost’s abilities, but the level of depth of understanding of what BM is TODAY, not before the nerf.
The best BM build has changed a lot. Best glyphs too.
The best shot priority has one shoot less than back then.
The best pet is not anymore what it used to be.
BM can be gemmed in 3 ways that are in the +/-1% dps range and i don’t particularity care about this, but it would be interesting to know it as well.
Was his Pet raid buffed with every buff ? Was he given pet STR food ?
And Lastly, why 5 minutes and not 3 ? or not 7 ? This has a big impact on CD due to the readiness/rapid fire bestial wrath CD that impact the comparison.
So in short i am curious to know the answers to the main points above (build, glyphs, pet, shoot priority Frost has used for the test) because the difference from the 15% he found and the 9% i see must be due to one to one or another of those factors.
The buff on PTR is not aimed/arcane, but steady/arcane. Essentially now FI is up 100% of the time (vs. something less than 100% of the time which is the case in live now – but let’s assume this effect is negligible) and additionally buffs Steady for 9% dmg on top of what it does already.
Steady represent for BM about 30% of the hunter dmg output (pet excluded) and 21% of the total dmg output. Hence the buff will be around 2% overall DPS, not enough for “parity” on “stand and shooot” targets but getting closer..
Plus on real fights, hunters move and lose DPS, pets dont. So in some real encounters the two spec
can actually compare 1:1.
Er, is it not possible that the difference between his 9% and your 15% could well be:
1. Gear
2. Skill
BM has fallen behind further with more gear disparity; if he’s sporting 264s and you’re in 251s, that can make a difference. (Spreadsheets for me have gone from parity in 226s to a win for MM in TOC gear to MM crushing BM in 251/264 pieces; the difference in game seems to match that.) For example, something as simple as a weapon upgrade gives my MM hunter a 400dps boost… but my BM hunter only half that.
An important thing to do when looking at MM is lining up procs with serpent sting refreshes; if you can line up procs of 3 things — two trinkets and a ring — and make sure you don’t let SS drop after that, you’re going to pick up a fair amount of DPS over the course of a fight.
Certainly fight timing plays a big difference — in fact, this is one of the reasons BM is killer on chain-pulling short-lived single-target trash; Rapid Fire + readiness vs. Bestial Wrath.
I don’t doubt that spec/rotation/etc. plays into that, but it’s seemed to me that the glyphs/spec/rotation for BM are pretty well-understood in the spreadsheeting / elitist jerks community; I expect the 53/11/7 BM build has a small enough set of variations that the differences between his 15% and your 9% are probably to be found elsewhere.
A lot of things you say are correct (gun update for example). But the spec/rotation and glyph of BM are not fully understood.
Actually i start to believe that there are some issues with the theorycraft at present.
The latest Shandara dps spreadsheet shows the best BM build at 1500DPS behind the best MM, with MM at 15000 and BM at 13500. But the BM build that seems to give the best results on the spreadhseet is not the best build in the real life. The glyphs in it are not the best glyph for the build etc. The problem is that builds that hare clearly better in real life gaming are not represented well there, like if the spreadsheet works for Bis object on thoeric mobs only but anywhere around that “space” the results are approximated.
Having said that the test Frost has made obviously has some degree of uncertainty, if the other hunter went from 8400dps at 7900dps between the two try there is reason to believe the whole raid was at +/-6% accuracy which would bring the difference between BM and MM more where i am saying it is, which is about 9-10%.
Did ICC 10 yesterday with 5.6kGS MM hunter I’m 5.1kGS BM 50/21 readiness build with wolf pet. I’m gemmed for armor pen. I was invited after they downed lord marrogar to fill a spot for deathwhisper. Did 3 attempts. All 3 attempts we both were the last to die. 1st try MM was 100 dps better, 2nd try we were less than 100 apart. 3rd try we subbed the lowest dps for well geared warrior dps and I beat the MM spec hunter by 500 dps. Also have a hunter friend with 4.8kgs duel spec BM/MM we raided TotC to test which spec he would do more dps in. He’s gemmed agi used ghost bear as BM pet and wolf for MM. After extensive testing in both specs there was no differance at his gear lvl.
So… All this means that I should stick to SV? Thanks for clarifying that for me.
“I have reasons to say that today the difference is at no more than 9%-10% (including Kill Shot) and after PTR 3.3.3 this gap will almost close leaving MM ahead by just 4-5%.”
If you have conflicting data, that’s certainly something we can discuss. How did you do your testing to get your info? Was it under controlled circumstances? What gear levels are you looking at? I’ve looked at very high gear levels and now mid (about ilvl 250) gear levels and the spread runs 15% – 30%, depending on the gear and the fight.
The current PTR buff is a flat 9% steady shot increase, which translates to about a 1.56% boost to overall hunter+pet dps. So I’m also curious to see how you’re going from 9-10% (numbers I can’t seem to replicate) to predicting 4-5% difference post patch. What’s the math?
The target dummy test was 5 minutes because that’s a reasonable boss fight duration, and a common enrage timer for dps race fights.
I used a wolf pet for the tests.
As I said in my article, I play as BM quite a lot. I’m familiar with the rotation and I’m pretty confident in my ability to handle the complexities of BM while I’m just sitting there staring at cooldowns :)
BM is kind of a mixed bag on movement fights. If you’re bad at managing your movement, BM will be very good to you, since a chunk of your dps will still be going while you’re running like mad. However if you’re skilled at movement management, BM is unkind since BM doesn’t have many instant shots that can be fired on the run, and requires a larger percentage of your time to be Steady Shots. BM also suffers in target switching fights.
However what we can measure somewhat reliably is a perfect stand-still fight of a fixed duration.
But again, I am always happy to argue science. If you’ve got some data, let us know what it is and how you got it!
Frost, i don’t have data yet due to time constraints and to the fact that i do not have the funds to re-gem nor got the opportunity of getting the Zod’s bow and i am running with the tank gun from ICC at present (i have been particularly unlucky as it dropped once in the 6 kills i had and i did not get it).
If tonight the lady drops the bow it is my intention to run the test in a controlled environment as i have the cash required for regemming.
The 9-10% figures i got comes from testing without changing gems (keeping the AGI ones) in real raid situations by averaging out results of 4-5 bosses of ICC 25m or the two main guys in VOA25m.
Are averaged values and the statistic science shows that averaged values often gives better info than punctual experiment as the outliers are filtered out more and the situational episodes have less relevance.
For MM i use the “cut and paste” spec that is widely known, while for BM i use the 54/11/6 build with just 2/5 mortal shots and no aimed, glyph of Steady, Serpent and Hawk. Pet is wolf, although is it not the best choice, best enchants/ apart from eng ench for back and hands. Priority is Rapid, BW, Blood fury (Orc), kill shot, serpent, arcane , steady.
One thing i do not like of your test is that the other hunter did 500 DPS less in the second try, 6% less. By the same rule your results of the second try could be 6% underestimated.
In any case as soon as i have my data i will let you know .. keep in mind will not be punctual data but averaged, hence it might take soem time to consolidate them.
Ah, okay, so what you have is annecdotal evidence – not something that was done in under any kinds of controls.
When we’re talking about average dps in raids, the variables are massive — not just gear and skill, but RNG for you and the others and boss mechanics, duration of the fight, raid composition, etc.
It only becomes really statistically significant data over a huge number of data points… in other words, looking at the top raid parses from thousands of parses. Which, of course, we’ve already done :)
sorry my reply to you came as an overall reply here below.
There is nothing more anecdotal evidence than the test you have conducted.
You talk about science, but the only two ways to evaluate scientifically the specs are:
1. Deterministic methods – via exact mathematical models. This is done via the Shandara spreadsheet which seems to be the most accurate deterministic model available which shows best MM spec at 15000 and best BM at 13500 – 10% difference
2. Statistical methods. Those needs to be based on experiments conducted under controleld evironment – like the one you did – but needs to include several samples and not just one as you did. Using just one sample completely ignores the fundamentals of experimentation. the basis of statistic, the concepts of Sample Size, Significance, Confidence Interval, Average, Standard Deviation and any other fundamental …
How can you possibly say your test, single, lonely and approximative can be science or facts ?
As far as Science is concerned that is just one sample, it’s significance, by himself, is zero.
At present the best possible assessment, given the lack of a large enough pool of experimental data, is still the deterministic models (spreadsheet). And those show a 10% difference.
He isn’t claiming his one test is statistical, the way I read it, he’s saying that the testing was done to create the deterministic mathematical model, and what his test shows is that these models are accurate for their comparison of BM and MM at his gear level.
Wow, never heard that deterministic mathematical model are created by using experiments.
Deterministic mathematical models comes from relation input/output well known on the basis of 1st principle rules/law (i.e. in this case the equations used by the Dev when they created the game – e.g. the theorycraft is what we have as close as possible to those equations).
All other mathematical models created by experiment and data analysis are statistical models, not deterministic. As such they are created by using statistical methods, such as DOE, C&E, Principal components analysis etc. etc.
1 experiment is no good for this purpose. Nobody with a tiny understanding of statistic would agree with me. Now if he did make 10 experiments per spec removed the highest and the lowest results, calculate the average and std deviation, could also have calculated the confidence level (whcih with 10 sample would still be at around 4%-5% accuracy) of the results “goodness” per each spec we would be comparing some more concrete data.
As far as science is concerned that experiment with one sample could have well be an out-lier in a way or another and the comparison be completely wrong, even loop-sided.
So far the only source of relevant info on this subject is still only the Shandara spreadsheet (and by any means i am not saying it is perfect, but is the least flawed).
But i start to have the feeling that i am wasting my time.
Er, I don’t think he’s saying that his test is anything more than one sample; however, it’s a (single) controlled experiment. I highly doubt anyone would object to further data points being collected.
Personally, I don’t have a 25-man group that I can convince to beat on target dummies for the fun of it. :) But the data that he’s collected seems in line with my experience and expectations.
For my personal gear — which seems to be on the same general level as Frost’s — the difference in DPS is 16.7% (10230 vs. 8766). Anecdotal evidence in the form of my personal experience competing against myself confirms that approximates reality. Frost’s controlled experiment confirms that this approximates reality for his gear as well.
You seem to be saying “Okay, the spreadsheet at BiS levels shows only a 10% difference — how do you explain your 15% difference?”, combined with “In actual raid instances, I see only a 10% difference — how do you explain your 15% difference?” Since I assume you don’t have all BiS gear, I assume that your personal spreadsheet difference isn’t 10% (though I don’t see that you’ve said what your difference is) — it’s just as fair to ask (as I tried to) why you think that your anecdotes are somehow more valuable to answering the questions at hand than the more controlled experiment that Frost attempted to perform.
If there was a fight where the entire time, the boss was dragged down the hall by a tank, BM would probably do better, because a (more) significant portion of the DPS wouldn’t be lost in BM. Does that mean that repeatedly showing that BM does better on that fight would ‘prove’ anything about BM?
The deterministic model shows different values for different levels of gear. For my gear, it shows a 17% difference. For your gear, it may shows something else (though you haven’t said). In addition, spreadsheets can be wrong for any number of reasons, and testing the spreadsheet with an experiment is obviously valuable. Frost tested the spreadsheet. You’re testing raiding. The difference between the two is huge, imho.
“For my personal gear — which seems to be on the same general level as Frost’s — the difference in DPS is 16.7% (10230 vs. 8766).”
10230DPS is 16% more than 8766DPS, but the difference between the two values is 14.3%.
Why am i talking about statistics here when even basic maths is an issue …
“I assume you don’t have all BiS gear, I assume that your personal spreadsheet difference isn’t 10% (though I don’t see that you’ve said what your difference is)”
I don’t have all the BiS items, i am at around 251 iLevel, the difference for me is at around 9% between the two specs, and theoretically the Shandalara spreadsheet gives me 9% difference. I did say all this on previous posts.
Keep it civil.
It is science. It’s just not *ideal* science.
Nobody has committed heresy against Science, here. The experiment and results are accompanied by disclaimers, and are interpretable by thoughtful hunters.
Further, the experiment serves as a useful verification of–and check against–the theoretical models, which give us incredibly precise results, which turn out to be wildly inaccurate in the game.
Gnort, i agree with you, if you would replace “experiment” with “experiments”.
Theory says 10% difference with BiS, lower gear the difference should be less. If the experiment was used to confirm the theory has failed.
My issue is that this experiment was used to promote a self fulfilling cause of demonstrating MM is further ahead than what the theory says from BM.
I have been using a devilsaur and i am really impressed with what im seeing. Worked really well in ICC but i changed back to MM for saurfang. Did a 25 toc and the deivlsaur pulled 2750 on Jaraxxus by itself. I have also seen some insane, for me, DPS at the beggining of boss fights where i can pop all CD’s and trinks. Just been alot of fun…..at the end of the day i am sure MM is the spec of choice for DPS but with the variables that happen with every raid i am sure that BM can be competitive.
Biggest drawback is when a pet dies…huge DPS loss for BM.