Hmm, good catch. Without FM-trading, 18/53/0 fireball appears to drop to ~6040 DPS in a 5-minute fight, and ~6200 in a 3-minute fight. That would narrow the gap enough for Rune of Razorice to basically deadlock the specs.
So, who do we have to harass to get Frost Vulnerability added to Rawr, so we can get some concrete numbers?
The real question is whether or not that 3% crit can provide more DPS than 3% personal crit, less than that, or about equal. I doubt other casters can make more dps out of 3% crit than mages, but it does need to be looked into more thoroughly. If you have more than yourself as a mage in the raid, though, and the other mage is equally geared/skilled, then it is as good as 3% personal crit and should be modeled as such.
On a more personal basis, if your DPS is noticeably lower than the top DPSer in your guild, and he's a class that benefits properly from crit, then it'll probably be worth *more* than 3% personal crit.
So, I went on ptr today, and had a play around with the Arcane specs.
I'm no maths whiz, so these are only preliminary findings, I'll leave it to Manly and the others to do the modelling etc. Consider this a 'Dummy's Introduction to Arcane'
I tried several specs and cast combinations, with varying armours. Numbers are below.
Most interesting thing I found was the feeling of it. Arcane, unlike the other specs, has a huge amount of player input. You need to manage things a lot more carefully than hit the macro button everytime your CDs are up. The management of all your abilities, and the variation between the cast sequences is great.
Mana. I stood in IF and nuked the Boss level target dummy. My gear is ok, 3T7.5, 1T7, Gothik's Cowl, Life and Death, but I have a few blues that need upgrading. For what it's worth, my mana regen was 75 with no mage armour, and 151 with. 2000 SP, 23% crit, 12.5% haste.
For all the below, I cast the sequence, using mana CDs when they'd return the max mana. NO raid buffs, and NO mana pots were used. When I was OOM, and there were no mana CDs available, I stopped. Even if Evocation was 10 seconds left, I called the OOM point the end of the trial. Therefore, I'd assume these numbers will be significantly lower than raid buffed ones, not just due to the flasks/buffs, but also due to the mana return abilities that would prolong the cycle, and allow more flexibility in Mana CDs, and therefore more stacking (e.g. Mana gem with AP/IV for dps, rather than every CD or OOM).
53/0/18 - AB/ABar cast, AM on proc. Mage Armour. No mana issues at all. I maintained this rotation for 13+ minutes, recasting gems after Evo. With raid buffs, this would be an absolute joke to maintain. DPS of around 2100, so 2360 if we factor in TtW.
53/0/18 - ABx3/ABar cast, AM on proc. Molten. Almost 3 minutes. 2700dps.
57/3/11 - AB/ABar cast, AM on proc. Molten. Mana was tricky on this. Cast for about 5 minutes, then mana CDs fell apart. TtW dps, ~2460
57/3/11 - ABx2,ABar cast, AM on proc. Mage. 5 minutes casting again, TtW 2460 dps also.
57/3/11 - ABx2/ABar cast, AM on proc. Molten. Just over 3 minutes of casting. TtW 2600 dps.
Glyphs were for FFB spec (Molten, FFB and Scorch), so you'd need to compensate for the AB/AP/whatever glyphs used. You'd also need to factor in the extra crit from Focus Magic (which I specced, but had no-one to cast it on) and any other factors that could influence dps.
From the testing, it appears that mana will be less of an issue than previously thought. If I can cast for 3 minutes solid of an ABx2/ABar with AM on proc rotation in Molten armour, with 0 raid buffs, it seems that will be highly sustainable with Wisdom/Judgement/Replenishment. It might even be possible to use a ABx3/ABar rotation in fully buffed raids. The main thing that ate my mana was AM. On 2xAB/AM, the AM was taking ~900 mana to cast, from my 19,000 mana pool.
One last thing I though may be useful. When I popped my CDs on a ABx1/ABar sequence, there was a time when the AB cast was clipping the ABar CD, so there was a wait time between AB and ABar. If this occurs, then it makes sense to use an ABx2/ABar sequence (possibly an ABx3/ABar) during IV.
Thanks for taking the time to do the research. I would be curious to see the results from the most recent "accpeted" rotation in this thread. Using a 57/3/11 build and the rotation of ABx3 > AM > ABar with the AM/ABar combo of course. If you have the time it would be great to see as I have not been on the PTR yet.
Chiharu: Could you be a little more specific with your description of testing?
How many repeats of each test did you do? What was the total damage count? How was your experimental crit-rate compared to your expected crit-rate?
I find it a little hard to swallow that swapping from Mage Armor to Molten Armor netted you 14.4% increase in DPS.
What I think you're observing is CD-caused result skewing. A 13m cycle can (did?) include CDs three or four times, while a 3m one will include one use, and a perfectly-stacked one at that. It's easy for this one CD-use to produce wildly skewed results if one of your auto-proc trinkets went off at the same time, or you happened upon a crit-spike.
It's also rather unlikely a test without mage-armor glyph will be accurate for the larger part of this, even if we are discussing non-raidbuffed output.
Each test I conducted twice per spec/rotation, and took the average. Hardly a normal distribution, I realise, but I only have a few hours a week to play. I just want to make it clear that I am not suggesting that my results are the be all and end all of arcane knowledge. As I said, these are just some tests that I did, that I thought people may be interested in. Apologies if people are taking them as gospel, that was not in any way my intention.
I have at home a spreadsheet I kept, with the exact seconds of rotation maintained, the total damage, the dps and the mana returned per rotation. Unfortunately, Recount doesn't count natural mana regen in its 'Mana Gained' section, so for the '30% regen while casting' talent, I looked in my character pane, and multiplied the while casting number by the time before I went oom divided into 5 second blocks. For example, if I managed to maintain a rotation for 260 seconds, with a 75 mp5, I had (260/5)*75= 3900 mana regenned from 'organic' regen.
Regarding the CD cycles, I was admittedly less than scientific on those. Each of the burn cycles was an absolute killer on the mana, and so I tended to pop the cycle when I was ~4700 mana down, which would pop me back to full mana (2T7), as well as being the best use of the CD stack, IMHO. From then on, I tended to use the CDs when I believed I could utilise the full burn without making myself OOM and ending with no mana regen CDs available (evo/gem).
I completely agree that the CD stacking did skew the results, there's no way to avoid this, unless I were to conduct the tests with all CDs unused. Still mana gems would skew, from the unavoidable use of mana gems (and thus the buff), but this would negate the impact of the CD stacking on the dps results.
I also completely agree that without the correct glyphs, any testing is dubious at best. Unfortunately, the AH on the PTR is exceptionally bare. However, even glyphed I'd be dubious of the results, as I know the enormous difference that raid buffs can make (especially to a mana tight spec), and try as I may, I can't persuade my guild to pop over to the PTR and buff me up in IF.
The general points that I made I believe still stand, about how Arcane is the single current mage spec with a 'mana -> damage' conversion that can be actively managed by the player. I still plan on speccing Arcane in 3.0.8, but I believe it will take a few weeks of raiding to get a real 'feel' for the mana situation, and the maximum burn that can be applied in each boss.
I'm still open to any more simulations etc that might suggest a better arcane rotation. I'm not ashamed to say that everything I've learnt about mage dps has been from these forums, from specs to rotations, so anything else you guys can throw my way would be appreciated, and I will happily test.
I'm still open to any more simulations etc that might suggest a better arcane rotation. I'm not ashamed to say that everything I've learnt about mage dps has been from these forums, from specs to rotations, so anything else you guys can throw my way would be appreciated, and I will happily test.
I would be interested in seeing similar tests but using the ABx3, AM, ABar (AB Shatter) rotation. It's been explained a few times in several threads here, but essentially the final ABar takes advantage of the AB 3 stack before the final AM missile hits the target. 57/3/11 spec with both Molten Armor and Mage Armor. If you have a chance to try this out, it would be much appreciated.
Okay so with the upcoming changes to mana regen I wanted to roughly model the differences between
the following
LIVE - 0-58-13 with Frost Chaneling
PTR - 20/51/0 With Medidation and clearcasting
PTR - 0-58-13 With Frost Channeling
I wanted to quantify the value of having two points in meditation.
Conclusion:
Spec #1
AVG of 140 seconds (2.3 min)
Spec #2
AVG of 160 seconds (2.6 min)
Spec #3
AVG of 120 seconds (2 min)
The theory:
Base Assumptions:
Gnome Mage (Base mana 3268,)
400mp5 while not casting (roughy what i have raid buffed)
16000 buffed mana, 1 x 4000 mana gem 1 x 4000 mana potion (24000 mana)
Evocation isnt used
50% Crit change buffed (and debuff on boss)
Spec 1
LIVE - 0-58-13 (13% reduced cost no mana regen)
Spec 2
PTR - 0-58-13 (10% reduced cost 30% mana regen)
Spec 3
PTR - 20-51-0 (10% Clearcasting, 50% mana regen)
30% mana regen of 400 mp5 gives 120 mp5
50% mana regen of 400 mp5 gives 200 mp5
Mana Costs and Cast times for spells
Spec Number 3 2 1
Base Cost Pyro Frost Ch Pyr + FC Cast Time % Base
Living Bomb 719 697 647 626 1.29 22%
Scortch 261 253 235 227 1.29 8%
Fireball 621 602 2.6 19%
Frostfirebolt 457 443 411 398 2.6 14%
Pyroblast 719 697 647 626 1.29 22%
I made a spreadsheet that simulates a fight. it simple casts in the following order
Using the RAND() function of excell you can simulate randomness,
if RAND() > 0.5 crit else no crit.
if RAND() < 0.10 Clearcasting else no clearcasting
1st LB
2nd SC
3rd SC
Following casts are based on a rotuine, if HS is up it casts Pyro, if SC or LB need refreshed it refreshes it
There is also a routine that calculates the mana cost.
if anyone wants the spreadsheet with the code in it PM me and ill email it to you, Currently google docs dosent like macros
I would be interested in seeing similar tests but using the ABx3, AM, ABar (AB Shatter) rotation. It's been explained a few times in several threads here, but essentially the final ABar takes advantage of the AB 3 stack before the final AM missile hits the target. 57/3/11 spec with both Molten Armor and Mage Armor. If you have a chance to try this out, it would be much appreciated.
One of the first things that I tried when I finally got on the PTR was the Arcane shatter. From the testing I did, I found it to be less than easy to pull off. I tried the old 'hammer ABar while AM is channeling' technique, as well as the 'clipping' technique. I didn't attempt any proper timed attempts with this rotation, but I can forsee two problems.
First, the always hated lag/latency. To pull off an AM/ABar shatter with all five missiles fired, requires an almost instantaneous ABar. A slight bit of latency would therefore negate the benefit from this combo, and you'd be pretty much ramping up AM, and then firing a unbuffed ABar, which I must assume is a dps loss.
Second, mana. As I mentioned in my first post, the main thing that drained my mana was AM. The cost of AM is not small, somewhere in the region of 900 mana (that number may have been eyegouged by me from the mana bar, while maintaining rotations, but it was a significant chunk nontheless.)
If you ramp up the AM for the shatter combo, then 'miss' the ABar shatter, you've wasted a lot of mana, AND gimped your dps. IMO a combination of those two things will make this a very dicey combination to try, especially in an effects heavy raid environment, and in a situation where outside factors will affect your casting (i.e. NOT patchwerk).
This sort of rotates back onto my point about Arcane Stability. If you were trying Sarth3D as arcane using the shatter combo, then with the drakes and the eles, one hit will knock x amount off your AM, which will both gimp your dps and dpm (due to the upfront mana cost of the channeling) and also ruin your shatter combo. Depending on the Ulduar content, and also the prominence of AM when the most effective Arcane rotation is finally determined, I think Stability will be a required, or at least a preferable place for possible loose points.
Last edited by Chiharu : 01/19/09 at 5:11 PM.
Reason: Error
This sort of rotates back onto my point about Arcane Stability. If you were trying Sarth3D as arcane using the shatter combo, then with the drakes and the eles, one hit will knock x amount off your AM, which will both gimp your dps and dpm (due to the upfront mana cost of the channeling) and also ruin your shatter combo. Depending on the Ulduar content, and also the prominence of AM when the most effective Arcane rotation is finally determined, I think Stability will be a required, or at least a preferable place for possible loose points.
Arcane doesn't have a lot of spare points to play with if you're already speccing 18 points into frost.
But I have to agree with you on the shatter combo. It's not an easy thing to pull off. The only way I can be successful in it is to clip the fifth missile in place of arcane barrage, and even then this can be tricky with a 2.2 second cast arcane missiles with missiles loaded to the end rather than the beginning of the cast.
Arcane doesn't have a lot of spare points to play with if you're already speccing 18 points into frost.
But I have to agree with you on the shatter combo. It's not an easy thing to pull off. The only way I can be successful in it is to clip the fifth missile in place of arcane barrage, and even then this can be tricky with a 2.2 second cast arcane missiles with missiles loaded to the end rather than the beginning of the cast.
I am by no means a macro expert, but would it be possible to make a macro that only casts spell X if spell Y is not being channeled? This could make the shatter work more often than not.
Don't know how to write it any clearer. How much damage impact ABarr has varies wildly by rotation, but I've never seen a situation where you lose dps by casting ABarr, even with 0 AB buffs.
I honestly wonder if the new arcane changes are worth all FOTM'ing it. Isn't it basically a whole lot more effort in the rotation for basically roughly the same dps when comparing to current FFB builds? I hope it's just not a situational thing like the old arcane/frost spec needing the world of raid support to do well. Just like in TBC I preferred doing 2300-2500 as deep fire on brut pre3.0 not having to spam mana potions, without needing innervates.
It's worth it because it can be fun. It requires more of personal finesse to be competitive. And for some of us the mere fact of it to be NOT Fire can be the sole reason to master it. And upcoming changes make such efforts worth the time.
The problem of arcane is that mages gear up themselves towards FFB spec. Optimal gear would give similiar dps, but clearly mages using dkp have limited choices. I am planning to spend some valors for some arcane fun myself, for farming contents.
With an Arcane spec, you give all your healers a free flask, which I consider quite a lot.
Magic Absorption might also be beneficial. Both can be brought by Fire Mages however.
Arcane also has a lot of situationally useful tools.
The main point however is that Arcane offers a vastly higher mobile DPS than any other spec.
It also has a high burst potential with low threat.
If you only want to look at your Patchwerk meters, go (Frost)Fire.
Arcane's strength are mobile and add fights like Heigan and probably Sapphiron or Gothik.
By the way, did anyone ever check on their character pane whether Pyromaniac is 3% fire or 3% spell crit now?
Don't know how to write it any clearer. How much damage impact ABarr has varies wildly by rotation, but I've never seen a situation where you lose dps by casting ABarr, even with 0 AB buffs.
I absolutely agree with this, with the tiny caveat of the following scenario:
Assuming a 3*AB-AM/ABarr shatter combo rotation that ignores whether or not MBAM procced, looks like it may be less DPS than a more convetional 3*AB/ABarr, MBAM when proc rotation. That is to say, if you're casting AM/ABarr shatter (MBAM or not) after 3 ABs as a fixed rotation, and failing the ABarr shatter, then you're probably not making any headway.
Strictly, what you mentioned, that ABarr is never a DPS loss, is correct. Technically what's causing the DPS loss in this scenario is the effective ignoring/missuse of MBAM and/or Non-MBAM AM.
I honestly wonder if the new arcane changes are worth all FOTM'ing it. Isn't it basically a whole lot more effort in the rotation for basically roughly the same dps when comparing to current FFB builds? I hope it's just not a situational thing like the old arcane/frost spec needing the world of raid support to do well. Just like in TBC I preferred doing 2300-2500 as deep fire on brut pre3.0 not having to spam mana potions, without needing innervates.
Current PTR arcane specs are currently much more sustainable than people think. Anyone who has played the spec on the PTR could tell you this. For starters the 53/0/18 spec has a 16% mana cost reduction along with clearcasting on top of this. Using gems and evocates as early as possible an arcane rotation is quite sustainable on a training dummy. Raid buffed and with raid support such as a ret pally it will definitely be sustainable. And if a fight is tricky on mana we can just switch to a less expensive rotation.
I just wanted to provide some quick damage data from my PTR testing with a little comparison to the live site. Note that these are all small sample sizes - generally in the 15-25 range.
I absolutely agree with this, with the tiny caveat of the following scenario:
Assuming a 3*AB-AM/ABarr shatter combo rotation that ignores whether or not MBAM procced, looks like it may be less DPS than a more convetional 3*AB/ABarr, MBAM when proc rotation. That is to say, if you're casting AM/ABarr shatter (MBAM or not) after 3 ABs as a fixed rotation, and failing the ABarr shatter, then you're probably not making any headway.
Strictly, what you mentioned, that ABarr is never a DPS loss, is correct. Technically what's causing the DPS loss in this scenario is the effective ignoring/missuse of MBAM and/or Non-MBAM AM.
Hm, I did an excel-sheet calculating the 3*AB-AM/ABarr shatter combo rotation with rawr numbers as base and it gave about 10% more dps than 3*AB/ABarr, MBAM when proc rotation. That calculation was however without the arcane blast glyph but with 100% succesfull combo.
I think this make sense also since the AM has a lot higher Coefficients than Abarr. Thus it should always give you more to ramp up AM than Abarr. About 60% of the rotations would also get the MB proc (0.8 ^ 4).
I would really like to see this rotation in one of the expert simulators (rawr, magecraft) though since I'm not sure I got all aspects right.
With an Arcane spec, you give all your healers a free flask, which I consider quite a lot.
Is there a new amplify magic thread? Or it is possible to make a new one in the mage forum? The old reck/amp magic thread really helped a lot in progress.
01/15/09, 8:35 PM
Similar Threads
