But what about the following for a great MPS rotation:
ABar,ABx3(CC:AM)
which means cast ABar, ABx3 and repeat. If you ever see ClearCasting proc, toss AM and start over. Because CC is 10% chance and MB is 20% chance, this will usually result in a CCMBAM.
Can anybody calculate the DPS and MPS of ABar,ABx3(CC:AM)? Kavan?
Which dps rotation is the best completely depends on your mana regen. In general you should view the range of cycles in the following way. Based on active regen like evocation, gems, pots find the sustainability level. Using that see where in the range you fall, your optimal cycle pair will be the one immediately above and the one immediately below the sustainability barrier. The only exception to this comes when you add in cooldowns in which case the simplified model can no longer give you everything you need to know and it's no longer the case that only two cycles are optimal.
Can you expand a bit on the CC proc? The reason why I omitted it from the model is that if it procs you won't know until during the next spell. And it will be consumed by that spell so the only way to react to CC is by cancelling casts. Is that what you are suggesting?
Which dps rotation is the best completely depends on your mana regen. In general you should view the range of cycles in the following way. Based on active regen like evocation, gems, pots find the sustainability level.
This is where I am falling short. Kavan would you mind elaborating (Or pointing to a post which explains it if one already exists) on how one finds their sustainability so they know where they fall in the cycle "table"?
I have taken a look at the dynamic cycle post and even though I have studied Markovs chains and understand quite a bit of what's in there I still got a bit lost. I am correct that it is the right post to look at or am I completely out and about here?
I want to run some tests on this, but I've noticed behavior like he's talking about. The Potency buff definitely disappears when missiles starts, and I don't seem to see an increased crit rate (should be over 60% easily) on CC'd AMs. Definitely need some good numbers on this.
If this is the case, though, I'd think that you would probably want to cast another AM following any AM that procs CC. The potency buff is gone either way when AM finishes, and the most damage you're going to get for no mana from that CC should be an AM cast. Can't play around with Rawr from work, but might be interesting to see if the mana savings from that could increase overall DPS in the same way manly is talking about with the ABx2 cycles.
Edit: Just saw Astraylian's response. That sounds exactly like what I was assuming was happening.
If that is in fact the case, then you wouldn't ever have Arcane Potency up after an AM (Astraylian says the last missile clears it). Therefore, the only time you will be able to react to a CC is after ABarr. If what you say is true, AM clears Arcane Potency on the original cast (negating it's effect), so you wouldn't want to cast AM again, and since ABarr was just used, you can't use ABarr again; this leaves only AB to cast which makes the whole discussion irrelevant because you will end up continuing your rotation anyway.
Which dps rotation is the best completely depends on your mana regen. In general you should view the range of cycles in the following way. Based on active regen like evocation, gems, pots find the sustainability level. Using that see where in the range you fall, your optimal cycle pair will be the one immediately above and the one immediately below the sustainability barrier. The only exception to this comes when you add in cooldowns in which case the simplified model can no longer give you everything you need to know and it's no longer the case that only two cycles are optimal.
Can you expand a bit on the CC proc? The reason why I omitted it from the model is that if it procs you won't know until during the next spell. And it will be consumed by that spell so the only way to react to CC is by cancelling casts. Is that what you are suggesting?
Kavan,
Yes, by ABar,ABx3(CC:AM) I am suggesting the usual Abar, ABx3 rotation with careful attention to CC procs instead of MB procs. When CC procs, tap forward then AM, then start the cycle over again. MB will be ticking down more often then not.
This proposed rotation seeks to optimize MPS by using CC for the most expensive (and largest damage dealing) spell. The slight loss in DPS due to reaction time is a given, but this is not meant to optimize DPS, but MPS.
What I hate is seeing CC proc on ABarr and then wasting it on the first AB. This rotation and attention to CC procs would prevent such a waste of the best mana saving proc.
I never had that much trouble with mana in 25mans doing ABx3-Abarr (mbam-abarr with 0 or 3 stacks). Is the only reason for ABx2-Abarr and ABx3-mbab-Abarr on procs only because of mana or does it actually do better dps?
Is it commonplace to not cast mbam-Abarr with 0 stacks anymore, but to wait and cast it on ABx3?
Sorry but rotations have been switching back and forth lately and has me questioning some things.
I've been using Simulationcraft with Wowhead to determine the dps increase of each piece of gear, but I'm starting to think its inaccurate with Arcane. I noticed that the rotation is hard coded to use Mage Armor for 57/3/11. Does that mean the scale factor for spirit (currently 0.61) is more inflated than it should be on fights where I'm using Molten Armor? I always thought 0.61 was awfully high compared to Intellect's 1.09, considering the many more benefits Int gives you over Spirit.
Would a simple calculation of 0.61 * (0.3 / 0.8) approximate Spirit's worth when using Molten Armor?
I can confirm this "bug" (I would call it a bug) with CC. When I gain the CC and Arcane Potency buff, casting arcane missiles consumes both. The AM is free, but the Arcane Potency is removed and the missiles gain no benefit from it. If AM proc's CC and Arcane Potency on it's own, however, (via the initial cast) Arcane Potency will affect the individual missiles giving them the increases crit and at the end of the spell the Arcane Potency buff is removed. The actual CC, however, persists to the next spell.
This is where I am falling short. Kavan would you mind elaborating (Or pointing to a post which explains it if one already exists) on how one finds their sustainability so they know where they fall in the cycle "table"?
I have taken a look at the dynamic cycle post and even though I have studied Markovs chains and understand quite a bit of what's in there I still got a bit lost. I am correct that it is the right post to look at or am I completely out and about here?
Well you can't really use that table explicitly since it's assuming numbers for a specific gear set and regen options. I think it's best to understand the relations between those cycles to be able to make decisions on the fly. For example if your evocation gets interrupted and you're suddenly in a mana starved situation you should know what options you have available and pick appropriately to keep you up until next regen cooldown. In the end my answer would be to use Rawr to get an answer for your specific case as it will take more things into account than the simple model used to predict the optimal cycle pairs (there should be a new release soon with all the new cycles).
If that is in fact the case, then you wouldn't ever have Arcane Potency up after an AM (Astraylian says the last missile clears it). Therefore, the only time you will be able to react to a CC is after ABarr. If what you say is true, AM clears Arcane Potency on the original cast (negating it's effect), so you wouldn't want to cast AM again, and since ABarr was just used, you can't use ABarr again; this leaves only AB to cast which makes the whole discussion irrelevant because you will end up continuing your rotation anyway.
No, you misunderstand me. I'm saying that because you don't have potency due to the first AM consuming it, this is the only time you can cast AM on a CC without it being a bad idea from wasting potency. Thus, in that situation, the only gains possible are from either using our most expensive spell mana-free or (if possible) clipping a barrage to receive the Potency buff rather than the last missile.
I've been using Simulationcraft with Wowhead to determine the dps increase of each piece of gear, but I'm starting to think its inaccurate with Arcane. I noticed that the rotation is hard coded to use Mage Armor for 57/3/11. Does that mean the scale factor for spirit (currently 0.61) is more inflated than it should be on fights where I'm using Molten Armor? I always thought 0.61 was awfully high compared to Intellect's 1.09, considering the many more benefits Int gives you over Spirit.
Would a simple calculation of 0.61 * (0.3 / 0.8) approximate Spirit's worth when using Molten Armor?
Someday soon I'll make it easy for people to push single-player profiles through the sim so that they can easily get custom scale-factor generation.
Until then, Rawr.Mage is by far your best avenue to personalized analysis........
Well it looks like I have some explaining to do. It appears I had a typo when I was typing in the transition matrix (the meaning of the typo was that it treated barrage procs at 1 stack as not yet visible, even if they should be, so there wasn't any error in cast distribution, just in spell selection logic). It didn't have any effect on the top optimal cycles, but it changes optimality in the lower mps range. Luckily I found this out right in time to add the missing optimal cycles to 2.1.9 release of Rawr (like almost not in time). Additionally I added a button in the advanced section that will compute the optimal range of cycles for your gear (using the control string notation I described before) so you can check if maybe some cycle is optimal for you that is not currently in the list used by Rawr.
The optimal cycles are (using the new Rawr notation, kind of a mix of some suggestions from here):
ABSpam03C (AB spam, MBAM-ABar at 0 or 3 stack): 5028.084 dps, 412.3207 mps
ABSpam3C (AB spam, MBAM-ABar at 3 stack): 5013.697 dps, 395.6343 mps
AB3ABar3C (ABx3-ABar, MBAM-ABar at 3 stack): 4801.682 dps, 237.0959 mps
ABABar3C (AB-ABar, MBAM-ABar at 3 stack): 4668.696 dps, 181.7985 mps
ABABar2C (AB-ABar, MBAM-ABar at 2 stack): 4566.665 dps, 145.9942 mps
ABABar2MBAM (AB-ABar, MBAM at 2 stack): 4500.556 dps, 128.9478 mps
ABABar1MBAM (AB-ABar, MBAM at 1 stack): 4445.778 dps, 116.023 mps
ABAM (AB-AM): 3743.657 dps, 73.89004 mps
And here's all unique cycles if someone wants to compare:
Luckily I found this out right in time to add the missing optimal cycles to 2.1.9 release of Rawr (like almost not in time).
He's not kidding, either. I was mid-upload of 2.1.9, when he said 'Wait!'. Anyway, CodePlex has been doing some maintenance this morning, so if it takes a while for the Rawr site to load, and for the new version to download, that's why; please just be patient.
ok, i like thinking of damage rotations as a big if-else statement. Maybe it's my background in programming, but I just think better this way. So help me make the best DPS arcane if-else statement!
We're just going to write it without taking into account DPS cooldowns like POM, Arcane Power, Icy Veins, trinkets, Managem, and heroism. We also won't take into account innervate from druids.
If AB_debuff = 0
If ABar is off cooldown
cast ABar
else cast AB
If AB_debuff = 1
cast AB
If AB_debuff = 2
cast AB
If AB_debuff = 3
If MBar = active
cast AM
else If Evocate is off cooldown
If Mana < 40%
cast Evocate
else cast AB
else cast ABar
My only real question about the rotation is what if you have 2 stacks of the AB_Debuff and you're in the middle of casting your third and MBarr has yet to proc. I know this was mentioned earlier in this thread but I can't seem to find a definitive answer.
The 2 options I see:
After the third Arcane Blast lands, wait a split second to see if MBar procs. If it does, cast AM, if not, cast ABarr.
Just Cast ABarr and if MBar procced, oh well, just wait until you get ABx3 and cast AM.
Do we have the proof of which is better? I'm thinking of just going with the 2nd because its easier to pre-empt it.
Yey!! Someone posting C programming Let me make it look better :P (with all due respect ofc)
P.S. And If you are casting 3rd AB and MBar is not procced yet, Kavan suggested best rotation as 3rd AB (MBar Procced)-> ABarr (didnt have the time to react) -> AB -> AB -> AB -> MBAM -> ABarr -> AB -> AB .... P.S.2 Should we evocate always if Mana < 40% and Evocate is not in CD? I think Evocation should rather be chosen by instinct.
Still have some problems with Arcane, but maybe because we realy have laggy Raids after the last 'improvements'. On most Bosses in WWS there are 5-7 Melees/Hunter top DPS, only after this we have a WL and me fighting for positions (Last Patchwerk 4200 DPS - 2 Weeks before with FFB i hav 4.7k). But meanwhile i change a little bit my equipp (More Haste/Spirit/MP5 instead of Hit/Crit).
One simply question: what class is best to give Focus magic? FFB mage? Las time i try with a Enh Schaman, it was pretty well.
I think the hardest part is still setting up mana regeneration for longer fights. But with the recent posted sugestion of using AB2Abarr i think most of us will get on it and have a good base for developing their propper rotations.
I think the hardest part is still setting up mana regeneration for longer fights. But with the recent posted sugestion of using AB2Abarr i think most of us will get on it and have a good base for developing their propper rotations.
Actually, according to what Kavan just posted, AB2ABar3C (Rotate ABx2-ABar, on proc, stack AB to 3 before MBAM-ABar) is no longer an optimal rotation. Instead if you can't manage AB3ABar3C you should be doing ABABar3C.
On a different subject, how are people generally starting off fights? I keep seeing people saying to save only evocate on the last tick of icy veins, but you also want to use icy veins as soon as possible. So are you spending the start of the fight doing the max dps combo, then at ~40% mana IV-AP-Gem, then use evocate and switch to a more normal cycle?
One simply question: what class is best to give Focus magic? FFB mage? Las time i try with a Enh Schaman, it was pretty well.
That's a very good question. It really depends if you're optimizing on one of two possible scenaria:
1) You want to have the maximum uptime for you.
2) You want the maximum Raid benefit.
In the first case, I'd suspect an E-shaman would be optimal, due to their higher cast-per-second than a FFB mage, though how much they crit to begin with may change this. Ultimately, you want to know how many crits-per-second the class generates in order to evaluate it correctly. However, an E-shaman critting will increase uptime of Elemental Oath, so now the DPS-return becomes dependent on how their buff influences your DPS too... Extremely complex.
Ultimately, I prefer to give it to whoever's struggling (for good reason that is, not whoever sucks hardest). Both E-shamans and FFB mages gain efficiency from the Crit so whoever is in the worst state with respect to mana would be my choice. In pure DPS terms, as in "who gains more from 3% crit", I'd be surprised if anyone surpassed the FFB mage, purely due to crit size modifier. Having said that, increasing a 60% crit to a 63% is a lot less increase than 45 to 48%.
Strictly, it's very hard to quantify the second, because the possibility of giving it to a healer makes it impossible to estimate. Shamans, Priests and Palas will all benefit in efficiency and mechanics with more crit, but how can you compare HPS and/or healer mana return to DPS increase?
Long story short? I'd give it to the Elemental Shaman. For an AoE encounter, however, I'd prefer a fellow mage. I'd also prefer a fellow mage (specced MoE) if the raid-setup was such that he'd be close to being mana-starved.
I'm still wondering how AB-spam MBAM-ABarr on 0 or 3 stack is a dps increase over AB-spam MBAM-ABarr on 3 stack only when ABx3ABarr MBAM-ABarr on 0 or 3 stack is a dps decrease to ABx3ABarr MBAM-ABarr on 3 stack only.
I wonder if someone could play around with Rawr and figure it out. I won't be off work for another 8 hours so I can't toy with it for a while. I can't think of anything intuitively to support this indiscrepancy; every stat should scale equally between them and the differences between the cycles and cast distributions only pertain to why AB spamming until MB procs is more dps than resetting your your stack with ABarr. It's actually bothering me a lot now.