Higher haste level meaning what, exactly? You can always get two CL's in between Lava Bursts (3.5s fits twice into 8 seconds). I've pointed out where you run into the cooldown, and you can work around it there. Nobody's going to force you to use a certain rotation and yes, rotations change as you reach higher levels of haste. But right now, with 15-20% haste in a boss fight, this is the best way I've found to consistently squeeze an extra nuke into the Lava Burst cooldown.
With the new changes coming to the new talent "Shamanism" Lightning Bolt will now be the higher damage spell even considering cast times.
The only time I see using CL on single target now would be if I had 1-1.5 seconds until Lava blast cooldown is up and wanted to get off the highest damage spell I could in that window.
@Ghando
No you can't put 2 CL cooldowns into 8 seconds
Even with enough haste you have 1 sec casttime for CL, after that, 3.5 secs cooldown, then next CL casttime of 1sec and again 3.5 secs cooldown - gives a total of 9secs - or: CL cooldowns overlaps the end of the LvB cooldown.
You have just to many cooldowns to consider. once you hit a cooldown there - and scaling with haste drops to ground. if youn put a spell more in, your scaling raises, but is on lower baselevel. with a rotation with just a CL you don't have this problem. you have only the 2 LvB cooldowns to consider. As the graphs shows, for every haste level above 6% there is a better 1 CL rotation than a roation with more CL.
And still: i don't believe you when you say you can focus on the fight and make your rotation as you describe. even if you do so as you say, you usually needs dozend or hunderts miliseconds to decide wich spells you take and this causes also "lag"
i mean, i can do 1,2,3 and mousewheel spamming. till i see LvB CD is finished and give real focus on bossfight and you do keep your attention only on CL and FS cooldown. my mousewheel spam reduces my lag between the spells nearly to zero
With the new changes coming to the new talent "Shamanism" Lightning Bolt will now be the higher damage spell even considering cast times.
The only time I see using CL on single target now would be if I had 1-1.5 seconds until Lava blast cooldown is up and wanted to get off the highest damage spell I could in that window.
That's what you're doing anyway. CL is higher DPS than LB now because their relative damage is a wash and CL is faster to cast. With Shamanism the spells will be a total wash, and the only advantage will be CL's shorter cast time. It's also worth mentioning that because CL is an important part of Elemental DPS (keeping us competitive on multiple targets) it might end up being included in Shamanism. In any case, the real point of putting CL into your cycle is to buy some time to squeeze an extra nuke in between Lava Bursts.
And Mitrandir, you're probably not trying to be obtuse but you're coming off that way. No shit, you can't get 2 CL cooldowns (3 casts) into 8 seconds. But you can get two casts (one cooldown) into one cycle, which is exactly what I said. I get that your cycle will change based on haste and at the ludicrous levels of haste you project to, things change even more. But in general you should be casting CL in lieu of LB every time the cooldown is up, to free up as much extra time as you can.
Also of note: with the changes, comparing 2LvB/FS/10LB with 2LvB/FS/7LB/4CL, you gain ~170 dps, at the cost of ~550 mps.
That's a 3.69% output increase for a 67% cost increase, assuming single target for CL.
The average rotation dps is still less then the (...) and therefore a shorter wait time is better.
I'm not sure what you mean here. I feel we have a bit misunderstanding, so just to clarify; there is two option: Either I don't wait, or wait for LvB cooldown. I say wait, when the time left on LvB cooldown is shorter than the solution of the equation. Of course if - due to your current haste and the used rotation - the CD left on LvB is shorter, don't wait that amount, shorter wait is better. I did not say wait that much amount before LvB all the time, there's nothing smart in "do this all the time".
Or, solving the problem from a different approach: make 2 spell rotations, one that does not wait, other that waits if it cannot fit another LB before LvB CD. The max DPS is the max of these two's DPS, and the break point is what I wrote up there, according to my spreadsheets.
There is a lot of discussion on the first rotation and how it should work, etc. I don't see much about the 2nd rotation, where LvB won't be the 2nd spell you cast, or necessarily even the 3rd. Most of these projections seem incomplete because of this.
FS-LvB-LBx3-CL-LvB-LBx2-FS
CL-LB-LvB-LBx3-CL-LVB-LB-CL-FS
LB-CL-LvB-LBx3-CL-LvB-LBx2-FS
The basic point is that these things can get sort've confusing, and it'd help to have some basic rules.
Ergo: FS always up, LvB always on CD, LB fill unless less than 1.5 secs on LvB cooldown or FS ticks, then CL (When that is also not on CD).
In my mind, it seems always more useful to have a spell on the way out rather than waiting on a CD. You never know when you'll have to move again, and then you'd have lost the move time + the time you figured you'd just wait on that CD.
Also, multi-target damage is nice, but personally, trash damage isn't that important in the face of mages, locks, and DKs, so throwing out as many CLs as you can in the face of the mana cost seems silly. You won't ever keep up with them in aoe situations, and given the quickness with which things die on trash, I find old school LBx3/CL spam to be more productive. Though maybe thunderstorm can put the lie to that. Another 8% mana every 45 seconds is actually pretty strong.
That's what you're doing anyway. CL is higher DPS than LB now because their relative damage is a wash and CL is faster to cast. With Shamanism the spells will be a total wash, and the only advantage will be CL's shorter cast time. It's also worth mentioning that because CL is an important part of Elemental DPS (keeping us competitive on multiple targets) it might end up being included in Shamanism. In any case, the real point of putting CL into your cycle is to buy some time to squeeze an extra nuke in between Lava Bursts.
And Mitrandir, you're probably not trying to be obtuse but you're coming off that way. No shit, you can't get 2 CL cooldowns (3 casts) into 8 seconds. But you can get two casts (one cooldown) into one cycle, which is exactly what I said. I get that your cycle will change based on haste and at the ludicrous levels of haste you project to, things change even more. But in general you should be casting CL in lieu of LB every time the cooldown is up, to free up as much extra time as you can.
I think LB will do more DPS for most of us by the time these changes go into place.
LB is 765 Base Damage with a 2 second cast time. Base .7142+.1(Shamanism)=.8142
CL is 1042 Base Damage with a 1.5 second cast time. Base Coefficient= .57142
I'm showing LB as recieving .2428 more base damage per SP than CL with a 277 base damage disadvantage which means they are equal damage at 1141 spellpower.
a 33% shorter cast time gives Chain lightning a 33% bonus to base damage giving it 1389 base damage or a 624 base bonus over lightning bolt. It would therefor take 2570 spellpower to completely remove CL from the damage rotation.
If you have totem of Hex you can reduce this to 2385 Spellpower which is what I'm sitting at raid buffed now. I'm expecting this patch to go in with the arenas some time in December so by then we can kiss the necessity of using CL goodbye except when we have between 1-1.5 seconds left until LvB cooldown is up.
I think LB will do more DPS for most of us by the time these changes go into place.
LB is 765 Base Damage with a 2 second cast time. Base .7142+.1(Shamanism)=.8142
CL is 1042 Base Damage with a 1.5 second cast time. Base Coefficient= .57142
I'm showing LB as recieving .2428 more base damage per SP than CL with a 277 base damage disadvantage which means they are equal damage at 1141 spellpower.
a 33% shorter cast time gives Chain lightning a 33% bonus to base damage giving it 1389 base damage or a 624 base bonus over lightning bolt. It would therefor take 2570 spellpower to completely remove CL from the damage rotation.
If you have totem of Hex you can reduce this to 2385 Spellpower which is what I'm sitting at raid buffed now. I'm expecting this patch to go in with the arenas some time in December so by then we can kiss the necessity of using CL goodbye except when we have between 1-1.5 seconds left until LvB cooldown is up.
If I follow your calculations correctly, you have accounted for the increased base damage through the faster cast time but have not accounted for the increased SP damage from the faster cast time (i.e. CL gets 0.57xSP per 1.5 secs while LB gets 0.81xSP per 2 secs). So whilst the coefficient for SL is only 0.57142 of spell power, in the same cast time as a lightning bolt (obviously averaged over many casts) it will receive 0.57142 x 1.33 = 0.76 SP.
This would mean that the break even point of the two spells is substantially higher.
My calcs show the break even at around 4700 SP, although you would presumeably drop the CL a bit before then due to mana efficiency.
After reading all of these posts, I think I have more questions then answers, but I suppose it is the nature of these things. I do have an idea to submit to the group, I'd like to turn the problem on it's head. Rather then asking what rotation is best at a given haste, why not instead ask what haste value should be maintained to optimize a given rotation.
Three casts of LvB can be effectively fit within the dot timer of FS (when glyphed). But, casting the 3rd will greatly affect the next rotation, as LvB will not be off cooldown by the time it comes up in the next rotation. In my mind, this creates a a static environment of 2 LvB casts per FS, then fit as many CL & LB casts possible within the cooldown of LvB. Rather then just equipping as much haste as possible, perhaps instead we should focus on a specific haste number to optimize that rotation. This has the added benefit of creating a natural gear weighting: hit to cap > haste to optimum > crit or spellpower to taste
The rotation I suggest is:
FS,LvB,CL,LB,LB,CL,LB,LvB,CL,LB,LB,CL,LB
This rotation would require 12.5% haste to be achieved. Assuming the presence of WoA, that would mean a requirement of 246 haste rating through gear. Though, pushing up over 250 might help offset latency and such.
After reading all of these posts, I think I have more questions then answers, but I suppose it is the nature of these things. I do have an idea to submit to the group, I'd like to turn the problem on it's head. Rather then asking what rotation is best at a given haste, why not instead ask what haste value should be maintained to optimize a given rotation.
<snip for brevity>
Thoughts? Am I totally off base?
I think it's a valid question but even with only the WoA down you will be struggling to get your haste as low as that I think. Then you add in 3% from the Boomkin buff, procs on trinkets (or possibly totems), BL and movement during fights and you really get into problems trying to maintain a consistent rotation. Better i think to use a priority system of casting: FS when the DoT runs out, LvB when off CD, CL when off CD, LB as a filler. You do need to keep an eye on CD timers as sometimes it will be better to use the LB rather than the CL (if there is, for example. 2 secs left before LvB comes off CD).
I think it is these variations in cast time etc. that lead the discussion to the best rotation at certain haste levels rather than the best haste for a given rotation as you have suggested - you almost need several rotations up your sleeve to cover the most probable eventualities.
This priority system over a rotation is exactly the conclusion that came out of the shadowpriest community in BC when they had to juggle varying DoTs and short duration CDs. To even consider we could really keep a solid rotation down in a real raid environment is completely naive. It is the extreme minority of fights that have literally zero movement, and our longest cooldown being 8 seconds makes even the smallest movement completely toss any sort of timed rotation out the door.
The bigger question isn't really whether we should be using a rotation or a priority system (its essentially the same thing, just one has it plotted down) but when we should hold our rotation for a higher priority spell. I've been having a helluva time trying to figure out the right way to work that out. It seems like there are plenty of gear levels that warrant waiting as much as .3 seconds on a LvB over a LB, but once you factor in human reaction time in altering the rotation, it tosses the benefit out the door. Our class has become a lot more intuitive than it used to be with the discretion that has been thrown into our rotation.
I've also been having trouble deciding when to replace my T6 bonus, I came up with a pretty simplistic value of about 140 spell power to break the 4 pc, but that isn't possible to achieve without breaking the 2 pc as well, which bumps the number up even higher. (5% base damage + 45 SP, 35 crit) Has anyone come up with good rule of thumb on that, or even broken it yet?
I tend to use kind of a priority system aswell, casting LvB each time i can and holding FS up at all times.
On bosses where there are just one thing that can be hit i dont use CL that often, or atleast not every cooldown as it dont seem effetive enough compared to the increased mana costs
I will probably keep 4p t6 untill i get 2xt7, unless i get something that is such a upgrade that i outdoes the 5% bonus(which most likely wont happen)
Using my current (3.0.4) EP values, I'm estimating 4pc T6 is worth 398 (SP is 1.142 for comparison). See: here
Sorry but just to clarify, is that the EP for both the 4 pc and the 2 pc bonus or just for the 4 pc? It seems about right for both bonuses with very good gear but very high for the 4 pc alone.
If I follow your calculations correctly, you have accounted for the increased base damage through the faster cast time but have not accounted for the increased SP damage from the faster cast time (i.e. CL gets 0.57xSP per 1.5 secs while LB gets 0.81xSP per 2 secs). So whilst the coefficient for SL is only 0.57142 of spell power, in the same cast time as a lightning bolt (obviously averaged over many casts) it will receive 0.57142 x 1.33 = 0.76 SP.
This would mean that the break even point of the two spells is substantially higher.
My calcs show the break even at around 4700 SP, although you would presumeably drop the CL a bit before then due to mana efficiency.
I'm not sure what you mean here, but I think you are trying to to add the bonus to chain lightning twice.
You can increase the CL damage by 33% either through base, or adjusted spellpower base but not both. It is a flat 33% increase unless I am misunderstanding something you are saying.
I balanced my equation around an equal cast time of 2 seconds(which is why CL got a 33% increase). The spell power coefficient is added per CAST and is not changed or affected by any haste or anything, it is merely added to base damage based upon spellpower.
The Balance point of CL and LB after this next patch will be 2570 Spellpower not including totem of hex or void.
I'm not sure what you mean here, but I think you are trying to to add the bonus to chain lightning twice.
You can increase the CL damage by 33% either through base, or adjusted spellpower base but not both. It is a flat 33% increase unless I am misunderstanding something you are saying.
I balanced my equation around an equal cast time of 2 seconds(which is why CL got a 33% increase). The spell power coefficient is added per CAST and is not changed or affected by any haste or anything, it is merely added to base damage based upon spellpower.
The Balance point of CL and LB after this next patch will be 2570 Spellpower not including totem of hex or void.
LB is 765 Base Damage with a 2 second cast time. Base .7142+.1(Shamanism)=.8142
CL is 1042 Base Damage with a 1.5 second cast time. Base Coefficient= .57142
Lets try this math with your numbers.
Base damage per second:
CL: 1042 / 1.5 = 694.66
LB: 765 /2 = 382.5
I balanced my equation around an equal cast time of 2 seconds(which is why CL got a 33% increase). The spell power coefficient is added per CAST and is not changed or affected by any haste or anything, it is merely added to base damage based upon spellpower.
I can see where you have normalised the base damage per cast but not the coefficient. The coefficient is a percentage of spell power per cast and if cast times are different then the damage added due to coefficient needs to be normalised appropriately.
Pitbuller's calcs do this and whilst his result differs from mine because my spreadsheet has a 'lag adjustment' included and I forgot to remove the T6 4Pc bonus (both of which bring the break even point down), I believe his method of calculation is correct and the result accurate.
LB is 765 Base Damage with a 2 second cast time. Base .7142+.1(Shamanism)=.8142
CL is 1042 Base Damage with a 1.5 second cast time. Base Coefficient= .57142
Lets try this math with your numbers.
Base damage per second:
CL: 1042 / 1.5 = 694.66
LB: 765 /2 = 382.5
Well, before we get to per second, lets expand out the spell damage
LB: (765 + 2.5/3.5*(SP*1.1))*Concussion*Overload = 884 + 0.9075SP => 442 + 0.45375SP per second
CL: (1024 + 2/3.5*SP)*Concussion*Overload = 1204 + 0.66SP => 802 + 0.44SP per second.
0.45375 - 0.44 = 0.01375 SP per second
802-442=360
360/0.01375=26182 spellpower.
Solving for SP:
LB: 12322
CL: 12322
So you were right that it won't be attainable in WotLK, but you were wrong as to what the figure was. It's important to note that Shamanism isn't a 10% bonus co-eff, but rather 10% extra spellpower for that spell.
Those breakeven points assume 0% haste, which is never going to happen in a raid situation. Every point of haste brings the breakeven point down as the two cast times get closer and closer together due to the effect of haste impacting the real cast time of the longer cast spell on a greater level than the lesser. Your formula is also not accounting for LO, which is also scaling faster on LB than CL.
I may have some flaw in my calcs as well, however.
SP = Spell Power
Co = Coefficient
LO = Lightning Overload
@ 2800 SP & 500 Haste rating, and including in 8% haste from buffs I have CL only beating LB by about 14%. When you factor in the high cost, it starts to look pretty shaky once Shamanism goes in. These numbers also assume perfect rotations, and the more variables you put in are likely to cause human error (both in rotation and play.) Even at 70 we were showing a lot of players doing more DPS cutting CL at high gear levels due in part to these factors, as the difference gets even smaller it seems the reward will diminish even further.
So you were right that it won't be attainable in WotLK, but you were wrong as to what the figure was. It's important to note that Shamanism isn't a 10% bonus co-eff, but rather 10% extra spellpower for that spell.
I think the exact mechanics for the the Shamanism talent are still open to interpretation. Unless I have missed some information, all we have had is a comment saying that "Your Lightning Bolt and Lava Burst gain an additional percentage of your bonus damage. We have not finalized the numbers, but it will probably be something like 10% for Lightning Bolt and 20% for Lava Burst at max ranks."
You have interpreted it as 10% of the bonus damage applied to your LB, Pitbuller and I have interpreted it as 10% of your total bonus damage. Either could be correct.
Those breakeven points assume 0% haste, which is never going to happen in a raid situation. Every point of haste brings the breakeven point down as the two cast times get closer and closer together due to the effect of haste impacting the real cast time of the longer cast spell on a greater level than the lesser. Your formula is also not accounting for LO, which is also scaling faster on LB than CL.
If we don't use lag when we counting cast times then haste scales equally with CL and Cl. Haste is percentage modifier. LO is percetange modiefier to or CoE. We dont' have to account any equal modifiers. It's just make math mrope complicated that it is.
Let's say that CL do 150damage and Lb do 200damage.
Also of note: with the changes, comparing 2LvB/FS/10LB with 2LvB/FS/7LB/4CL, you gain ~170 dps, at the cost of ~550 mps.
That's a 3.69% output increase for a 67% cost increase, assuming single target for CL.
I just wanted to throw out a comment here.
I am planning a FS,LvB,CL, LBx(however many you can squeeze in until LvB is almost back), LvB,CL, LBx(until LvB is almost ready again) Repeat.
Doing this gets the added DPS from CL, but it will be garunteed to be at 40% reduced cost due to the LvB crit.
So really the best of both worlds, slightly better DPS, and not as heavy on the mana as postulated above, not to mention much more simple of a rotation to keep up with.
I think ill plug in some haste ratings to find where the sweet spots are for this rotation. Although I do hate having to plan for specific amounts of haste...
So looking at haste 100 haste rating at 80 is about 3.1% haste +wrath of air + boomkin = ~11.1% haste
a rotation of FS,LvB,CL,LBx2,LvB,CL,LBx3 = 17.58 seconds, +40ms lag per cast (too much?) = 17.98 seconds
100 haste is low low low... but hey it works perfect in this rotation
To add in another LB to the rotation we need about 475 haste rating or about 22% haste with totem and moonkin added in.
FS,LvB,CL,LBX3,LvB,CL,LBx3 with 40ms lag = ~18.01 second rotation
So yeah pick your poison 100 or 475 haste rating... on second thought, nevermind back to the drawing board.
Doing this gets the added DPS from CL, but it will be garunteed to be at 40% reduced cost due to the LvB crit.
But will it? Clearcasting will only proc when you crit, which you won't have done yet when you start casting CL because LvB won't have landed yet due to it's "air time". Much better to FlS>LvB>LB>CL>LB(til LvB is up) to be sure.
I agree with Bink that on single target, LB is the better choice in terms of MPS vs DPS. I don't have any of the new and excellently itemised MP5 gear yet so I'm having troubles with continual AOE pulling in 5mans without any kind of replenishment buff and not being allowed to use Thunderstorm for fear of blowing mobs out of the AOE if I'm a bit too close...
Anyway... If/when I use CL, I'll make sure I have clearcasting up first
11/26/08, 4:02 PM
Similar Threads
