If No totemsdwn then drop totens
Else If no FSDot And no FScooldown Then FS
Else If Heroism And No EMCooldown Then Elemental Mastery
Else If Exhaustion And No EMCooldown Then Elemental Mastery
Else If FSDottime < LvBCasttime and LvBCooldown < FSCasttime then FS
Else If FSDot And no LvBCooldown Then LvB
Else Lightning Bolt
The line "Else If LvBCooldown < (LB Casttime minus ~0.3s) And No Heroism Then CL" may be useful, but I really question the practical application of something like this. Who can make the split second choice to use CL to save that extra three tenths of a second? (Unless you're operating on a fixed rotation where you've mapped everything out beforehand)
Not really a rotation per se but a little tip that has helped me in my rotations. I changed from a loooong time nurfed UI and nurfed clone to XPERL unitframes. There is a little feature that shows your personal debuffs/and dots as a much larger icon with a countdown... its much easier to see and if you haven't tried it I am sure you will like it. I haven't missed a FSdot/Lava burst hardly at all since this.
If No totemsdwn then drop totens
Else If no FSDot And no FScooldown Then FS
Else If Heroism And No EMCooldown Then Elemental Mastery
Else If Exhaustion And No EMCooldown Then Elemental Mastery
Else If FSDottime < LvBCasttime and LvBCooldown < FSCasttime then FS
Else If FSDot And no LvBCooldown Then LvB
Else Lightning Bolt
The line "Else If LvBCooldown < (LB Casttime minus ~0.3s) And No Heroism Then CL" may be useful, but I really question the practical application of something like this. Who can make the split second choice to use CL to save that extra three tenths of a second? (Unless you're operating on a fixed rotation where you've mapped everything out beforehand)
your spreadsheet days it hurts me but william shatner says it's necessary. Now who should I believe? Binkenstein or William Shatner?
It's a question of theoretical vs practical. Theoretically, casting CL when you have 1.2 seconds until LvB comes off cooldown would be a DPS increase over just casting LB, but in practice telling the difference between 1.2 seconds and 1.6 seconds on the LvB cooldown is harder than just mashing 2 again - to the point where it's arguable whether or not it's even worth trying.
ok maybe not mecessary but you made it pretty clear that CL is still really useful in single target.
It depends on your haste rating, and the differences can seem pretty small. For example, if we have a Moonkin I'll use 5 LBs in between LvBs because it fits perfectly. Without a Moonkin I need to throw in a CL to keep things airtight. I know everyone's all about this "priority system" but it really doesn't make sense to take it to the level of a series of if/then logical statements. Bink's right when he says that you don't have time to effectively make the decision based on the rules you've set forth, but you shouldn't have to. You know how much haste you've got. You know whether you've got a Moonkin in the raid. I suppose you can't account for latency on any given evening but it's really not difficult to figure out, given your particular haste rating, what your rotation needs to be. Mine is "set in stone" in the sense that I know my ideal rotations with and without a Moonkin and I'll stick to those during a fight.
Exceptions and difficult decisions start to crop up once you add things like movement and adds into a fight, but those are hard to quantify or theorycraft on a napkin. Even so, I know what my ideal rotation is and I account for other circumstances by thinking of them as deviations from the ideal. If I'd be using 5 LBs in a given cycle and I figure I spent 2.5s running, I'll cut one LB off in my head and substitute CL for another one. It doesn't have to be perfect; anyone with enough game experience should be able to handle those decisions.
Would it be viable to get enough haste rating for 33% without bloodlust and only Wrath of Air, making Lightning bolt and GCD a 1 second cooldown?
Would that make it possible to fit in 7 lb in between lava burst cooldowns and therefore doing great dps on bosses?
With that said, i dont plan on gimping my gear, was just wondering if its worth to get that much haster rating (about 1000) of it was possible. To maximse rotation.
Would it be viable to get enough haste rating for 33% without bloodlust and only Wrath of Air, making Lightning bolt and GCD a 1 second cooldown?
Would that make it possible to fit in 7 lb in between lava burst cooldowns and therefore doing great dps on bosses?
With that said, i dont plan on gimping my gear, was just wondering if its worth to get that much haster rating (about 1000) of it was possible. To maximse rotation.
33% haste would make GCD = 1.128 secs and LB cast = 1.504 secs. To make LB = 1 sec cast you need 100% haste.
Would it be viable to get enough haste rating for 33% without bloodlust and only Wrath of Air, making Lightning bolt and GCD a 1 second cooldown?
Would that make it possible to fit in 7 lb in between lava burst cooldowns and therefore doing great dps on bosses?
With that said, i dont plan on gimping my gear, was just wondering if its worth to get that much haster rating (about 1000) of it was possible. To maximse rotation.
Erm... 33% haste rating would give you 1.5s LBs, not 1.0s ones. The calculation is [base cast time] / (1+[haste modifier]), in your case 2 / 1.33 = 1.504s
Personally, though, I wouldn't go past 1.6s LBs (25% haste raid buffed) without Heroism. It gives you nice 5 LBs for each LvB cooldown and requires normally attainable 558 haste rating (in Ulduar). Even then, gemming is not worth it, as spellpower gives us more damage overall.
Creativity requires the courage to let go of certainties.
Erm... 33% haste rating would give you 1.5s LBs, not 1.0s ones. The calculation is [base cast time] / (1+[haste modifier]), in your case 2 / 1.33 = 1.504s
Personally, though, I wouldn't go past 1.6s LBs (25% haste raid buffed) without Heroism. It gives you nice 5 LBs for each LvB cooldown and requires normally attainable 558 haste rating (in Ulduar). Even then, gemming is not worth it, as spellpower gives us more damage overall.
A lot of us here, in full 25m gear, have the required haste with a Boomkin (or Ret Pally) to reach the 5 LB mark.
It is a very simplified rotation (and good thing because you can't "test" it on a training dummy) You will drop below the required haste for 5 LBs.
I totally agree with the above quote that one should not really go beyond that level of haste. During Bloodlust I am at 0.908 GCD / LvB. Unsure how much of a dps loss this is, would be relative to what percentage of the fight you benefit from Bloodlust. I would presume none for the moment due to how items are "budgeted". If we see more "two stat items" and can hold our present haste and increase SP / crit. Currently these are cloth and designed for Mages / Warlocks.
LvB cooldown = 8 seconds
8/5 = 1.6 LB second cast
2/1.25 = 1.6 LB second cast
25% haste required
25% haste - 5% WoA buff - 3% moonkin or ret pally buff = 17% haste required
33 haste = 1%
17%*33 = 561 haste required to fit 5 LB inbetween a LvB cooldown
That's not taking into consideration server lag, etc so something closer to 600 haste would be required.
Why are shaman who do not have 561 haste saying they can fit 5 LB inbetween their LvB? I'm not hating, just asking.
LvB cooldown = 8 seconds
8/5 = 1.6 LB second cast
2/1.25 = 1.6 LB second cast
25% haste required
25% haste - 5% WoA buff - 3% moonkin or ret pally buff = 17% haste required
33 haste = 1%
17%*33 = 561 haste required to fit 5 LB inbetween a LvB cooldown
That's not taking into consideration server lag, etc so something closer to 600 haste would be required.
Why are shaman who do not have 561 haste saying they can fit 5 LB inbetween their LvB? I'm not hating, just asking.
Haste is multiplicative not additive when dealing with different sources of haste. The only additive haste is haste associated with ratings.
16% Haste through gear(precisely what I have) will get you a theoretical 5 LB rotation.
1.16*1.05*1.03=1.25454
2/1.25454=1.594
Theory never works out actual numbers however as everyone experiences delay and latency to some degree.
Currently I am exactly able to model 4, 2 second casts and a 1.5 second cast inbetween each lava burst at 16% haste.
Since we can never remove flame shock from our rotations I think this is the ideal place to be. After chain Lightning gets nerfed however, it may end up being worth it to extend the lava burst cooldown by the extra half second every other rotation by continuing to cast lightning bolts.
Fully geared in the next tier of gear from 25 man ulduar we are likely going to be sitting at 20% haste from gear giving our lightning bolts a 1.54 second cast time. which isn't going to put us in a position to increase our DPS on paper....but considering latency should perfectly model that 5 LB between lava burst rotation.
Hi, first time poster here - hopefully a very mod/addon specific post isn't quite out of place here.
Originally Posted by doogless
It's a question of theoretical vs practical. Theoretically, casting CL when you have 1.2 seconds until LvB comes off cooldown would be a DPS increase over just casting LB, but in practice telling the difference between 1.2 seconds and 1.6 seconds on the LvB cooldown is harder than just mashing 2 again - to the point where it's arguable whether or not it's even worth trying.
Wondering about how practical it is to make split second decisions, I began investigating what, if anything, an elemental shaman specific mod could do. While my XML/LUA skills are limited, I've managed to put together something that is vaguely useful and gives me decent information regarding if I'm going to need to use a CL or LB in my rotation without having to make the decision in a split second.
A small FRAPS (19 megs. It does stream, but the quality is terrible. About a minute of thrilling target dummy DPS): eleshammod.wmv - FileFront.com
So that the mess of bars makes some sense, notes:
1. The top two bars are scaled to be "8 seconds" wide (LaB CD).
2. The "FS" bar basically shows the Flame Shock debuff on my current target.
3. The bar with the green beginning shows my LaB cooldown. The green section is an attempt to project how much "spare" time I'll have if I continue to spam LB. The lighter shade white part of the bar shows my current cast embeded into my LaB cooldown.
4. The more traditional cast bar, at the bottom, has a short red bar at the beginning. That's just a convinient place to show the length of an LB cast (scaled to the same length as the top two bars) for quick comparison to the green bit.
It makes sense to me, though it is probably some awkward amount of information overload, at first.
While it is a bit rough around the edges, I usually have a very good idea of what selection of spells I'm going to want to use as my LaB is close to coming off cooldown. With some practice it isn't too tough to tell, on the fly, what I need to be casting in order to make the green bit the length I want it to be.
While haste procs throw a small spanner in the works, the mod immediately adjusts for current level of haste and with a close-by timer (say for Embrace of the Spider), it was easy enough to keep that in mind, along with everything else.
Sadly I've been using a delayed-FS rotation, not a priority system. I don't know if a more developed addon could work well for a priority system but I figured I'd throw the idea out here. A bottom line for me, at least for now, is that I don't have to make split-second decisions.
I have noticed something odd about my rotation, which may have wider implications for server latency etc. (or may not).
I am using a rotation of:
FlS, LvB, CL, LBx4, LvB, LBx4 ... and repeat
What I have noticed is that there is a difference in the first and second parts of my rotation. This is assuming my current haste, M.Aura and WoA.
Between my first two LvBs, I cast:
CL, LBx4 - which gives me a very smooth gap of just a fraction over 8secs
Between my 2nd and 3rd LvBs, I cast:
LBx4, FlS - which gives me a gap of noticeably less than 8seconds, meaning I am tapping my toes for a few 100ms until LvB is ready.
Can someone with decent knowledge of server mechanic explain this? I imagine it has something so with latency and how our spells are processed, but on the face of it it seems odd. Oon paper the 2 sets of 5 spells I cast between LvBs have the same cast time, but in practice they absolutely do not.
Hi, first time poster here - hopefully a very mod/addon specific post isn't quite out of place here.
Wondering about how practical it is to make split second decisions, I began investigating what, if anything, an elemental shaman specific mod could do. While my XML/LUA skills are limited, I've managed to put together something that is vaguely useful and gives me decent information regarding if I'm going to need to use a CL or LB in my rotation without having to make the decision in a split second.
A small FRAPS (19 megs. It does stream, but the quality is terrible. About a minute of thrilling target dummy DPS): eleshammod.wmv - FileFront.com
So that the mess of bars makes some sense, notes:
1. The top two bars are scaled to be "8 seconds" wide (LaB CD).
2. The "FS" bar basically shows the Flame Shock debuff on my current target.
3. The bar with the green beginning shows my LaB cooldown. The green section is an attempt to project how much "spare" time I'll have if I continue to spam LB. The lighter shade white part of the bar shows my current cast embeded into my LaB cooldown.
4. The more traditional cast bar, at the bottom, has a short red bar at the beginning. That's just a convinient place to show the length of an LB cast (scaled to the same length as the top two bars) for quick comparison to the green bit.
It makes sense to me, though it is probably some awkward amount of information overload, at first.
While it is a bit rough around the edges, I usually have a very good idea of what selection of spells I'm going to want to use as my LaB is close to coming off cooldown. With some practice it isn't too tough to tell, on the fly, what I need to be casting in order to make the green bit the length I want it to be.
While haste procs throw a small spanner in the works, the mod immediately adjusts for current level of haste and with a close-by timer (say for Embrace of the Spider), it was easy enough to keep that in mind, along with everything else.
Sadly I've been using a delayed-FS rotation, not a priority system. I don't know if a more developed addon could work well for a priority system but I figured I'd throw the idea out here. A bottom line for me, at least for now, is that I don't have to make split-second decisions.
The addon looks pretty solid, any chance we could get an upload?
Between my 2nd and 3rd LvBs, I cast:
LBx4, FlS - which gives me a gap of noticeably less than 8seconds, meaning I am tapping my toes for a few 100ms until LvB is ready.
Are you sure it is the best to wait those 100-300 ms for the next LvB instead of putting in another LB or CL? I've got the same gap due to my disadvantageous haste rating of 520. If I dont have to refresh FS, the 5 LBs fit very nicely between the LvBs, but 4 LBs + 1 FS cause a small gap and I'm not sure whether to wait or to simply cast 5 LBs + FS. Is there a value of milliseconds that is worth to wait for LvB?
Are you sure it is the best to wait those 100-300 ms for the next LvB instead of putting in another LB or CL? I've got the same gap due to my disadvantageous haste rating of 520. If I dont have to refresh FS, the 5 LBs fit very nicely between the LvBs, but 4 LBs + 1 FS cause a small gap and I'm not sure whether to wait or to simply cast 5 LBs + FS. Is there a value of milliseconds that is worth to wait for LvB?
I'm fairly certain based on my own spreadsheet that with any gap of up to 200ms its better just to wait. Even though I naturally dislike spending any time doing nothing in a fight. I have never calculated at what point waiting becomes sub-optimal, although I am sure others have.
With respect thats slightly OT though. My issue is that there shouldn't be a gap, or a difference between the time to cast the two halves of my rotation. CL and FlS should have an equal cast time/GCD, so why don't they?
I'm fairly certain based on my own spreadsheet that with any gap of up to 200ms its better just to wait. Even though I naturally dislike spending any time doing nothing in a fight. I have never calculated at what point waiting becomes sub-optimal, although I am sure others have.
With respect thats slightly OT though. My issue is that there shouldn't be a gap, or a difference between the time to cast the two halves of my rotation. CL and FlS should have an equal cast time/GCD, so why don't they?
For whatever reason any time I do a wait-rotation my DPS is much lower than when I squeeze in another LB/CL.
Last night I hit 5310 DPS on a patchwerk that lasted 4 minutes and 44 seconds. We were trying to do a 20 man immortal(got screwed by a freak Glutch screwup and the raid went downhill from there) and we had all of our melee sit out just to be safe.
My rotation was sloppy as hell, my first flame shock missed because the Spriest didn't have Misery up in time and my following Lava Burst did not crit due to it. The only thing I made sure I was doing was casting continuously and using no wait-methods for LvB.
With as much improvement as I could have had last night I'm surprised people still are having difficulty breaking 5K.
The big metric I'm using to determine whether a higher DPS parse is a result of gear or skill is average LB hit (this doesn't take crit rating into account but I find there's not a huge disparity between Shamans there and the impact is relatively minor). Ezareth's LBs hit about 200 harder on average than mine despite a paperdoll spellpower advantage of about 20, which gets erased with a flask on because I'm an alchemist. The only explanation I can offer, the only really meaningful difference between our gear, is the second trinket spot (MotWP vs. Dying Curse). That proc is massive with a reasonable internal cooldown (45s, right?) I lose rolls to people who have double epic trinkets so I can only dream.
It does make sense that clipping your LvB cooldown is higher damage than waiting a fraction of a second for the cooldown, because if you've got to wait 0.2s or clip LvB by 0.2s you're better off using those 0.2 sub-optimal seconds to generate damage rather than sitting on your butt.
P.S. Not linking last night's WWS because Naxx was a laggy shitstorm and everyone's numbers were embarassing, but I landed 2nd on a non-laggy 3-drake Sarth. Aggressively switching between Magma and ToW worked very nicely.
The big metric I'm using to determine whether a higher DPS parse is a result of gear or skill is average LB hit (this doesn't take crit rating into account but I find there's not a huge disparity between Shamans there and the impact is relatively minor).
Doesn't Lightning Overload wreck numbers from WWS parses?
It skews the results downward, but everyone has the same LO chance so it should even out. My average LB hit stays pretty consistent over multiple parses so I'm willing to write it off as a wash. In any case there's no good LO-based explanation for Ezareth's LBs hitting over 300 harder than mine. If my LO was really proccing so much that it skewed the average hit numbers that much I doubt the DPS disparity would be so large.
Also calculated Ezareth's average casts/minute to see if this is a haste/rotation issue. They're nearly identical; he's pushing out one more cast/minute than I am. I had a much higher Bloodlust uptime, which is disturbing because it feels like my casts/min should be much higher, but he used a Haste potion and casts CL more often. This doesn't seem too out of whack given his advantage in haste rating; there's also one spot in my rotation where I have to wait for half a second. Also, looking at my Patchwerk parse I fucked up by assuming I'd have time for a second trinket; turns out he died before I could use it a second time so I wasn't able to stack EM/trinket/BL. That hurts my numbers quite a bit. So I'm chalking the difference up to haste disparity (small), better cooldown use, slightly more efficient rotation due in part to haste disparity, and Dying Curse proc (10s buff 700+ SP buff, 45s internal cooldown).
02/09/09, 6:23 PM
Similar Threads
