Hello all, I have a question regarding starfall and delaying it to gain greater benefit against some bosses (most against LK and deathbringer). The question I ask is, how long can you delay the starfall before its a dps lose. I cant seem to get my heads around the calculations.
Starfall every second blood beasts cast, starfall's damage is unaffected by their AOE reduction aura so it's free add DPS. Period.
On LK it's debatable. I generally don't bother waiting for trinket procs or the like to cast starfall, my philosophy is that you will do more DPS casting it the moment it's off cooldown than you will delaying it waiting for procs and getting one less cast per boss. Ghoul / shambling horror DPS is largely irrelevant in either normal or heroic LK. For heroic, DPS on the boss matters most (must push phase 2 before a third shambling horror or guaranteed wipe) so getting two starfall casts off is rather essential then a third around the second raging spirit spawn (so you're cleaving and also hitting the first one which likely still has some HP), meaning it'll be up again right around the first valkyr spawn. So the "cast it off cooldown" philosophy generally wins on that fight as well.
There are some fights where casting it requires awareness - heroic rotface you need to make sure you don't cast it if a mage / priest gets the first infection, as if they used mirror images or fade, the little ooze drops threat on them and will immediately run to you when stars hit it.
So, I dropped BS for engineering. Mostly I did it for boots enchant, but I would really love activating my gloves on time.
To keep it simple, let's imagine a macro I would press each time Lunar Eclipse procs:
Now, what happens to me, is that when i press this macro it all works perfectly, but if I am smashing buttons, like most of us do when dpsing, the gloves won't be used. Imagine you have 3 seconds left until your gloves come off CD. You begin smashing this macro non-stop, Starfires are cast, but gloves are never used, 5 seconds or even an hour later. I believe this happens because when I press macro gloves are unusable because my GCD is busy, but next spell is still qued and fired right after current one, so when i press macro next time it all repeats again
Currently as it is for me: I get lunar eclipse, I have go check if gloves are off CD (another thing to look at ain't doing good for overall dps), if they are on CD I continue to dps as usual, if they are off CD I have to make a pause after casting that last wrath, make sure gcd is over (my wrath is <1s) and only then press gloves macro. Please share your experience with me, or tell me what I'm doing wrong, as for now this was a dps loss profession for me (
Yes you've spotted the quirk with it. The macro doesn't operate with spell queueing or inside GCD. What I do is before the first Starfire of the Lunar Eclipse I wait a tiny bit longer and spam the macro as the cast bar reaches it's very end. That way it doesn't fail. From my experience the drawback is that I lose 1xMyLatency of cast time per Lunar Eclipse, but that *appears* to be an acceptable loss (i have fairly low latency, <100ms). The quirky operation I don't find a problem and I still manage to commit it to habit and not actively track the CD of Hyperspeed. However, there may well be differences in how we spam our keys that might make it more annoying for you to use it like I do. I try to spam my keys in a quick flurry as my cast bar goes over half way. I can't spam constantly (carpel tunnel). If you spam in a non-stop stream it might seem more clunky to use, also if you have considerably more haste than me (i have 650 ish) you might feel the pause more. It's the same situation as trying to drink a Speed Pot at the start of Lunar so it's not really a new phenomenon.
Apologies for not mentioning this in my first post.
Of course the absolute best mileage from the gloves is to stack it with Power Infusions whenever possible. Often I found the Disc Priest unwilling to monitor my buffs closely enough to infuse on Lunar and not Solar, but it is possible via either Power Auras (2 linked auras) or with other UI tweaks (like a filtered Satrina Buff Frame showing only my Eclipse buffs and a cast [@Kaitain] macro), to make his reacting as effortless as possible. With my gear Hyperspeed + PI + NG = 1.3 sec Starfires throughout Lunar, and I typically would get that lining up like that 2 times per fight. That was until my PI buddy rerolled a Paladin anyways :-(
Starfall every second blood beasts cast, starfall's damage is unaffected by their AOE reduction aura so it's free add DPS. Period.
On LK it's debatable. I generally don't bother waiting for trinket procs or the like to cast starfall, my philosophy is that you will do more DPS casting it the moment it's off cooldown than you will delaying it waiting for procs and getting one less cast per boss.
Starfalls main star damage might be unaffected (though I'm not sure it is), but the splash dmg IS affected. I try to sync it out of beast-time
Your second point has little merit, since it's not guaranteed that delaying for X seconds means you lose an extra starfall. If you delay for X seconds, you can still end the fight with a cooldown on SF thats higher then X.
I agree that IF you lose an extra SF then its most likely a dps loss, but you can't assume this happens every time.
Remember that a Starfall on a single target shoots 10 stars in 10 seconds, and on more than a single target shoots 20 stars in 10 seconds. So (ignoring AoE damage reduction for now), a Starfall on multiple targets does twice as much damage (more if there's any potential splash) than one on a single target. So it's worth waiting for multiple targets, even if you miss an entire cycle in the process.
The splash is only around 1/4 of the damage on a single target. So you still get at least around 1.5 Starfalls by waiting to use it on Beasts at Saurfang. At Lich King you definitely want to use it every other Val'kyr wave, getting double damage + splash every 90 seconds.
1.5 a worst case when saving it for beasts. Saurfang has a decent sized hit box compared to the distance the beasts spawn from him and any splashes onto him will still hit for full damage. What more, the splashes will still hit the beasts for 10% and with 5 targets out gathered so close together you'll probably splash a beast or two and so still get 20-30% of a single target splash out of them for each star.
As far as waiting for trinket procs and whatnot, the math isn't too hard. Starfall gets nearly all of its power from really good scaling, so there is the potential that waiting for a spellpower proc will be worthwhile. Using some really simplified numbers that aren't too far from accurate, let's say the druid has 4k base spellpower and gets another 1k from the trinket proc and we're only considering single target. I'll also ignore any persistent damage multipliers or crit which don't change the relative power.
Starfall has a base damage of around 7000 and a scaling of 430% of spell power. That means that a 4k spellpower starfall does 24200 damage and a 5k starfall does 28500. 28500/24200 = 1.18, or about 18% more damage. That means you can wait up to 18% of the cooldown (10-11 seconds glyphed) on average before waiting goes from a dps gain to a dps loss. Interestingly, this creates a pretty clean generalized rule that I think can apply fairly accurately to any spellpower proc or effect. That is that for every 100 spell power your starfall will gain by waiting for the buff (averaged over the 10 seconds of starfall) you can delay your glyphed starfall by about 1 second.
That kind of makes me realize something interesting about DFO. The actual average spellpower for it depends on what 10 second period in the proc you're looking at, with the average spellpower from casting starfall at the start of the DFO proc equal to about 315 and the average from the last 10 seconds being about 840. The growing spellpower from DFO only grows at about half the rate it would need to in order to make it worth waiting for the later half of the proc to cast starfall. Using the rule above, it would seem you would only be able to wait about 3 seconds for a DFO proc in order to cast starfall. Of course, because starfall updates spellpower on every star afaik a starfall cast 3 seconds before the proc still gets about 170 average spellpower from your DFO, so the number is actually more like 1-1.5 seconds, which is under the average time to proc for the DFO even if it is already off cooldown. In other words, delaying starfall for DFO related reasons is always an average dps loss.
Applying similar logic to phylactery, once you get under 10 seconds until it procs the last stars from your starfall would have benefited from the buff whether you wait or not, bringing the average gain from waiting down by 107sp for every second you wait, which means you're roughly breaking even by waiting. If you wait and for some reason phylactery doesn't proc within 10 seconds that's a dps loss. Heroic phylactery has a small chance to be worth waiting for, though for it to be any gain at all you'd need to cast starfall pretty much the instant your trinket procs and the gain would be small.
So updated generalized rule for waiting on spellpower procs to cast starfall which is really just a particular case of the first rule: A spellpower proc needs to be well over 1000 spellpower to be worth considering delaying starfall to wait for it and given currently available trinkets, just don't.
All the same logic should be applicable to %damage buffs, just changing the 100sp/1sec ratio to around 1.66% damage/1 sec, with the caveat that I'm not sure that starfall updates %damage buffs on a per star basis. If it does not then you have to take that into consideration when you figure the %damage your starfall will gain by waiting for the buff. A second target can just be looked at as a special case with a damage buff of 100% or greater.
It should go without saying, but it seems foregone that someone will feel the need to say it anyway like they've stumbled upon some great hidden truth - That time for delaying a spellpower trinket only applies as an average over many fights of indeterminate length or when you have no idea how many fractional multiples of a starfall cooldown are left in the fight (i.e. the margin of error on your estimation of the remaining time in the fight is 30 seconds or greater). In fact, what I'm really calculating here is the extra damage you get from a trinket proc weighed against the odds that waiting will cause you to miss a starfall in a fight of indeterminate length. There's no way delaying works out to be a dps gain if it causes you to miss a starfall. Similarly, as long as it doesn't cause you to miss a starfall, you could theoretically wait up to 50 seconds for a trinket proc and gain dps.
I didn't subtract the amount some other spell would have gained from being case under the proc instead of starfall, but the coefficient per casting time of starfall dwarfs all other spells (It's 7-8x as much) by so much that that wouldn't be a huge effect.
I took a major shortcut on my generalized rule for trinket delaying by only looking at two points and inferring a linear relationship with a 0 y-intercept to describe all points. At 4k spellpower the spell power coefficient on starfall is already dominating the base damage to such an extent I don't think a more accurate formula would lead to substantially different results though.
Fights with really long periods of no or much less important dps like sindragosa should basically be treated like multiple shorter fights when considering whether delaying is going to cost you a starfall cast.
If you have a trinket that really outshines the rest of your gear such that your trinket proc gives significantly more than 1/4 as much spellpower as you have without the proc then it may be worth delaying a short amount of time. Same thing if your starfall isn't glyphed.
You can potentially use some other tricks, which makes waiting for a trinket proc relatively more attractive.
We basically have two ways to increase the damage of a Starfall without relying on procs: i) a wild Magic potion and ii) Gotw to force a OoC proc.
If it is worthwhile to use those tricks during a Starfall that has increased damage due to a trinket proc, then waiting for a trinket is relatively more attractive, since both a OoC and a Magic potion scale with a SP proc.
Since both of those can just as easily be used without the trinket proc the effect of them on the decision to wait is small and doesn't really change the rule.
Basically, multiply the % boost you get from the controlled burst by the % boost you get from the trinket and use that number in the calculations from before.
At current crit rates the 4% or so you get from wild magic is around a 2% damage boost to your starfall, using my made up numbers from before that makes waiting for wild magic give about a 2% x 18% boost, or 0.36%, which is worth waiting maybe 1/5th of a second, or not at all. The spell power portion stacks additively with a trinket proc so actually slightly discourages waiting.
Omen of Doom gives around a 1.6% additional boost by lining it up with a trinket proc, which increases the amount of time it is worth waiting by around 1 second. In a perfect theorycrafted situation where you would guarantee your heroic phylactery proc within 10 seconds and, more importantly, use starfall the instant phylactery procs this is probably enough to get a very small dps gain out of delaying your starfall. However, say you're casting a starfire when phylactery procs and delays your starfall by a second, that puts you back to breaking even. Anything else that causes you to delay and you're back to a dps loss. This is all within the rounding errors and small effects I ignored in that last post.
Practically speaking, it still stands that the only situation where you delay your starfall for a trinket proc is when you know you will still get the same number of starfalls off during the fight either way.
Last edited by Videl : 05/19/10 at 5:37 PM.
Reason: I made some kind of math error I had to track down.
Thank you very much for clear Starfall explanations.
I still have got one thing unsettled, and unfortunately I am unable to model it out in wrath calcs. At which point stats-wise, spaming Starfire is benificial during Heroism/Bloodlust regardless of an eclipse proc? What about extended moonfire? Currently, I cast Starfire 1.1s during Bloodlust/Heroism + Haste Pot + Hyperspeed Accelerators, obviously that's a no-brain situation since it's lunar proc , but then I could still spam 1,5-1,4 Starfires, and I have a feeling switching to Wraths would be a dps loss... I am unsure though. I have studied WoL on Lord Rebrad(H) since it's closet thing to Patchwerk we can get, and have seen different choices by top boomkins, both netting good results.
Same question goes for that Power Infusion during Solar Eclipse, what should i do ( except spanking Disc Priest)? In multitarget situation -Lich King Valks- I choose to get extended moonfires on all of them. In single target scenario I generaly keep spaming wraths (and curse priest if target is about to die, and forgo rotation if target has plenty of HP to work with.
Please share your experience or better yet help me find a way to model these things out. It might seem like a small dps gain/loss but for me every single tiny bit counts when fighting that Lich Kings as we are still working on a perfect phase 2 PS as usual, be kind and forgive my bad english
I did some calculations on this a while back. My findings were that, given my gear, during a Bloodlusted Solar a Wrath is better than a Starfire. However, a MF combined with 3 Starfires is better than a spamming Wrath. Any starfire beyond the 3th is worse than Wrath. Off course you might be able to improve by trying to cast the MF during the Lunar > Solar transistion or use the MF and 3 Starfires near the end of the Solar to prevent having to refresh MF in the upcoming Lunar.
I am not sure if it is a good idea to go for extending MF on different targets like the valkyrs. You might not benefit from the extension if one of the valkyrs goes down fast. From a less selfish perspective it might even be a good idea to put FF on all valkyrs. I haven't done any calculations on this. It will depend on the expected uptime of the valkyrs, which should be that long in normal LK, but perhaps long enough in heroic LK.
Does anyone have a good estimation of the raid-dps gained by having FF up?
Edit: Response to Videl
I agree that those effects are very small and that is most likely not worth it to delay Starfalls to benefit from procs. Moreover, there is also not much to be gained by delaying Starfall in order to get more Lunar Starfires. Therefore, Starfall usage in single target fights mainly depends, as you point out, on length of the fight. (Delaying a Starfall for a few seconds during a Bloodlust + potion of Speed Lunar might be an exception.)
Something I still need to calculate is if it is worthwile to: i) pop GotW during Starfall to force Omen of Doom ii) pop a Wild Magic potion during Starfall.
Option ii) is a bit hard to evaluate, since you have the option to wait for a particular Starfall that coincides with trinket procs, Eclipse, etc. The longer the fight, the more attractive this option becomes, since the probability of having a one or more of those buffs up during one of the Starfalls increases in time.
The alternative is off course a potion of Speed during Bloodlusted Lunar.
The problem with evaluating option i) is mainly calculating the extra uptime of Omen of Doom due to a GotW cast, where the uptime of Omen of Doom depends on what you are doing during Starfall. For example, during Bloodlust the uptime of Omen of Doom is relatively higher (please correct me if am wrong on this), which makes casting GotW relatively less attractve, while also the opportunity costs of the gcd are higher.
Isn't FF being applied or not a function of how much HP the add has and what % of that HP is nuked off by physical DPS, rather than the time it is alive?
Time alive and hp are usually heavily correlated though
To estimate the value of FF, I could for example take some rogue and calculate for him how much his dps goes up by having FF on his target. However, on adds burst damage is likely to play a more important role, which implies that by this method I would undervalue FF.
Wowhead says the Valkyr Shadowguard's are ??. IFF is a 3% boost to casters at the hit cap (unless the target will have Misery anyway). It is also a 5% armor reduction. I'm not sure of all the math after Arpen, but I'd assume that is a on the order of a 1% boost to physical DPS (but ferals could provide that at no cost).
It costs you a GCD, and likely causes you to lose NG (unless Starfall is up). Overall, that is a penalty of 10-15k damage at Heroic LK 25 levels.
Assuming no Misery and no Feral, it is probably never worth casting on a target with less than 300k health (unless you are just looking for another instant-cast spell). It is probably always worth casting on a target with over one million remaining health (I'm not sure of their health in heroic 25).
Note that if your raid doesn't have Ebon Plaguebringer, E&M is a much more important debuff.
First off, as our tactic includes heavy clustering it also includes running away from each other as the valk spawns (since defile will come any time druing the phase).
So what I do is running in the direction of the edge where the valk is flying (or the 3 are, if 25M) and on the run MF/FF/IS (that order, because I might turn my back against them and MF must be cast facing them). I do not FF all three, but thats because we have a feral and in 25HM a second owl will be there, so that is distributed well.
Secondly though I am pretty sure they are 82 and NOT ?? (at least I am sure thats what tukui shows me).
But since you may, depending on tactic, be running after the spawn: Use it to dot/ff her and then start bursting her down.
On 10 man though, if you are out far enough by the time the MF-gcd is over it might be not worth casting it (I skipped it since we only had a rogue melee her on our 10HC kill and that guy stunned her anyways so not full burst here).
So I would really make it situational on lineup, way you have to run and finally the fact if the hit is needed (maybe I am wrong about the valk lvl). If hit is needed, the situation is clear: Cast it, it will be worth it, even in 10 man.
Imo, you should always keep FF up on main targets like bosses, the only excuse for a moonkin to not put up imp FF is on a tank and spank fight where everyone stays single target, liike BQL for instance.
This is ofc still only if you have a feral druid and SP in the raid.
For bosses like LK where you are swapping targets all the time it's essential that you use that 1 gcd, so the 3 % is up 100 % of the time. Sps missery isn't a long lasting buff like FF, so might ware off on encounters like that. Another good example on this is Lady Deathwhisper.