If you have a Disc priest in your raids is it optimal to use his/her Power Infusion in place of your arcane power when you want to pop all your cooldowns + haste potions + heroism? At what haste rating is it not optimal?
Lets also assume abyssal rune and lightweave just procced, you used the Talisman of Resurgance, and you gain these cooldowns right before your 5th arcane blast. IF it really seems that if Arcane Power is additive, PI must be way better depeding on your current stats.
Also how might a potion of wild magic affect it differently?
If you have a Disc priest in your raids is it optimal to use his/her Power Infusion in place of your arcane power when you want to pop all your cooldowns + haste potions + heroism? At what haste rating is it not optimal?
Lets also assume abyssal rune and lightweave just procced, you used the Talisman of Resurgance, and you gain these cooldowns right before your 5th arcane blast. IF it really seems that if Arcane Power is additive, PI must be way better depeding on your current stats.
Also how might a potion of wild magic affect it differently?
Potion of Wild Magic is really only useful once you're hitting the GCD, do the math and see whether you're there or not.
Time to re-sim arcane? Major Bug with Arcane Potency.
I'm fairly confident that the information in this post is both new and correct, forgive me if this is already a known issue.
Simply put, Arcane Potency currently does *almost* nothing for Arcane Missiles. This is not the first time the two spells have had a wonky interaction, but now the pendulum has swung away from the mages benefit and now, it seems, that you only get the 30% crit bonus on 1 out of 5 missiles.
Here's a quick summary of what I found:
1) Clearcasting now procs on spell impact rather than spell cast. This is the source of the current AoE bug where AoE effects have a much higher chance than normal to proc clearcasting since AoE spells have multiple impacts. Because it takes 1 second for the first missile to travel, and as much as a 1 second of travel time, this means that clearcasts proc about 1-2 seconds into an AM cast depending on your distance from the target.
2) When clearcasting does proc, you'll notice that the arcane potency that procs is immediately consumed on the next missile. You might expect it to therefore benefit all the remaining missiles, but I do not believe this is the case, read below for a more in depth explanation of that. Also, this means that next spell does not have Arcane Potency and, as far as I can tell, behaves exactly as you'd expect it to (no crit bonus).
3) With this under consideration, you'd expect only the last ~3 missiles of AM to even have a CHANCE to benefit from Potency, but it seems like only 1 of them does. I tested this on a target dummy for about 30 minutes. Keep in mind, my sample size is small because ONLY missile waves that procced clearcasting were counted. The vast majority of my casts were tossed out.
I managed to proc clearcasting on 26 waves, that's 130 missiles. Of those 130 missiles, 33 crit. Only 33. That's 25.35% crit rate. My paper doll's listed crit rate was 24.08% (using mage armor instead of molten armor). However, if things were working as intended, each of these missiles should have had a bonus 30% crit and thus my *expected* crit rate is 54.08%. Even with a small sample of only 130, a 25.35% crit rate vs a 54.08% crit rate is a huge gap that cannot be explained by the standard deviation for a sample of that size. While I'm happy to do more testing to increase the size of this sample if there are skeptics, I believe this should satisfy most here that the issue is indeed real.
The most likely explanation for this outcome is that per clearcasting proc only a single missile is gaining the +30% crit rate, not all 5.
If you want some "anecdotal" evidence to support this Hypothesis, go to the target dummies and do this simple test. Cast arcane missiles into you proc clearcasting, and watch. You'll see the arcane potency buff does indeed disappear pretty much instantly. Now do the same test with Blizzard. With Blizzard, each wave can proc clearcasting, and each subsequent wave (but not every wave, so even Blizzard is still short-changed) benefits from that Arcane Potency and the buff does not disappear until the very last wave of Blizzard. The different behaviors between these two spells pretty clearly demonstrates how arcane missiles SHOULD work if it were functioning correctly (potency buff only disappearing on the final wave). But even if it were functioning "correctly" as Blizzard does, it still does not get 100% of the expected benefit. The first missile or two would still not be included just as all waves of Blizzard prior to clearcasting proccing are excluded from the benefit.
This seems like a serious issue for Arcane mages, and in light of this information, I'm beginning to question whether the spec holds up vs Fire on non-incanters-absorption-friendly fights. I welcome any feedback/comments/questions.
I'm fairly confident that the information in this post is both new and correct, forgive me if this is already a known issue.
This exact issue occurred regularly during the Beta from the time they introduced it up to shipping. The proc timing was on cast so only the first missile would crit and consume the buff. It usually got fixed in the next patch or two, though I seem to remember it being hotfixed one time.
I noticed today that Arcane Potency has been changed when using Blizzard as far as I can tell. Casting blizzard after gaining Clearcasting consumes Clearcasting but not Arcane Potency, it instead fades as soon as the channel is completed or cancelled. If you reproc Clearcasting again during the channel the Potency buff will still fade once the channel is complete but you will still have Clearcasting left for the next spell. It could be that it's been working like this for a while and I just haven't noticed, but then why haven't they made this change for Arcane Missiles aswell? Besides being aoe both are channeled spells with individual missile hits.
edit: It appears that you can break the Blizzard channel with a instant cast spell and still get the benefit from Arcane Potency.
With a 33.8% paper doll crit rate, using Blizzard only when having procced a new Clearcast (with Potency) I kept braking the channel before it finished using Frost Nova. Since it seems pretty conclusive I didn't bother gathering more data than this:
This exact issue occurred regularly during the Beta from the time they introduced it up to shipping. The proc timing was on cast so only the first missile would crit and consume the buff. It usually got fixed in the next patch or two, though I seem to remember it being hotfixed one time.
Yes, I realize it's not new in the sense that it's never happened before. I can recall quite a few potency bugs that have popped up since it's inception. In the very first version, the buff was tied into the clearcasting buff and PoM buffs themselves instead of as a separate buff. In that very first version, you could cast Presence of Mind then cast all the AM/Abar you wanted and all would get the +30% crit cause none of them would ever consume PoM. Then, once that was resolved, there was the issue wherein each clearcasting proc benefit not 5, but 10 Arcane missiles -- the 5 from the wave that procced it and the 5 from the wave that consumed it. That even meant that if a wave both procced and consumed a clearcasting, it would have +60% to crit instead of +30%.
After that, they implemented it was a separate buff, wherein this issue first popped up and then was fixed the next patch.
What I meant when I said "new", was simply that knowledge of its *current* reemergence was not already widely known and that Simcraft/Rawr were not currently factoring this bug into their game model. As a result, I believe it's quite possible that the narrow margin of 57/3/11 over 20/51/0 may well be a lie (except for fights wherein Incanters Absorption is strong). If I am wrong about that, please let me know.
Another potency bug, while I'm on the topic, is that you can currently cast a presence of mind then cancel it (thus starting the spell cooldown) or even switch specs and still keep the 30% potency for your next cast. This means you can technically have Potency for the first cast of any spell in any spec. Whether or not that's a worthwhile expenditure of time, I suppose, depends on how dedicated you are to min/maxing each encounter.
I just stumbled across a thread on the EU WoW-Europe forums regarding the behaviour of the Missile Barrage proc during Arcane Power. Apparently it still costs in the region of 150 mana, even during the free mana proc. I'm assuming the game is making the Missile themselves free but still charging the 20% mana Arcane Power tax.
Anyone had any experience with this? Either way I'd presume it's a bug? Applying a 20% mana charge to a spell that costs zero mana should logically still be zero mana.
The actual text on the Missile Barrage talent may have something to do with it, it states that it reduces the mana cost of the spell by 100%, but doesn't explicitly say the spell will 'cost no mana', as in the Brain freeze talent.
It could be possible that, for Missile Barrage, the mana cost is determined by:
X - (X*1),
X of course being the mana cost of Arcane Missiles. Then adding on Arcane Power (assuming that it is indeed additive) you'd get:
X - (X*1) + (X*0.2).
It would seem strange to word it in a different way to Brain Freeze if the mechanic was exactly the same, and I assume that in the case of Brain Freeze, X is set to zero (X being fireball in this case).
Missile Barrage: Gives your Arcane Blast a 40% chance, and your Arcane Barrage, Fireball, Frostbolt and Frostfire Bolt spells a 20% chance to reduce the channeled duration of the next Arcane Missiles spell by 2.5 secs, reduce the mana cost by 100%, and missiles will fire every .5 secs.
Brain Freeze: Your Frost damage spells with chilling effects have a 15% chance to cause your next Fireball spell to be instant cast and cost no mana.
Last edited by Rugz : 10/05/09 at 1:16 AM.
Reason: Talent texts added for easy reference.
Well with the fix to MB we're back to where we used to be. Here's the new old cycles in case you forgot (with the same setup as for previous numbers). Note that these are optimal regardless of whether you have ABar.
For an extra two seconds of Evocation to be worth it, you need to be able to use that mana to fund sixty seconds - an entire minute of cast time - of extra Arc Blasts. Over an actual fight, proposing that you already have enough mana to fund ABMBarr, the amount of remaining combat time available for Evocation (a single tick, even!) to be worth casting becomes astronomical.
I think what Aramezzet is saying is that the dps difference between rotations with different mps constraints are minimal enough that using a less mana efficient rotation and then evocating becomes a dps loss almost regardless of how much mana is returned by evocate.
It is easy to misinterpret what is being said here - having 0 dps for 2 seconds over the course of a minute is still only a 3.33% dps loss. However, for that (albeit small) dps loss, we could have switched from ABSpam to ABSpam4MBAM and had higher dps for insanely less mana consumption. This is not limited to these two rotations.
Also, since there is a 54 mps difference between the two your 2 seconds of evocate would need to return 98 * 54 - 125*2 = 5042 mana for mps to also break even. I think this pretty well illustrates the reasoning behind choosing the rotation which allows you not to evocate over the one which allows you to continue using the same rotation, and then evocate when you go oom.
I could, of course, be completely misinterpretting this post, but I believe this was the point that was intended, unless I'm misunderstanding Kavan's numbers.
Last edited by Casstor : 10/05/09 at 1:35 PM.
Reason: Changed time interval to clarify the result
I could, of course, be completely misinterpretting this post, but I believe this was the point that was intended, unless I'm misunderstanding Kavan's numbers.
You're interpreting the post correctly. But there is one very important point that some seem to forget. Those numbers were computed under a special case with artificially lower haste so the numbers are not representative. The way dpm tradeoffs work that doesn't affect optimality of cycles, but it does affect the choice regarding Evocation. In practice it will very much depend on your specific gear. For example in the setup I have here I get:
One important difference between normal spell cycles and evocation is that evocation is limited. You can interpret the numbers like this. Let's take a cycle and lets say it's not sustainable. We know dpm conversion tells us for each mana missing to make it sustainable how much damage we're losing by having to use the lower cycle. Similar for the dpm conversion against evocation. Overall damage will be higher if we take the one with lower dpm. So for the above you can see from ABSpam first switch to ABSpam04MBAM before going to evocation. But for all the other cycles we get a better deal by using evocation first before moving to next cycle.
Now I don't want this to turn again into people thinking this is ground truth. These are very specific numbers for very specific gear and buffs. Check Rawr to determine where evocation fits for your specific case.
Then I wondered - with clearcasting now proccing on all targets individually and granting a high uptime of +30% crit rate, would a simple one-button Arcane Explosion spam (a la FoK) with an arcane build not only be as high DPS and also much better mana efficiency with the free casts? Just stand in the center and smash AE...
Quoting a post from the FFB-thread since this thread seems more appropriate.
I've been experimenting abit using Arcane Explosion up until I proc Clearcasting and then start channeling Blizzard breaking it much like the old "Arcane Shatter" right before (or after if I feel latency is on my side) the last wave with another Arcane Explosion to double dip on Potency (behaviour described here). Since Arcane Explosion seems to keep 10% proc chance per mob, unlike Blizzard that anecdotally seems to have 1/8 of that per wave, already at 5+ mobs you stand a high chance (40%+) of reproccing Clearcast for another wave of +30% crit Blizzard.
If anybody feel like helping out with the math I'd appriciate it, going to take a stab at it tomorrow when I get time over.
Though I agree completely that evocation is something to avoid, I don't believe it to be something to make a rule. Obviously, if you're fighting patchwerk, don't evocate. If you're fighting heroic jormungar or icehowl, burning mana is advantageous before the end of the fight. Evocating during worm burrows or while icehowl is dancing in the middle of the room, even if mana/ticks are wasted, is zero dps lost for a potential ton of mana gained. I potentially lose overall dps on jaraxxus and twins so i have more dps on shields / portals with early mana munching and evocs, since they have loose enrage timers. FC's if you're doing anything but AB spam until you think your mana wont last till the first mob dies, then you're doing it wrong. [edit] Most of this is considering the "warm up" of the beginning of cycles (0 to 3 stacks) being quite low on DPS compared to 4 stack dps, and in burn phase situations you cant expect to fully complete every cycle before the phase ends.
I think all these "this is how you dps as arcane" things get too drilled into peoples' heads, and they dont consider the versatility of actually having the ability to have a temporary dps increase at the cost of an overall loss that is either insignificantly small or just doesn't end up mattering for your kill.
Anyway, This blizzard potency bug sounds interesting, if i get a chance I may mess with it tonight.
Due to the way Blizzard made Arcane Missiles work i'm pretty convinced you actually do get the 30% spell crit bonus, but it gets eaten your first crit of arcane missiles, ( hit 1 = 55% chance 2-5 = 24% chance ) judging by that your still a little low but not ridiculously so.
You're interpreting the post correctly. But there is one very important point that some seem to forget. Those numbers were computed under a special case with artificially lower haste so the numbers are not representative. The way dpm tradeoffs work that doesn't affect optimality of cycles, but it does affect the choice regarding Evocation. In practice it will very much depend on your specific gear. For example in the setup I have here I get:
One important difference between normal spell cycles and evocation is that evocation is limited. You can interpret the numbers like this. Let's take a cycle and lets say it's not sustainable. We know dpm conversion tells us for each mana missing to make it sustainable how much damage we're losing by having to use the lower cycle. Similar for the dpm conversion against evocation. Overall damage will be higher if we take the one with lower dpm. So for the above you can see from ABSpam first switch to ABSpam04MBAM before going to evocation. But for all the other cycles we get a better deal by using evocation first before moving to next cycle.
Now I don't want this to turn again into people thinking this is ground truth. These are very specific numbers for very specific gear and buffs. Check Rawr to determine where evocation fits for your specific case.
I was aware of you artificially lowering your haste. The way I see it, if haste affects dps and mps the same way, then adding 10% more haste to the same numbers increase dps, mps, and the time spent casting for each tick of evocation by the same amount. Thus, if I scale down my time period (using one rotation for 100/1.1 seconds and the other for 98/1.1 seconds), but keep the same method of analysis, the numbers I see will be similar. Is this a faulty assumption?
Edit: Of course, your point still stands that this tradeoff may be worth it for people with certain gear sets. I understand that. But if haste doesn't change the tradeoff, and stats which increase the damage of your spells (hit/crit/SP) but don't increase the output of evocation are the ones mages gear and gem for - doesn't this make the tradeoff worse the higher your level of gear? That's provided you are unable to gear almost exclusively for haste I suppose.
First of all, please excuse my bad english.
I have a true problem with playing the arcane-rotation. The situation of interest is, when I'am casting AB4 and have no procc ready(same for 5,6 and so on). If I spam the AM now, hoping to get a procc, my rotation is "AB4AM04MBAM" and I risk to cast AM without procc. If I just spam on AB, I risk to cast AB5 with a ready procc after AB4, what means a much higher manausage on ABs.
So how do you play the rotations "ABSpam04MBAM" & "ABSpam024MBAM"? Do you wait after AB4,5,6 to see if you have a procc and let the latency+brainlatency lower your dps or do you just always cast one AB to much, when the situation occurres?
In my opinion, these ABspamxxxMBAM rotations are theoretical constructs and unplayable without dps loss or much higher Mps-values.
Any tips for me? Because AB4AM04MBAM is just a very little dps gain compared to fire and then arcane is imo not worth loosing the additional possibilities fire offers.
But if haste doesn't change the tradeoff, and stats which increase the damage of your spells (hit/crit/SP) but don't increase the output of evocation are the ones mages gear and gem for - doesn't this make the tradeoff worse the higher your level of gear?
I believe you are on the right track but drawing the wrong conclusions from this line of thinking.
Your point is that crit/hit/sp increase the mana efficiency of your rotation, which I agree with. Each one of these stats increase not only your overall dps but also your dpm, since you're using the same time/mana for each cast but doing more average damage.
However, the key here is that as your dpm increases, the value of evocation in a rotation that is otherwise unsustainable increases because you are able to do more damage for each point of mana available to you. Again, this assumes that your rotation is unsustainable, since unused mana is wasted, but the key here is that crit/hit/sp increase the value of mana as they increase your efficiency of it, which may be somewhat counterintuitive.
So how do you play the rotations "ABSpam04MBAM" & "ABSpam024MBAM"? Do you wait after AB4,5,6 to see if you have a procc and let the latency+brainlatency lower your dps or do you just cast always one AB to much?
In my opinion, these ABspamxxxMBAM rotations are theoretical constructs and unplayable without dps loss or much higher Mps-values.
I believe these rotations take into account the fact that you will always cast an additional AB after MBAM procs (assuming MBAM did not proc from 4pt8 or from a ABarr cast on the move, which is accounted for by AB0MBAM).
This is why you do not see AB1MBAM as part of these rotations; if you do not wait to see if MB procs (which is a proven dps loss) then you can never expect to cast an AB1MBAM.
While we're on the subject, does anyone have an explanation for why this is the case? I'm sure the sim is correct, but I can't come up with a plausible explanation for why AB2MBAM is a dps increase over AB3MBAM but a dps loss over AB4MBAM. It seems nonlinear.
I believe you are on the right track but drawing the wrong conclusions from this line of thinking.
Your point is that crit/hit/sp increase the mana efficiency of your rotation, which I agree with. Each one of these stats increase not only your overall dps but also your dpm, since you're using the same time/mana for each cast but doing more average damage.
However, the key here is that as your dpm increases, the value of evocation in a rotation that is otherwise unsustainable increases because you are able to do more damage for each point of mana available to you. Again, this assumes that your rotation is unsustainable, since unused mana is wasted, but the key here is that crit/hit/sp increase the value of mana as they increase your efficiency of it, which may be somewhat counterintuitive.
Actually, my point was that if your mps stays the same, you require the same amount of evocate time. If your dps stats change such that your dps increases by 10%, then if the damage of the rotation without evocate was higher before, it'll be higher after as well, and the amount more damage it'll deal is increased by 10%.
Assuming Rotation 1 is the rotation with evocation and Rotation 2 is the one without. Damage after looks like this:
If the difference is positive before the 1.1 multiplier, it'll be positive after.
Of course, this ignores the fact that different rotations scale slightly differently with stats, which may or may not contribute to any differences.
Also, honcl, in response to your post starting with "While we're on the subject...", what you quoted and what you said were too different things. The later you use your MBAM procs, the less mana efficient you will be and more dps you will do. This seems obvious, as MBAM is your cheapest spell (mostly 0), and stacked AB is your most expensive spell, but stacked AB does slightly higher dps. It is nonlinear and should be, because in some cases (i.e. the case where MB procs during your third AB cast) all rotations beginning with "ABSpam" function the same.
While we're on the subject, does anyone have an explanation for why this is the case? I'm sure the sim is correct, but I can't come up with a plausible explanation for why AB2MBAM is a dps increase over AB3MBAM but a dps loss over AB4MBAM. It seems nonlinear.
From ABSpam04MBAM to ABSpam024MBAM the only difference is that for 40% of the cycles (those where MB procs on the 1st Arcane Blast) you are cutting off early with a MBAM. This lowers dps (since you aren't waiting for 4 stack AB to cast MBAM), and lowers mps (since you aren't casting the high cost 2, 3, and possibly 4 stack Arcane Blasts). Same thing for changing from ABSpam024MBAM to ABSpam0234MBAM.
Originally Posted by Crowl
If you have to control a robot dinosaur that fires lazers and there's a time when you shouldn't be shooting those lazers then the encounter is clearly flawed beyond hope of fixing.
What I mean by nonlinear is I would expect the rotation from highest dps down to look like:
ABSpam04MBAM
ABSpam034MBAM
ABSpam024MBAM
Or something similar, not skipping AB3MBAM and going directly from 4 to 2.
It makes sense that it must have something to do with the chance for a MBAM proc being consumed by a later AB stack, but I don't understand why this is more relevant going from 2->3 than it is from 3->4, given that the proc chance and damage increase are both linear.
What I mean by nonlinear is I would expect the rotation from highest dps down to look like:
ABSpam04MBAM
ABSpam034MBAM
ABSpam024MBAM
Or something similar, not skipping AB3MBAM and going directly from 4 to 2.
It makes sense that it must have something to do with the chance for a MBAM proc being consumed by a later AB stack, but I don't understand why this is more relevant going from 2->3 than it is from 3->4, given that the proc chance and damage increase are both linear.
This should be how they rank. ABSpam034MBAM is not included in Kavan's dps cycles, likely because ABSpam024MBAM is the optimal choice for a second spot to use MBAM from a mps perspective.
10/03/09, 3:59 AM
Similar Threads
