Akron

Doubt AM ticks are classified as periodic damage. The fireball DoT is though, although I've never seen it proc anything.

 

Keyne

I personally was not talking about warlocks specifcally. I was trying to speak generally about the way that casters (people that use spell hit, crit, dmg, and haste) gear themselves for a boss encounter. I am unsure that many people know how to properly do this, if any at all (myself included). It just so happens that locks value the same things we do...

 

epiphenom

Hmm. Kavan, using your formula from the int/spirit thread, I don't see how you're getting 450 mp5 while casting. Unless I'm applying the formula wrongly, it looks like you're going to have 667 mp5 OO5SR, and Mage Armor plus Arcane Meditation is 60% of that number, or about 400 mp5.

Obviously, it's still a huge increase from what we're getting now, but I'd still like to see what I'm missing.

 

Kavan

I'll try to explain how I got to the numbers for Intellect posted before. I mostly do numerical analysis, but I'll try my best to backtrack those and break down where the value comes from. All the evaluations here are based on my gear and a typical 5 min boss encounter. The numbers will change in different situations, but the basic idea is the same.

Intellect has the following values: base mana pool, evocation mana regen, spirit mana regen, crit, spell damage. Base point of Intellect has the following multipliers: 15% from Arcane Mind, 10% from Blessing of Kings, 5% from Gnome racial, together 1 Int stat results in 1.32825 Intellect. This means 1 Int gives 19.92375 base mana and 11.95425 mana from evocation. (Note that the value of evocation is already included in the base mana evaluation)

For spirit mana regen we have mps = k*spi*sqrt(int). Assuming 870 int, 485 spi base, we get a partial derivative of k*spi/(2*sqrt(int)). So a base 1 Int results in 1.32825*k*485/(2*sqrt(870)) = 0.10185 mps. How much mana exactly this is depends on how much time you spend in 5SR which depends on fight characteristics and spell selection.

Crit is simpy 1.32825/80=0.016603125%=0.366597 crit rating, and damage is 0.25*1.32825=0.3320625. For non-gnome mages all values are simply 5% less.

What remains needed to compare all these values is equivalence between mana, mps, dmg and crit. At the core is dps/mps for AB spam and AB cycle, the ratio of which is determined by equating starting mana pool with mana consumption, mana regen and external mana regen. One could write dps as a function of these factors, do derivatives on the variables and compute actual gradients. It becomes impossible to handle once you add in all the cooldowns, that's why I'll just use the numerical evaluations I have.

So here's the values for different stats. 1 mana = 0.0157 dps, 1 mps spi regen = 2.0183 dps , 1 dmg = 0.6589 dps, 1 crit = 0.5731 dps. Using the breakdown of int contribution we get the following values for 1 point of int: 0.31 dps from base mana pool (and evocation), 0.2 dps from spirit regen, 0.21 dps from crit and 0.22 dps from dmg. Total this is 0.94 dps per point of int or 27% increase over its previous value. A direct numerical evaluation of int comes to 0.96 dps, which is close enough to be explained by rounding errors and synergistic effects of stats. Compare this to 0.6589 dps for 1 point of damage and the difference is quite significant.

 

Kavan

You're right. The 450 mp5 figure includes mp5 from my gear and buffs.

 

Muphrid

Kavan, you're assuming mana always limits DPS, which it doesn't.

 

◊ Vontre

So I plugged in the new formula from the thread for my spreadsheet, and it looks like spirit based regen is about twice as effective as previously. I'll be putting up an updated version soon.

www.magegraf.com

Raiding is full of challenge. Sometimes there is fire. You have to not be in the fire.

"We agree with Communism." - Greg Street 2009

 

Kavan

I'm not. I'm just not close to that point. I'm using 175 mp5 sh priest and el shaman in my calculations. Obviously with better sh priest it would come much closer to that limit. Right now for 2.4 AB spam is roughly -260 mps after all regen. If we can close this gap it only means it is no longer needed to use AB cycles, but I don't think we're close to that point yet or that we will be anytime soon.

 

Muphrid

But by using an AB spam model, you may end up correctly pointing out that int is the most valuable stat for AB spam...while ignoring that AB spam may produce less than maximum DPS precisely because you are limited by mana, even in this 5 minute scenario.

In other words, you may be saying something useful about a rotation that is, at present, useless.

 

Alvira

I guess the question then would be rather. Does AB scale as well as fire if you consider it as an average of 1.5 second cast in DPS terms.

Another question I don't know whether has been theorycrafted, is whether at high enough levels of haste, and assuming the mana to sustain it, would AM spam also become competitive. There wasn't enough haste gear previously, but with more being added in sunwell, it may be possible. The new T6 and gear allows stacking of both stats, +dmg, haste all at one go, with more options of where the 4 slots of T6 will be (to get the 5% bonus to AM).

 

Etherealz

Haste scales with fireball and AM at the same rate up till 200% where fireball reaches the 1s gcd (and 50% for scorch of course). So, it won't change am spam scaling.

 

Searix

Wow, it just occured to me we can yet again use 2 peice t5.

Someone theorycraft this, using 4 peice t6 and best 2 peice t5

Edit: Vontre's doesn't model the new global cooldown and assumes haste doesn't help ab spam :/

 

Jarlyn

My comments were slightly off-base before, largely due to the fact that I had only been loosely following the Int->Spi thread. I had grossly underestimated the amount of regen provided by the 2.4 changes, and the math I had been doing was quite wrong because of it.

However, I still believe this discussion does ultimately boil back down to Fire vs. Arcane. Our mana regen will be significantly boosted in 2.4 and this does shift the "balance of power", so to speak, slightly back in Arcane's favor - this is indisputable now. But I think there's a fundamental flaw in this argument, in that more mana does not directly correlate to more damage. This assumption is true only if the encounter is of sufficient length to warrant the added mana, and that there are no other extraneous factors affecting mana consumption (Kaz'rogal marks, Shaz mana drains, RoS P2, et al). A more accurate assumption, I believe, is that more mana is more potential DPS.

It's common knowledge that AB spam was, is, and will still be viable over the short term, and that because a lot of encounters in T6 content are comparatively short, Arcane compares favorably to Fire/Frost through large chunks of Hyjal and a decent bit of BT. But as fight length increases, Arcane becomes less and less competitve. Everyone knows this. My point is this: it is unlikely that we will ever reach a point where pure AB spam is possible. TC says this is likely the case, but speaking practically, were we ever to be in that situation, the factors making endless AB spam possible would almost certainly be nerfed.

So given that to be generally true, 2.4 changes nothing in regards to whether Arcane is viable. The new regen effects will make the max DPS cycle of Arcane sustainable for a longer period of time, perhaps even substantially longer. But sooner or later, the constraints placed on your DPS by the incredible mana consumption of AB will become too great, and Arcane becomes an inferior spec to Fire. Which brings me back to my original point - unless you absolutely, positively know that you will be able to do maximum DPS as Arcane for the entire duration of the fight, then you shouldn't be speccing Arcane.

 

Alvira

I am not good with math as you guys. But I did get some math models downloaded from this forum. Vontre's excel spreadsheet (mage_dps). I took some liberties with the numbers.

I input 700 into INT (Karan showed it can be done if you gear, enchant and gem for it). I then raised spirit to 600 to approximate the new regen rate. (I don't know if this is being too off base or not). The rest were stats of my current gear (not the best possible), with some reduction taking into account regemming and enchants.

stats inputted as follows:

Intellect (unbuffed)*: 700
Spirit: 600
Spell damage (all schools): 0
Crit Rating: 329
Hit Rating: 133
Mana/5 sec. 28
Fire Damage: 1000
Frost Damage: 950
Arcane Damage: 1000
Haste Increase (rating): 0

I then created a cycle called AB spam. (I put in AB 3 to simulate AB spam throughout).

Based on vontre's spreadsheet, AB spam gave me 2055 dps. Yes, it only lasts for 1 min 50 seconds. BUT here's the interesting thing. During that time, I would have outputted 372,930 damage

I don't know about the rest of you guys. But most wws I come across, the average total dmg down over the course of a boss would be around there (300,000 to 400,000). Here's the thing. If AB spam can be kept up for 2 minutes, it will output that amount of damage, in just 2 minutes flat! Seems amazing to me.

Of course, maybe I set the spirit too high. But as I said, I am just using some approximations to gauge some ball park numbers. Not trying to be so exact. If someone can come out with exact numbers, all the better of course. It is possible, because Karan stated that those levels of gear (for int) can be achieved. And he said in his post it would allow him to spam AB for 2 minutes.

AB spam has been limited by three things. GCD on haste, mana and spell pushback.

This upcoming patch solves two of those problems. haste and mana. Think its worth some exploring on whether arcane can now be competitive with the changes, especially on bosses with no spellpushback.

 

Alvira

Spell Haste will now help AB spam right? Cos it can go down to as low as 1 second? Why won't it help AB spam?

 

Muphrid

I think this 600 value is what needs scrutiny. Until the tools can catch up with the new change, it's useful to know that...



...where s_eq is the total spirit in 2.3 required to provide the same regen as s spirit and i intellect in 2.4, and k = 9.327E-3.

We can flip this around and see that for s_eq = 600 and i = 700, s = 329 or so. Is this consistent with the gear setup you were trying to use?

 

BrTarolg

He just means vontre's model doesn't account for it yet.

But christ, this is some odd TC. I also suppose it depends on the practicality of the encounter you are facing - will all the new sunwell encounters be a lot of tank and spank gear checks? Or will they be long, complicated fights?

Also, you have to consider what you're going to do when *not* AB spamming. Since you can dump your mana pool *so* fast, for a more practical purpose you could consider going all out with AB at the very start, switch to fireballing (with a 40/21 spec or something similar) for max DPM, and then back to arcane when you think you can finish the boss off (assuming this is some kind of 4-5 minute fight which allows for that)

Probably the simplest way to calculate this - find how much MP5 you can bring in, and how much DPM you can get off arcane blast (1.5 seconds), work out the "prebuffer" you get from your origional mana pool (so thats how much of a headstart arcane gets in terms of damage over fire), and then add on the MP5*DPM (as DPS) from then on (probably relatively low)
Or you could do it in a graph so you can see the spikes you get (from mana pots and evocation and the like)

Another interesting thought is - a fight with LOADS of movement, but short periods of standing still, arcane will benefit very heavily from, simply because fire will not be able to do *anything* during those moving times (other than fireblast perhaps, or maynbe the odd fireball here and there, but more likely just keeping up scorch debuff), whereas arcane will be constantly regenning superhigh mana throughout.

 

Alvira

Based on vontre's spreadsheet, with the inputted numbers I mentioned, I can fireball infinitely. So, yeah, theoratically, if the fight allows for it, it may be possible to AB spam until you have like just 1k mana. Then just keep on scorching or fireballing indefinitely after that. Sure, your DPS after the second minute will be lower than pure fire spec. But its not like you are wanding. Plus you would have already dumped close to 400,000 dmg onto the boss in the first 2 minutes. So, you are already ahead.

Previously, pure AB spam never worked because it ran oom just way too fast, and didn't do enough damage in total even after it ran oom. If enough damage in total is done, its fine.

My current gear has slightly less spirit than 329. But its not a lot less. With T6 or sunwell gear (that has tons of spirit too), it could probably approach that.

 

Alvira

Not really. Too much movement is not good for AB spam, its good only for AM spam. You need a bit of time to build up to a 1.5 second AB. So, if you only get to cast a few ABs before you have to move again, then your DPS won't be as high.

But a fight where you stand still and spam AB without interruption for quite some time, then you stop for a lull, then again you can start again would probably be good. My guild fights Anetheron by asking all mages to stand out of range of his carrion swarm. We only DPS the infernals. Arcane is quite good in that situation. No spell push back. Intense period of DPS followed by a lull before the next infernal drops.

Fire actually builds up to a faster maximum DPS mode than arcane if you have 2 or 3 fire mages. Cos, each fire mages cast sorch once or twice, and they are in full DPS mode. Arcane needs to cast arcane blast 3 times before getting into full DPS mode.

 

Pintofbrew

Alvira, in a number of your posts now you point out non-numerical facts which are not TC. "Arcane needs to cast arcane blast 3 times before getting into full DPS mode." . How can you compare scorch ramp-up to AB ramp-up? They're two utterly different things. The fact that they're both preambles to the main nuke is irrelevant and the way they work is so radically different that comparing them is totaly chalk and cheese. Two firemages will make scorch ramp a lot less duration than AB ramp and scorch will need 1 cast refresh every 21-27 sec while AB will need a full-ramp whenever it falls off, which will be very often for certain high-mobility fights.

"Then just keep on scorching or fireballing indefinitely after that. Sure, your DPS after the second minute will be lower than pure fire spec. But its not like you are wanding.". How can you say this? Arcane spec scorching is pathetic, and unless you're planning on going 40/10/11 fireballing isn't particularly better than wanding either. The goal is not to "do 300k" within 1:50 and then not wand, the goal is over a given amount of time to do the most possible, as fast as possible. Yes 2055DPS is great but what we're trying to understand is "Is the mix of AB as-much-as-possible and AB-AM rotation the rest of the time better or at lease competitive with Fire-IV Cooldown stacking?"

Kavan: You seem to have neglected to add BoW to your overall raid regen. Also, do your calculations assume JoW on boss to resolve value of int?

 

Kavan

I do include BoW in the calculations. For the numbers I posted I'm not using JoW, but the values do not change that much. JoW on a 10 min fight would increase AB spam uptime by about 40 seconds. For the values I have (175 mp5 sh priest) ABAMx3+Sc rotation is +3 mps without JoW and +34 mps with JoW, in both cases you're gaining mana when you're not AB spamming.

 

Selun

As I see it Kavan posted the value that int will gain in 2.4 and since BoW (a flat mp5 increase) and JoW (on spellhit proc chance for mana gains) doesn't affect intellect they shouldn't be taken into consideration.

 

Alvira

A fireball cast from an arcane mage will do a lot higher damage than just wanding. Based on what I entered previously into vontre's spreadsheet, its 1230 dps for fireball spam. Scorch spam does 1073 dps, not fantastic, but definitely a lot higher than wanding too.

 

Alihandras

I want to pose an entirely new rotation to the table I'm still toying with the Haste numbers but I have an idea. Last night I took out all my ZA haste gear got my spell haste to around 230ish, specced a werid as hell 46/0/14(+1). Put the rotation together of ABx3 AMx1 Frostboltx1. With the haste at 230 I was able to get in a 2.6 fireball instead of the frostbolt, however due to the coefficients of damage isn't it wiser to throw a frostbolt as I will not have Ignite in order to get IV, plus specced for 100% critical dmg from frost increases the off hit that much more. From the haste set I was able to maintain a 4.45 AM with a 2.23 base AB, and a 2.17 frostbolt.

It's a bat out of hell I know, but with this now I'm at a 1200 solid dmg base, with 20% crit, mind you this in mostly ZA gear with 2/5 T5 and off T6 pieces. The question I want to pose as I'm going to test the dmg viability in a raid enviroment this week in BT, is that with Sunwell coming out and the massive amounts of spell haste that are attainable now, can Arcan become viable again? I'm also going off a ruling out of mana pool, Still not sure on that equat as far as the SP durability in some fights for us.

What I'm trying to do is get to the point where instead of scorch we have a huge nuke in place. At the same time with IV up stacking beyond 2 AB's is wasted as the 2nd cast is under the GCD your at 1.61 sec cast after 1 stack with it up so the rotation I was playing with while under the affect of heroism or IV was AB AB AM Fbolt Fbolt repeat. With IV/Hero Stacked, It was viable to AB AB AM AM AB Fbolt repeat, The only conern to a point becomes the GCD however with 2.4 that will be reduced as well with our haste.

I know its a longshot but my image without putting it all together, is attaining the Grand Magister's Staff of Torrents and gemming 3 +10 Spell Hit. With that one piece of gear you are given 80 spell hit (arcane cap met). Your pieces can now focus souly on Spell DMG, Crit, and Haste entirely. With the arcane spec you'll be able to stack a very large amount of spell haste and crit vs. deep fire being limited to stacking hit, and most likely passing the staff for the MH/OH (+100dmg). It's something to think about that I haven't seen conversed yet anywhere and would like to recieve any feedback on. Thank for your time.

 

Pintofbrew

I'm interested to see some data of how much haste exactly is the max one can wear before the AM+Fbolt become too close to causing the subsequent AB to finish casting -without- the debuff having worn out. Some analysis of 150-200-250ms latency fluctuation would also be nice.

Then we can possibly look into AB*3/AM/Fbolt rotations which will perform the same accepted function as the good'ol AB*3/AM/Sc rotations used to: To capitalize on 1st AB with 3x debuff but cost only 0x debuff. It is also arguable that a 13-talented frostbolt is much more useful than a ignite-only scorch, and a 47/0/14 is much more eye-pleasing than a rather ineffective 40/10/11 or 42/8/11 or whatever malarkey.