Kavan

The only difference between using strictly two cycles and interweaving is the ramp-up cost. The cost of ramping up comes up very frequently so it would be useful to get a good grasp of the actual cost. We don't want to look at dps lost. What we really want to do is treat the ramp-up as a dpm "cycle" and see what is the AB spam tradeoff compared to other fillers.

First a straight ramp-up as ABx4 assuming the debuff delay. For arcane/frost without CoE ramp-up is 3.46 dpm, Frostbolt is 3.4 dpm, optimum cycle is 3.04 dpm. Adding CoE Frostbolt improves to 2.99 dpm and optimum cycle to 2.73 dpm. Using a interweaved ramp-up of (AB-FrB)x3 we have 3.18 dpm for ramp up without CoE and 2.94 dpm with CoE. So it seams to me that the interweaved ramp up is even more efficient than just Frostbolt. This is with my gear so it's a bit biased but still.

Let's look at opportunity cost of using a suboptimal cycle for a specific amount of time. If we remember what the tradeoff means, it is the damage lost for each point of mana we're missing to sustain full AB spam. If we expand our formulas for multi-cycle situation we have:

T = tab + tc1 + tc2
M = tab * abmps + tc1 * c1mps + tc2 * c2mps
D = tab * abdps + tc1 * c1dps + tc2 * c2dps

Let's assume tc1 is a known quantity.

M = (T - tc1 - tc2) * abmps + tc1 * c1mps + tc2 * c2mps

Solving for tc2:

M = (T - tc1) * abmps + tc1 * c1mps + tc2 * (c2mps - abmps)
tc2 = (M - (T - tc1) * abmps - tc1 * c1mps) / (c2mps - abmps)

D = (T - tc1) * abdps + tc1 * c1dps + tc2 * (c2dps - abdps)

D = (T - tc1) * abdps + tc1 * c1dps + (M - T * abmps) * (c2dps - abdps) / (c2mps - abmps) + (tc1 * abmps - tc1 * c1mps) * (c2dps - abdps) / (c2mps - abmps)

D = T * abdps - tc1 * abdps + tc1 * c1dps + (M - T * abmps) * c2trade + tc1 * (abmps - c1mps) * c2trade

D = T * abdps + (M - T * abmps) * c2trade + tc1 * (c1dps - abdps) + tc1 * (abmps - c1mps) * c2trade

D = T * abdps + (M - T * abmps) * c2trade - tc1 * c1trade * (abmps - c1mps) + tc1 * (abmps - c1mps) * c2trade

D = T * abdps + (M - T * abmps) * c2trade + tc1 * (abmps - c1mps) * (c2trade - c1trade)

With this in hand we see that if we set tc1 to 0 we get the already known equation. Now we want to see what is the opportunity cost of having a tc1 > 0 which would correspond to the time spent on ramp-ups. First tc1 * (abmps - c1mps) is how much more mana we would have to spend to get to full AB spam. Typical values would be -24.35 mps for ramp, 262.39 mps for AB spam, duration of one ramp-up 12.93 sec. Together ~3700 mana for one ramp-up.

So in the non-CoE case a ramp up when using the optimum cycle costs us 3700 * 0.14 = 518 damage or for CoE 3700 * 0.21 = 777 damage.

So the real cost of having to do one ramp-up is less than 1k damage.

 

Wizeowel

It would be really cool if you could add these to the list on Rawr.mage. Give them their own spell cycle subtype 'RampUp' and force-include them into the spellcycle solution and at the start of the reconstructed sequence.

 

Alvira

Hmmm, if the cost of doing one ramp-up is less than 1k damage, I think I would stick to two cycle. Trying to keep up AB debuff to make sure it doesn't fall off is challenging because there are a fair amount of boss fights with movement and such.

More importantly to me. You won't really have a very low mana cycle because as long as you don't let the AB debuff fall off, your ABs are going to be the highest cost one. It gets harder to gauge your mana on whether it is enough or not.

Two cycle theory is more clear cut to me. You have the high DPS, High mana cycle where you are spamming AB. Then when you run low, you have the lower DPS, low mana cycle (your filler sequence) where you are trying to regen mana to the stage where you can go back to the high DPS cycle again.

If you interweave your filler through out. You don't really have a so called low mana cycle at all. It can essentially become just one long AB rotation throughout the whole fight. Like 6 AB, one frostbolt for example. But its a high DPS, high mana "rotation" throughout.

And if you misjudge how much frostbolts you really need to cast. You may intervweave in more frostbolts or more fillers than you truely need. Example: Say it looks like the fight would need some frsotbolts thrown in, total AB spam is not possible.

The mage using two cycle might find out at the very very end that he has to cast a couple of frostbolt in his low mana cycle. But it would probably be nearer the end of the boss fight when this happens. He can see more clearly that its almost over. He might then end up casting say 6 frostbolts during his low mana cycle and he quickly then switches back to AB spam because the boss is that 5% health at that stage.

The other mage doing the interweaving in of frostbolts. He knows he can't spam AB 100% of the time. Yet, he will need to go into his for example 5 AB, 1 frostbolt rotation very very early in the fight. Say when the boss is at 70% health still. How well is he able to gauge how many frostbolts he truely needs to weave in for the whole fight at that point? To get the maximum benefit, he would have to know the exact number of frostbolts to cast during the whole fight almost from the very start of the fight. Its too late to start weaving in when the boss is at 20%. Because a rotation like 5 AB, 1 frostbolt takes a fair amount of time to run through just one frostbolt.

Either you will run out of mana even with the weaving in anyway, because you started it too late. In which case you are forced back into a low mana filler sequence cycle anyway. Or you over compensate. You go with something like 5 AB, 1 frostbolt throughout the entire boss fight and end up casting more frostbolts than necessary and finishing with excess mana.

With two cycle. You are setting yourself up for sucess. Because you spam AB all the way at the start. its only when you find that you get low, where you change. And usually its when the boss is at lower health, so its far easier to gauge mana management at that point.

With weaving in. You are setting yourself up not to cast AB 100% of the time from the very get go. Because you start weaving in frostbolts much much earlier, since the entire cycle to weave in even the frostbolts is going to take a much long time than just the immediate ramp up that the two cycle arcane mage does.

And in the end, it is just to squeeze that extra less than 1000 damage from saving each ramp up. With the real possibility that you may end up doing too many frostbolts so you lose back the damage anyway.

Once you get into a routine of doing say 5 AB, 1 frostbolt. It gets very comfortable to just go with it for the whole fight. You are still burning mana at a high rate, and you and doing good DPS also. Indirectly, you are being "conservative". And to get the best DPS. You have to get the exact number of frostbolts cast throughout the whole fight correct. which is a challenge.

Anything less than optimal, and you are going to go back to two cycle casting anyway. Because say you started with 5 AB, 1 frostbolt gauging that that would be about right for the whole fight. If it lasts longer than you anticipated, you then have to lapse into 4 AB, 1 frostbolt, or some shorter lower mana sequence. On the flip side, if the fight lasts much shorter than you thought. By the time you switch over to full AB spam, you probably would have cast more frostbolts weaved in that you really should have. In which case you did less damage than the two cycle mage. Because in the shorter than expected fight. The second cycle for the two cycle mage is extremely short. Or it never even happens at all. Whereas the "weaving filler spell in" mage doesn't know this at the start of the fight and has already started to actively build in "weave" frostbolts.

Hence in the end, to me, if its for less than 1000 damage difference each ramp up saved. And given the more clean cut, easier to predict mana nature of going two cycle, hence I prefer two cycle.

 

Kavan

Well the way I see it it is no more difficult to do mana management with two cycles or with weaving. No one said that weaving should use a rigid cycle all the time. You can imagine weaving as a two cycle solution where your high dps cycle is AB spam and your filler cycle is AB-FrBx2 where AB is fully debuffed. Yes the filler is higher mps than FrB spam, but both cycle pairs are equivalent up to ramping differences and will produce the same result.

The real question to me is whether to use AB spam + FrB spam solution or AB spam + AB overlap cycle. As far as I can tell there is no one answer to this question and will depend on exact fighting conditions and buffs.

 

Alvira

The issue is that you probably won't be getting net positive mana gain if you are doing AB, frostbolt x2 with AB fully debuffed. Unless your group is that stacked. So, you are still losing mana. This means you may be at boss 40%. Gauge that you can't finish with AB spam, but at the same time, guage that you can finish with AB-FrBx2. And when you find that you misjudged even that, you have to lapse into an even lower mana cycle.

Frostbolt spam as a low mana cycle is very straight forward. You know you should be having net positive mana regen while doing it. So, you can afford to go really low on mana. And you are using the frostbolt spam cycle to regen mana. So, you could be like "Opps, I have AB spam all the way, boss is still not dead at 20%, but I am at 1000 mana left. So, ok, frostbolt spam time. Then at boss 10%. hey I now have 3000 mana, time to go back to full AB spam".

Unlike with AB-FrBx2 with AB fully debuffed, where your mana is still continually going down even with that sequence. So, its more like "I have 5000 mana, but with boss at 35%, I think that I can AB-FrBx2 until end of the fight and end it with zero mana. Opps I misjudged it, boss is at 10% and I am so low I can't even keep up AB-FrBx2 anymore. Time to go into frostbolt spam anyway."

 

• Intermission

I admit I'm new to the mage class in general, but I dont understand this talk regarding AB/Frostbolt. I thought it just goes like this:

Fight started, boss is at 80% and you're spamming AB. Due to knowing the vague time required in the fight, you start to cast:
AB AB AB AB AB Frostbolt AB AB AB AB AB Frostbolt etc
Looks like you wont have enough mana to keep doing this method, so you drop a few AB's relative to Frostbolts:
AB AB Frostbolt AB AB AB Frostbolt etc
Oh no wait, you got an Innervate or you misjudged your mana pot, you can go balls out for a little bit longer than planned:
AB AB AB AB AB
Ooops, I'm down to 10% mana and 30 sec away from gem:
AB Frostbolt AB AB Frostbolt

Isnt that how a fight would play out? What is the advantage or disadvantage of doing AB spam for 3 minutes and then 2 minutes of AB AB Frostbolt AB AB Frostbolt, compared to simply doing AB AB AB Frostbolt AB AB AB Frostbolt the entire time, then changing as needed? I'm not sure what the discussion is about :S

Am I way off track? Or is the goal to LET the AB debuff fall off for increased mana efficiency on the first/second AB on the new stack? I would of thought casting 6 Frostbolts in a row would be a terrible waste. You have to work your AB stack back up, either copping the haste/mana penalty on the chin, or using a [AB Frostbolt AB Frostbolt AB Frostbolt AB] rotation so you dont get charged more mana for a non-hasted AB. Why not just split those 6 Frostbolts over the duration of the fight. Or at least put an AB inside them to keep the stack.

 

Alvira

Well, if you are so familiar with the fight you know it will requrie exactly 6 frostbolts. Then by all means interspace out 6 frostbolts throughout your whole fight and you are done. The thing is boss fights are not exact in length and situations may cause them to be longer or shorter than you originally estimated.

For me, Starting out with AB spam assumes I am going for 100% AB casting from the very get go. I am starting from max DPS, and working backwards as the fight progresses.

Starting out with the idea of weaving in 6 frostbolts somewhere means that at 90% of the boss health when you throw in your first frostbolt, you are already accepting that you won't be able to spam AB 100% of that fight. You are already setting yourself up for 90% AB casting as opposed to 100%. If you then realise you have more than enough mana and you want to ramp up, you would have already weaved in a certain number of frostbolts.

 

• Intermission

Well my point wasn't really about inserting 6 frostbolts pre-emptively, and I agree with you on your example. (Although if I did know I needed exactly 6 Frostbolts, I would cast 3 of them at the start staggered with AB's to cancel the mana/haste penalty.)

Anyway I think we're just talking semantics here. So long as you dont loose/rebuild an AB debuff by doing something like 6 bolts in a row, it doesn't make a difference. At first I thought you meant you would cast AB, then swap to *purely* frostbolts, then go back to AB spam again once you realise you can go balls out with 10% to go. When obviously you can just throw in an AB between every 2 bolts during that frostbolt period, but start it a few seconds earlier, and not loose the stack.

I was mainly asking the question to make sure I was correct in my assumption that we dont want to loose the AB stack. I know that a non-debuffed AB is very mana efficient, and I thought there may have been something based around that which I had not thought of. I suppose my method of learning is arguing with someone who knows more than me, hehe.


Thanks for this thread by the way! I had always liked the idea of arcane but I never really knew where to start.

 

Tinybronco

When you talk about "AB-FrBx2 where AB is fully debuffed" does that serve as a net mana gain or rather uses less mana than AB spam for that given amount of time. I haven't taken the time to look into this and atm still use AB-FrBx2 where AB is rank 1 as my mana regen cycle (AB spam resort to this when I cannot AB spam till next mana cd). Is there a way to use rawr to find out if given the buffs how much mana is gained using AB-FrBx2 where AB is fully debuffed, given selected buffs and gear as it applies to you. Like Intermission I'm getting a bit confused as well.

For the moment I can generally AB spam for the majority of the fight length given my guilds progression, if I can I'll use AB-FrBx2, where AB is fully debuffed, which i'll try tonight instead of my AB(1)-FrBx2 rotation. Although the mana regen from frostbolt spam is good I didn't really care for the change in tempo and DPS and am unsure whether the extra regen from frostbolt spam, allowing for more AB spam in the future, would outweigh using AB-FrBx2 instead. I guess it depends on fight length and buffs etc.

Can you elaborate more on this? I'm very much new to the TC side of things.

 

grayrest

It doesn't matter. If you did misjudge a 2 FrB cycle and had to drop to pure FrB spam, then you would have had to cast the same number of FrB in the low part of the pure AB->pure FrB cycle, except you would have had to predict it before you ramped the ABs again.

The point that everybody but you is making is that there little practical difference between weaving and using your all AB and then all Frostbolt method. The only real differences are when you spend the mana and that doing an all frostbolt low cycle forces you to ramp AB every time you start up the AB part of the cycle.

Nobody is suggesting you start with a cycle. You start with AB spam to prevent your mana from capping out and to get your pot/gem cooldowns going, spam down to a comfrotable mana level (I drop to just under 4k mana unless I'm holding mana for something) and start cycling. All this does is move the ABs you'd cast after a ramp forward in time to the low dps part of the cycle. It's still two cycle, but the low part of the cycle is higher dps/mps.

If Blizzard's cast system wasn't bugged, you started with 100k mana, mana didn't cap out, and the target had infinite health,

40 FrB, 20 AB
20 AB, 40 FrB
20 AB->2xFrB
10 AB, 10 AB->2xFrB, 20 FrB
10 FrB, 10AB->2xFrB, 10 AB, 10 FrB

Would all take the same amount of time, use the same amount of mana, do the same amount of damage because they all only ramp AB once. The only difference is when the mana is spent. Because Blizz's cast system is bugged, you start and are capped at a realistic amount of mana, and the target will die at some point, some of those options are more likely to produce higher damage than others (namely the ones starting with FrB would be worse because you'd lose regen mana due to being capped; start with AB spam or a ramp cycle).

Weaving is just an easier way of determining the number of frostbolts that must be cast. You say "is the fight going shorter/longer than I expect?" and then decrease/increase the number of frostbolts. If perfectly played, either method will cast roughly the same number of ABs (only difference being the ramp ABs), regen the same amount of mana, spend the same amount of mana (modulo ramps), do the same amount of damage (ditto). Most people just find weaving to be an easier way to approach the theoretical maximum number of ABs. Assuming, as always, that true drop-the-buff-AB cycles aren't better, which is a problem I'll leave to the more rigorous.

 

Alvira

Guess it then boils down to personal style and preference then. Two cycle doesn't have that much ramp up of AB really. There's just one ramp up ideally after the low mana cycle phase. And as Kavan calculated above. The difference in doing one ramp up less is lower than 1k damage.

So, having the AB debuff drop off and having to reapply it just that one time compared to weaving in frostbolts to ensure it never drops off is a difference of less than 1k damage. (Assuming in the end, both sides cast approximately the same number of frostbolts).

Two cycle with frostbolt spam at very very low levels of mana just front loads more AB. While Two cycle with the weaving in of frostbolts at the later part means switching to such a weave cycle earlier and spreading out the ABs more at that point.

 

Kavan

Alvira I don't see why the filler has to be mana negative. It seems like you rely on that because you go all out until oom and only then change to filler. Doing that will almost never be optimal and under normal circumstances you won't have enough mana to do the AB spam burn on second AP/IV. I'll agree that planning for second burn phase is more challenging, but proper mana management is at the core of playing arcane to the max.

You can't directly, but you can compute it as (2 * FrB mps * FrB cast time + AB mps * AB cast time) / (2 * FrB cast time + AB cast time).

Based on the available buffs one of the following will be optimal for arcane/frost. For dps cycle you'll always use AB spam, the only difference is in the filler. One option is FrB spam (or equivalently weaving it in with AB spam), the other is (ABFrB)x3+FrBx2 overlap cycle. When I'm trying things out in Rawr I can only get FrB spam to be optimal filler if I stack things heavy in FrB favor (winter's chill, no cos). Some have reported it being better under normal circumstances also, but I'm not clear under what conditions that happens.

The overlap cycle looks like this: AB(overlap)-FrB-AB(1)-FrB-AB(2)-FrB-FrB-FrB. The only difference between this and FrB is the ABs. The most important one is AB(overlap), but the other two are also important. If we could choose how many AB debuffs we have then the ultimate two cycles would be AB(3) spam and AB(0) spam. If you go compute AB tradeoff for AB with arbitrary debuff you'll see that (ignoring differences from proc effects) all have the same tradeoff which is equal to dpm of non-debuffed AB divided by 5.6875. In the example arcane/frost setup I'm using this is 2.74 dpm tradeoff compared to 3.12 dpm on frostbolt. It's important to note that tradeoff for AB(overlap) is 0 as it has the same dps as AB spam (this just says you're not losing any damage compared to AB spam, but you are actually gaining mana). So all the ABs in the overlap cycle have better AB spam tradeoff than FrB, which makes it clear why the overlap cycle as a whole has better tradeoff compared to FrB.

The usual concern about overlap cycles is that a pushback can cause you to lose the AB debuff and you get the full 2.5 cast on AB. Well the way to look at it is that the shorter cast for AB overlap is just a bonus. As I've said all ABs no matter how many debuffs there are have the same AB spam tradeoff. So AB(0), while worse than AB(overlap) is the same quality as AB(1) and AB(2). And all have better tradeoff than FrB so even if you mess up each overlap you'll still get better tradeoff than FrB spam. The reason why you want to use full overlap cycle is that AB(overlap) in that case has the best tradeoff (that is 0), the ones you would get at lower debuffs aren't as good (by definition they can't be better than that as they are all lower dps).

 

Wizeowel

I can't actually think of any encounter in MH/BT where I'd actually want to try to do overlaps. Either the fights are short enough to only AB spam, or there is enough pushback or movement going on to make overlapping too risky. So I tend to stick with AB/frostbolt weaving just for simplicity. It's very easy to look at your last week's WWS for approx length of fight, and then plug that into Rawr for expected number of frostbolts.

For Brutallus fight parameters, Rawr suggests I use [(ABFrB)x3, FrB...Sc] in an overlap cycle. This for an extra 17dps over frostbolt weaving. I am wary of trying this, but since the fight has no movement and the pushback is very predictable, maybe I could do it. On the other hand, it does seem a lot of effort for only 17dps. Theoretical dps.

So for the moment I will be keeping the AB-cycles unchecked in Rawr.

 

Kavan

This is exactly my point. People don't like AB cycles because they think that messing the overlap will somehow make it worse. As I've shown even if you don't get any of those overlaps you would still do better. Most are saying that one should never drop the AB debuff. What I've shown is that dropping the debuff is actually good as ramping ABs have better tradeoff than FrB.

 

Wizeowel

I'm glad you said this. It made me read the last two paragraphs in your previous post again.

So as I understand it now, paraphrasing for my own clarity. You use frostbolt as filler for 2 reasons. Firstly to break up the ramping of AB so that you're not affected by the bugged AB debuff stacking, because that would indeed worsen the trade-off. Secondly to allow for the overlap to happen at all. You don't look at the overlap so much as a cheap AB(3), but more as a way to get back to AB(1) which has a better trade-off than frostbolt on it's own.

So I guess the moment was short-lived and I will reapply the AB-cycles option in Rawr

 

Webbers

Hi Guys, I’ve read your forums over the past year or so and can’t say how glad I am that you’ve opened it to the public. I would like some advice here to non Mage players in the position of Guild / Raid leaders. From my limited experience mages can either be your best Damage dealers or your worst – and the difference is not always the player.

I’m mage leader and was all about Fire. Then 2.4 dropped and a 2 pc T5 mage respeced to Arcane. I gave her an ultimatum, Stay within 2% of my total damage and you can raid as arcane. She bet me by 3% twice in a row (don’t ask). Quickly I had another 2 pcs T5 mage to go arcane. We dropped them in a 2 Shadow priest, Resto Shaman Group.
We feed them innervates like candy.
The Resto shammy is marked with a circle and has a macro emote “Mana tide down”

Why did we do that??? TPS. As DPS there isn’t much close to and Arcane mages Damage Per Second – Threat per Second Ratio. I haven’t seen the TPS argument anywhere. To me it is one of the great strengths of an arcane mage. The fact is threat limits total damage output, Period.

Another spin off we’d not thought about was that when the Arc Mages started topping damage meters by a mile – All our other DPS got GREEN with envy and behold they started really pushing, every buffs, haste pots, destro pots – the works. We sent them here for advice to stop the QQing about the OP group. The result is that SSC/TK is a breeze. MH is easy and BT doesn’t seem that tough.

Anyway my point is - if you’re leading a raid please think about a SP SP RS AM AM group. Healers don’t really need the SP – they can live on pots. Pots are cheap i.e 5g. A Raid wipe due to low dps costs more than 80g (not to mention mental anguish)

On the theory crafting side - I’d really like some Math on the threat outputs of IV,AP,AB,AB,AB over FB,FB,Fb,FB,FB,Fb. Sure your tanks should always be able to hold threat off your DPS but we’ve all been threat capped in the past. It might help some mages choose Arcane, and help some raid leaders support them.

 

Keldarn

Our best tank outputs about 1200 tps, which equals my tps with fire spec. Add in one or two ticks of invis and you're always fine.

 

grayrest

This is simple. Fire up rawr, pull in an arc mage, set up the buffs and read the tps off the cycles tooltip. I can't particularly remember a time I've been both threat capped and had invis down regardless of spec.

 

Webbers

I’d love to but I don’t have rawr at work. Also by the time I get online (in 40 hours or so) this thread will have left me behind o.O

If anyone could add the TPS and DPS for a arcane mage doing a solid AB spam and compare it with any other class (casters or melee). I’d really appreciate it - I've never done the math and seriously doubt my ability to do so successfully

Why waste your time? It may help others value Arcane’s TPS – DPS ratio. Not all tanks are created =

And yeah invis is a powerful tool – if only it wasn’t a wasted 5 seconds (+ time to remove the buff).

 

bynop

As fire, I usually find myself at the threat threshold on any fights where I don't have to move as much. I will almost always be required to use invis on nalorakk and gorefiend.

but invis is meh, i never do the complete aggro dump, just the first few ticks will drop enough aggro to matter for the rest of the fight.

what you do is take advantage of the way invis is removed when any action is taken, including spellcasting.

Stop casting
pop invis
wait 2 secs
start casting scorch (invis fades on it's own)

 

• Anaxo

If your mages are limited by the TPS of your tanks, it's not the fault of your mages, it's a problem with your tanks. Between misdirect and the changes to warriors, I only pull aggro my main tank if I blow all my cooldowns within the first 5 seconds. Low threat is a bonus for arcane spec, not a selling point. Yes, threat limits total DPS output, but not just of your mages. Having your rogues/hunters/shadowpriests/mages/warlocks take the spec which has a bigger thread reduction talent is approaching the problem from the wrong angle.

Back in the days of AQ40 my guild had a tank who was nicknamed 'Rogue Killer' because the rogues would outaggro him without even pushing hard. The rogues didn't go splat with our other tanks, just this one. Once that warrior finally got a Thunderfury, said rogues no longer died when he was tanking. Which begs the question of why, because nothing had changed with the exception of Thunderfury. The tank just had threat production that was that terrible, and only Thunderfury's proc allowed him to cove