Haste and Arcane blast don't go together too well if you combine AB with further haste cooldowns. It just means that you go under the GCD which means wasted haste.
So yes, for socket bonuses, if ur arcane and are haste capped the int and spi would be better I believe.
Well aside from the totally awesome mana vaccuum it creates, haste is effective up to approximately 263 rating (rounded up from 262.554) at which point you hit 1.0s with IV up, without IV you will not be at the GCD. The only time you will drop below GCD and waste haste (so to speak) is with Heroism/BL and you can reduce this cap if you'd like to account for that (and further reduce by 80 if you have 4 leatherworking fiends with drums!).
Well aside from the totally awesome mana vaccuum it creates, haste is effective up to approximately 263 rating (rounded up from 262.554) at which point you hit 1.0s with IV up, without IV you will not be at the GCD. The only time you will drop below GCD and waste haste (so to speak) is with Heroism/BL and you can reduce this cap if you'd like to account for that (and further reduce by 80 if you have 4 leatherworking fiends with drums!).
Edit: I forgot to mention I never stack IV with BL/Heroism, I always stagger my IV's around Heroisms.
Well aside from the totally awesome mana vaccuum it creates, haste is effective up to approximately 263 rating (rounded up from 262.554) at which point you hit 1.0s with IV up, without IV you will not be at the GCD. The only time you will drop below GCD and waste haste (so to speak) is with Heroism/BL and you can reduce this cap if you'd like to account for that (and further reduce by 80 if you have 4 leatherworking fiends with drums!).
Because everyone is debating about the cap, I'm gonna show you my math here so that I'm not just throwing around a random number. I did this all with very simple ratios and cross multiplying for instance:
1.0s = 80%
x = 100% Solve for x by doing (1.0s x 100%) / 80% = x
assuming 1.0s is ideal timing for ramped AB cast we know that approx 1575 haste = 100% and we know that the 20% increase from IV is what (in my model) would put us at 1.0s so therefore our cast time needs to be at a time that will reduce to 1.0s when IV is activated.
so assuming that 80% of our target = 1.0s then 100% = x (the target cast time). thus 100 / 80 = 1.25s. This means we need our base cast time (1.5s) reduced by .25s.
if 100% cast time = 1.5s we need to determine what amount of haste will reduce us by .25s. the ratio we're using is 1575 rating = 100% reduction. so we have to convert .25s into a percentage of 1.5s. so (.25 x 100) / 1.5 = 16.666...%
if 1575 = 100% reduction, then (575 x 16.6666...%) / 100 = haste rating needed = 262.5 hate rating.
Yes, of course the socket bonuses come into play. The issue I was addressing was how to evaluate whether 10 haste or 12 spell damage is better for you. Naturally in something like [Gloves of the Tempest], if you evaluate 10 haste to be better, you should use a [Quick Lionseye]. If you evaluate 12 SD to be better, you're better off with a [Reckless Pyrestone].
Because everyone is debating about the cap, I'm gonna show you my math here so that I'm not just throwing around a random number. I did this all with very simple ratios and cross multiplying for instance:
1.0s = 80%
x = 100% Solve for x by doing (1.0s x 100%) / 80% = x
assuming 1.0s is ideal timing for ramped AB cast we know that approx 1575 haste = 100% and we know that the 20% increase from IV is what (in my model) would put us at 1.0s so therefore our cast time needs to be at a time that will reduce to 1.0s when IV is activated.
so assuming that 80% of our target = 1.0s then 100% = x (the target cast time). thus 100 / 80 = 1.25s. This means we need our base cast time (1.5s) reduced by .25s.
if 100% cast time = 1.5s we need to determine what amount of haste will reduce us by .25s. the ratio we're using is 1575 rating = 100% reduction. so we have to convert .25s into a percentage of 1.5s. so (.25 x 100) / 1.5 = 16.666...%
if 1575 = 100% reduction, then (575 x 16.6666...%) / 100 = haste rating needed = 262.5 hate rating.
Except that you are Wrong
Very simple, Haste rating adds to itself and then multiplies with haste effects (Icy Veins and Bloodlust).
Originally Posted by Rounced
To expand on the notion of haste and how it works/doesn't work with a 40/0/21 spec I ran some numbers to check on things.
With Bloodlust active you need 244 haste to get AB down to a 1 second cast.
With Icy Veins active you need 393 (392.5 to be precise) haste to get AB down to 1 second.
With Icy Veins and Bloodlust active you need 384 to get Frostbolt down to 1 second.
So even if you have the Skull of Gul'dan you would be able to make use of Spell haste up to a rating of 209 in regards to stacking cooldowns.
I'm relatively new to these boards as a poster, but I would like to share my thoughts on the stats evaluation for arcane. Being an (the only) arcane mage myself in a relatively high end guild (working on Twins), I found it very difficult to judge how much exactly 1 int/spi/haste/crit is worth compared to +1 dmg. So I decided to make a relatively simple model (in Mathematica) for a concrete fight like Brutallus. I'll explain the model briefly, and then show the results that were quite surprising to me at least.
As arcane you use almost exclusively 2 spells, which is frostbolt and 3stack arcane blast. Given a fight of length T, you can calculate, based on your personal stats, the total amount of mana M that you have available to spend in the fight. If we call the amount of ab's you cast x, and the amount of frostbolts y, there is a best value for x and y that maximizes your total damage done (subject to the time and mana constraints). This is a simple linear programming problem which can be solved easily. As an example, if the fight is very long, you'll be forced to cast relatively many frostbolts to save up on mana, but if the fight is short you can spam ab exclusively. The relative worth of each stat is then calculated by changing that stat slightly and noting the change in total damage done.
Using as input my own stats, and assuming all standard raid buffs and consumables, and a relatively "standard group" (for me at least ) of spriest/resto shaman, I get the following results:
1 dmg := 1
1 int = 1.312
1 spi = 0.981
1 crit rating = 0.732
1 haste rating = 0.997
Two things surprised me in this. Haste is not quite as bad as I thought it would be considering arcane is such a mana intensive spec. I deliberately passed on all haste gear in Sunwell so far because I believed it wouldn't add a lot to my dps. The second remarkable result is that int is very effective. Int adds more damage through mindmastery, more crit, more total mana, more passive regen through the new 2.4 regen formula, and more mana back from evocation/mana tides. Based on this information, I'm probably going to socket all yellow slots with 10 int gems from now on.
I should note that this is also assuming that you use mana pots/gems on cd, evocate once, and receive 1 innervate from a druid. But without an innervate it would only increase the value of int/spi even more.
I also compared my results with Rawr. It more or less agrees for the value of haste/crit, but it values spi/int much lower. I don't know why this is. I'm not too familiar with the workings of the mage model of Rawr, so I can't comment on this.
I also compared my results with Rawr. It more or less agrees for the value of haste/crit, but it values spi/int much lower. I don't know why this is. I'm not too familiar with the workings of the mage model of Rawr, so I can't comment on this.
Your results almost exactly match the results of Rawr, when the situation is fully defined, for my wife's mage.
I am very suprised with this result too, can you tell us the factors you included in your program?
I remembered a post in the 2.3 thread showed 1 extra mana is about 2.2 extra dmg (by replacing FB with AB), which made me think spirit/int is not the way to go.
Things I included which might not be clear, because I went after our usual raid composition for Brutallus, is JoW and Imp Divine Spirit. Basically I included everything that has a long term effect on the average damage/mana cost/casting time of both ab and fb. Things I didn't include is short term buffs like Arcane Power/Heroism/Icy Veins.
Things I included which might not be clear, because I went after our usual raid composition for Brutallus, is JoW and Imp Divine Spirit. Basically I included everything that has a long term effect on the average damage/mana cost/casting time of both ab and fb. Things I didn't include is short term buffs like Arcane Power/Heroism/Icy Veins.
Considering how important those cooldowns are to a mage's dps, any simulation that does not include them is going to be of questionable value.
I remembered a post in the 2.3 thread showed 1 extra mana is about 2.2 extra dmg (by replacing FB with AB), which made me think spirit/int is not the way to go.
If you just look at the mana added to your pool, it's not impressive. It shows up as high ranking because it affects every damage stat besides haste. Socketing a 10 int gem at 750 int and 320 spi--fairly low stats, if you're stacking int/spi, you'll be in the 870/450 range with raid bufs--gives (assuming BoK):
* 12.65 intellect (non-gnome)
* 189.75 mana
* 3.43 OOCmp5 or 2.06 mp5
* 113.85 mana from evo (151.8 2t6)
* .16% crit
* 3.16 +damage (4.05 +dam spellfire)
Assuming a 6 minute (300 second) fight, the total contribution to your mana is 189.75+113.85+2.06*(300/5) = 436.2 mana, or about 2/3 of a fully stacked AB, which is a significant amount from a single gem, not to mention the +dam/crit equivalent to a [Potent Flame Spessarite].
Stacking int, in turn, makes spirit a better stat. Socketing a 10 spirit gem yields:
Naturally if you never run out of mana (you're chain-innervated, fight is short, extended interruptions, etc) and can spam AB indefinitely, the relative value of spirit drops like a rock, int is reduced to its pure damage stats, and the dps contribution of the various stats comes to look more or less like it does for fire and frost.
Just out of curiosity, is there a reason behind calculating the spell haste cap for 1 second AB spam while under the effects of IV and not for BL/Heroism? For short enough fights the time spent under the effects of Bloodlust and Icy Veins are equal assuming that one of the abilities doesn't have remaining duration at the end of the boss fight. In longer fights assuming that mages only receive one bloodlust (an arcane mage could arguably handle several effectively assuming they were able to get an innervate.) and the number of uses of IV increases it probably makes more sense to increase your spell haste beyond the threshold for Bloodlust.
Does anyone know if Rawr or any other tool has a module that's able to calculate the gain/loss of increasing your spell haste beyond any of the caps? Basically all it would need to do is add up the amount of time spent below the global cooldown waiting to cast again and compare it to the amount of time gained while not under those conditions. With that data we could probably derive a simple formula to take into account exactly how much the value of spell haste decays when compared to other stats.
It's probably not as much of an issue now, but as more gear becomes available in Sunwell it becomes somewhat difficult to avoid stacking spell haste.
* 12.65 intellect (non-gnome)
* 189.75 mana
* 3.43 OOCmp5 or 2.06 mp5
* 113.85 mana from evo (151.8 2t6)
* .16% crit
* 3.16 +damage (4.05 +dam spellfire)
Intellect also receives multiplier from clear casting, and if you have, mana tide.
Spirit becomes better when you are human, receive innervate, or when the fight has high down time.
Arcane is so situational that one may need to switch gear for different fights/group setup.
It seems every arcane mage need to use Rawr to optimise his/her gear, however one result from Rawr troubles me, the % of dmg done by Frb or cycle is very high compared to all my raid experience (4/5 Mh, 3/9 Bt). Usually I do 0-25% dmg with Frb, while Rawr shows 40%+.
It seems every arcane mage need to use Rawr to optimise his/her gear, however one result from Rawr troubles me, the % of dmg done by Frb or cycle is very high compared to all my raid experience (4/5 Mh, 3/9 Bt). Usually I do 0-25% dmg with Frb, while Rawr shows 40%+.
This is usually because in real raids your dps time won't be 100%. Have you tried lowering your dps uptime setting in Rawr?
Alright, this may not be the thread for this question but I figure that this is probably the best place to get mages that know what they are talking about to answer it so here it goes. My guild, like most, has always given mages a shadow priest over other casters dating back to the days of 2/5 T5 arcane mages with mana problems. However, recently I watched our Brutallus kill video from one of our fire mages POV's and noticed that with a shadow priest and resto shaman in the group he never went below 50% mana and this was while using Molten Armor and Destruction Potions (0 mana potions). We run a ret pali and have JoW up constantly which probably helped him too.
So my question is this, do full-T6 level geared fire mages really need a shadow priest? From what I can tell, if your not having mana problems standing still nuking for 6 straight minutes on Brutallus you probably wouldn't have them on most other fights. Is there something I'm missing here, or maybe he just really lucky in the video with JoW procs/mana gem crits?
Now that gems and odd gear choices have been addressed somewhat, just a quick note regarding weapon enchants: It appears that many people running a 40/0/21 or such choose a +40 spell damage enchant.
However, let's make an odd choice here as well: It is clear that (very naively put) as soon as around 80% of one's casts are arcane blast, Sunfire is just as good*. This totally leaves cooldown stacking out of the picture. If Arcane Power/Icy Veins (Icy Veins just increases the benefit from AP in this case) is taken into account, I imagine the break even point to be clearly lower.
Now the only thing to do is to take a look at individual WWS reports and make an educated decision?
*:
50*x + 0*(1-x) > 40*x + 40 * (1-x) where x is the number of arcane blasts.
x>40/50
(Edit) Note: Actually the calculation is not complete. Frostbolts "receive" slightly more Damage bonus (per cast) because they crit for 225%, however on the other hand they receive a 5% penalty due to their chill effect. Even if we take a maledicted CoS into account this slightly shifts the favor in the direction of the +40 Spelldamage enchant. Nevertheless the point remains.
Alright, this may not be the thread for this question but I figure that this is probably the best place to get mages that know what they are talking about to answer it so here it goes. My guild, like most, has always given mages a shadow priest over other casters dating back to the days of 2/5 T5 arcane mages with mana problems. However, recently I watched our Brutallus kill video from one of our fire mages POV's and noticed that with a shadow priest and resto shaman in the group he never went below 50% mana and this was while using Molten Armor and Destruction Potions (0 mana potions). We run a ret pali and have JoW up constantly which probably helped him too.
So my question is this, do full-T6 level geared fire mages really need a shadow priest? From what I can tell, if your not having mana problems standing still nuking for 6 straight minutes on Brutallus you probably wouldn't have them on most other fights. Is there something I'm missing here, or maybe he just really lucky in the video with JoW procs/mana gem crits?
They would go oom without the shadow priest. Letting them flamecap and destro pot is a signficant amount of dps, especially if you lust beloe 20% boss hp. If he was using mana gems, which I doubt, then he was wrong to use them. Now youre talking about Sunwell, so do mages need a spriest to decurse, max dps on brutallus, and aoe on Felmyst? Yes. Where else is the spriest going to go that benefits the raid more? Not really anywhere.
@Prod
Warlocks receive a tremendous boost from shadow priests. So raising the question of whether it's a good idea to put the fire mages in the spriest group is a good one.
@seferis
Interesting... And have you done the calculations on Spellsurge? How many arcane mages in a group would have to have spellsurge to make it worthwhile? Especially if you use the weapon swap addon? I'd do it myself, but I'm not arcane, and don't understand the spec as well as you, I suspect.
@Ammanas
In my guild, we usually have the destro locks/arcane mages with the spriest, and I find as a fire mage, I'd rather be using my party slots for boomkin/resto shaman (or elem if possible). While these may not be optimal for a raid, some raids use them. And given the choice between an spriest and a resto shaman, I'll go with the resto shaman (WOA, BL, and mana tide, coupled with not always having 100% dps time due to movement), and leave the spriest for the other classes. So I would answer your question by saying "fire mages *benefit* from shadow priests, but you really have to thoroughly examine your raid composition to get a meaningful answer."
Smog - Mana gained by Vampiric Touch - 23,985
Figs - 18194
Manly - 20694
No other outside source brought more than 6000 mana (manly's JoW, excluding that 4000)
You may see a mage with half mana, but do tell me where he can find 20,000 or so mana. A smart evocation is 8k mana (~100 dps), 3 super manas is 7200 (~80 dps loss).
Yes, you could technically put them in group 5, but get ready to lose 180 or so dps just in terms of mana compensating.
Strictly speaking, you should be getting an SP -and- a shaman at the same time. Failing that, I'd rather have a shaman than an SP as a firemage. With the exception of Brutallus.
While answering the question "do firemages really need an SP?" is valid, the question you should be answering is "how can the raid change to accommodate better RDPS" or more likely "swap firemage with X".
Or respec all your mages arcane, drop CoE for another CoD, SP/Rshaman for everyone, with SP/Eshaman for the locks.
Edit: Searix does have a point in terms of quantity of mana returned. The fact that with an SP you're never "low" doesn't mean you'd be "above 0" without one. At this level it truly is the case that prologued SP regen is many times more than any other comparable regen. It's such that we're either overflowing with one, or strained without one.
I personally raid as Arcane/frost, gathering up gear to go Fire versus Brutallus, as all my models hint at an increasing superiority of fire as ilvl 156 gear steps into the game, making 2 pieces of T5 obsolete.
However, with the recent changes in mana regeneration mechanics, and talking to our resto druids, i've found out Innervate is not as important to healers as it was pre-2.4.
I therefore asked one of our druids to put Innervate on me during bossfights where he doesn't need it, with a result of a net increase in my dps, allowed mainly by the fact i can spam AB during all cooldowns, opposed to the ability of doing it only for the first AP/IV use or so.
By having 2 innervates this became even more clear, as i hardly had to use any frostbolt during these fights.
Even though anedoctal evidences are not something i've ever trusted, this made a question arise: is it possible, and how many innervates are needed for an arcane/frost spec to deal more DPS than a fire one - given the best gear available?
@Prod
@seferis
Interesting... And have you done the calculations on Spellsurge? How many arcane mages in a group would have to have spellsurge to make it worthwhile? Especially if you use the weapon swap addon? I'd do it myself, but I'm not arcane, and don't understand the spec as well as you, I suspect.
I think spellsurge has been addressed somewhat in other threads. It is a 15% Proc with a 30sec cooldown.
Anyway, an option is to just switch weapons when casting frostbolts. Then the issue boils down to whether you want 40 spelldamage or spellsurge.
Obviously, since the 20% of your casts that are Frostbolts
- are likely cast when you are short of mana
- and you will not stack cooldowns during these 20%
it is an interesting choice to swap to a spellsurge weapon when casting Frostbolts.
04/25/08, 11:07 AM
) of spriest/resto shaman, I get the following results:
Similar Threads
