State of the Specs (Feb 2, pre-BRF)

[Komma]

Tomorrow, Blackrock Foundry is scheduled for release on Normal and Heroic. Similar to last time, this post is meant as a short summary of how the specs perform, relative to each other. There have only been a few Mage related hotfixes since the release of Mythic Highmaul, which means that things have not changed much. The only major difference is the inclusion of set bonuses.

Most of the following is not news, and just summarizes what we've found over the last few weeks. If you've been following our forums and chat closely, most of it will sound familiar.

As is the case with most of my posts, this is based off extrapolation from SimulationCraft results and my personal experiences. The following simulations use T17 Heroic (680) sample profiles. They are meant as suggestive data, and do not represent accurate absolute differences between specs. 6.1 PTR data is NOT used, due to it having a later release date.

Spec comparisons
Before going into comparisons, it is worth mentioning that all 3 specs are looking very strong. Among ranged classes, Mages more than hold their own in any sort of encounter, and it is not hard to justify your raid slot regardless of what you decide to play.

Single target: Arcane > Frost >= Fire

As Highmaul demonstrated, Arcane likely provides the highest single target DPS in the entire game. In Highmaul, Frost was a strong performer as well, while Fire was slightly below average. With better itemization options and very strong set bonuses, Fire catches up quite a bit in BRF.

It is worth noting that this ranking doesn't change if we're talking about boss damage on fights with a lot of adds, similar to Mythic Imperator. Arcane Orb and Thermal Void with increased FoF generation mean that Arcane and Frost gain a lot of single target damage when adds are involved.

Cleave: Fire > Frost > Arcane

With two target cleaves , Fire and Frost look very close on the charts. However, Splitting Ice's unreliability, and Fire's stronger performance on additional targets put Fire solidly ahead in rankings. However, there are rumors of possible improvements to Splitting Ice, so keep an eye out. On the other hand, Arcane is unsurprisingly left in the dust. Arcane's T17 set bonuses affect Arcane Blast and Arcane Missiles, neither of which offer any cleave value.

AOE: Fire >> Frost >> Arcane

Nothing has really changed in terms of sustained AOE. Fire is still a mile ahead, effectively playing the role of a "ranged Enhancement Shaman". Arcane once again suffers on multitarget fights.

Set Bonuses
It is worth looking at the set bonuses in detail to see the effect on overall DPS. SimulationCraft developer Collision was nice enough to provide a chart based on PTR data, which includes the 4T17 Fire set bonus nerf.

To summarize, Arcane's set bonuses provide a strong boost to single target DPS, but does nothing for AOE. Fire's set bonuses, especially the 2-set are extremely strong for all situations. There are also weird tricks that involve Fire's 2-set, which might inflate its value. Frost's 2-set is rather weak, while the 4-set is a strong boost for both single target and AOE, due to Frost Bomb.

Predictions
Overall, I am expecting Mages to be among the strongest ranged regardless of spec.

Arcane will likely be the King of all fights that focus on boss damage. Fire will provide the strongest ranged AOE in the game. This however puts Frost in a weird spot, because Arcane and Fire seem to cover everything and leave no niche for Frost Mages to shine. For this reason, I believe most top performers will be playing a mix of Arcane and Fire.

For the last 16 raid tiers, single target DPS has always defined the strongest spec on most of the hardest encounters. This is because you spend a long amount of time beating up on big bad, even if it's at the expense of cleave or AOE. Blackhand looks to be no different, which leads me to guess that Arcane will be the top choice for many bosses.

Rough and biased guess: Arcane >= Fire > Frost

FAQ
Q: I frequently hear "Fire becomes OP with a lot of gear". Does that mean Fire is the strongest single target spec with full mythic gear?
A: No. Even 10 gem slots and full mythic warforged T17 gear does not put Fire ahead of Arcane.


Q: With so much more critical strike, does Kindling become the strongest talent?
A: No. Kindling in fact doesn't "hyperscale" with crit at all. Kindling will continue to remain the worst talent for single target, but a strong performer cleave/aoe fights. For a more thorough explanation, check Frosted's post.

Q: I don't have a lot of mastery or crit gear. Does that mean Frost is better for me?
A: Not really. Your gear doesn't matter nearly as much as the fight. If it's a mass AOE fight, Fire will perform the best even with full multistrike gear. If you want maximum boss damage, Arcane will probably give you the best numbers regardless of your enchants. Players often overemphasize having the "best stats", even though the difference between fully optimized stats and unoptimized stats is only a few percent.
That said, it's been mentioned that mastery is very strong for both Arcane and Fire. This explains why a lot of mages opting to dual spec Arcane/Fire have chosen to gear for heavy mastery.

Q: You rate Arcane so highly! Doesn't it suffer heavily from movement?
A: Not really. With Ice Floes, Arcane honestly doesn't suffer that much more than Fire and Frost do. While Fire and Frost have instant procs that are usable on the move, proc generation is often dependent on casting filler anyway. Even on hectic fights such as Mythic Imperator, boss damage charts show that Arcane is clearly ahead of Fire and Frost.

[zhengma]

If it's a mass AOE fight, Fire will perform the best even with full multistrike gear...Arcane will probably give you the best numbers regardless of your enchants...mastery is very strong for both Arcane and Fire.

The following discussion are not backed up by sim/parse, which means I'll be grateful if someone can provide some.

If I just spam Flamestrike, it's true that low crit isn't a big issue, and Mastery could possibly be better than crit due to the extra Ignite we can stack on the targets. However, if I seek to spread Combustion from a main target to multiple mobs, I guess low crit will make it much harder to brew a good Combustion. And in single target (which we'll do for at least some period in whichever encounter), crit not only increases damage but also makes the rotation more smooth. The question then is: how many opportunities will we have in BRF to spend the majority of time on spamming Flamestrike?

As for Arcane, before I can obtain more Mastery from the new gears, can I do something e.g. exit the burn phase later to relieve the penalty of lower Mastery? I'm under the impression that before Mastery is adequately stacked, spending as much time as we can in conserve phase isn't that profitable.

I'm asking these questions because I'm obviously in full Frost gear/enchant/gem at the moment, and I'm seeking to switching out during BRF, but before I can get some BRF loots I'll have to work with the gears at hand for some weeks, regardless whether and how I'll re-spec.

EDIT: btw, I just realized that the two of us haven't conversed for months. Are you mad/bored at me for whatever reason that I haven't realized?

[Berlinia]

Q: You rate Arcane so highly! Doesn't it suffer heavily from movement?
A: Not really. With Ice Floes, Arcane honestly doesn't suffer that much more than Fire and Frost do. While Fire and Frost have instant procs that are usable on the move, proc generation is often dependent on casting filler anyway. Even on hectic fights such as Mythic Imperator, boss damage charts show that Arcane is clearly ahead of Fire and Frost.

Also, Arcane Orb!

[Kams]

As for Arcane, before I can obtain more Mastery from the new gears, can I do something e.g. exit the burn phase later to relieve the penalty of lower Mastery? I'm under the impression that before Mastery is adequately stacked, spending as much time as we can in conserve phase isn't that profitable.

Mastery doesn't have any break points that would change the effectiveness of your burning or conserving. Right now everyone has access to the same gear in Highmaul that does't have a ton of mastery and Arcane still does well and the rotations you have seen are based on this gear.

[Bashlow]

In my opinion Frost will be much more reliable in terms of movement than you mentioned. For example Hans&Franz. I think people bellow the 99% percentile underestimate the effect of simplicity of a spec in progress and learning a new encounter and therefor failing less and offering high amount of damage on a sustainable basis.
Frost is also much more forgiving in terms of rotation misstakes while keeping your dps up and acting to a certain encounter mechanic.

Maybe I only try to believe that, because I will go Frost/ Fire for BRF, as the Ice is running in my Veins and Frost being the spec I can play eyeless. We managed 6/7 mythic with 3 raiddays

Thanks again for all the awesome discussions and information you provide in this forum, casuals like me do highly appreciate it.

[zhengma]

Mastery doesn't have any break points that would change the effectiveness of your burning or conserving. Right now everyone has access to the same gear in Highmaul that does't have a ton of mastery and Arcane still does well and the rotations you have seen are based on this gear.

Code: Select all

call_action_list,name=burn,if=time_to_die<mana.pct*0.35*spell_haste|cooldown.evocation.remains<=(mana.pct-30)*0.3*spell_haste|(buff.arcane_power.up&cooldown.evocation.remains<=(mana.pct-30)*0.4*spell_haste)

Just wondering how these "0.35", "0.3" and "0.4" and what-not are derived. Trial and error?

I'm a physicist and am used to approach a question by examining the hypothetical, simplified limits first. Hypothetically, if one has 0 points of Mastery, I don't see any point of having a conserve phase at all. Hypothetically, if one has no other secondary stats on any gear but Mastery, as Mastery -> ∞, a burning phase will eventually becomes detrimental regardless. Since all the realistic cases fall between these two limits, and if we apply the continuity assumption into WoW theorycrafting, we should expect that when Mastery makes up a smaller portion of total secondary stats, it's beneficial (or at least, not as detrimental) to burn more.

These are just thought experiments of a physicist, though. If there's evidence that the line I quote from the APL applies to a wide range of Mastery percentile (in the total secondary stats), I'll just buy it.

[TLTeo]

Mastery isn't what makes us play a burn-conserve style; Arcane charges (or AB stacks) is, and in fact burn-conserve has been a thing for most of Arcane's history since the introduction of Arcane Blast during TBC, MoP being the only exception.

The whole point of it is that the best dps rotation is AB spam (with eventual am/mbar procs), but that would send the mage oom really quickly, so we need some way of conserving mana in between these burst phases. The key then becomes to maximize your time spent in spamming ab, while having enough mana to cast spells for the whole fight.

In TBC you could spec arcane/frost and weave in frostbolts with low-stack ab casts. In WotLK (if possible, i seem to remember that burn/conserve wasn't as important then) you used evo and gems to manage your mana properly and try to get some ab spam phases, especially towards the end of a fight.

Being a physicist myself - mastery adds a lot of numerical complexity to Arcane. When mastery was announced, there was a thread on EJ using markov chains to figure out exactly how it would change the spec, and i remember also seeing a thread using Lagrangians to model it but i can't find that anywhere. It's one of those situations where the system is so complex, your only choice is to use numerical simulations to study it.

[Berlinia]

Zhengma, your method ignores mana as a spell resource. Even at 0 mastery we'd still conserve to have activity uptimr.

[Frosted]

In my opinion Frost will be much more reliable in terms of movement than you mentioned. For example Hans&Franz. I think people bellow the 99% percentile underestimate the effect of simplicity of a spec in progress and learning a new encounter and therefor failing less and offering high amount of damage on a sustainable basis.
Frost is also much more forgiving in terms of rotation misstakes while keeping your dps up and acting to a certain encounter mechanic.

Maybe I only try to believe that, because I will go Frost/ Fire for BRF, as the Ice is running in my Veins and Frost being the spec I can play eyeless. We managed 6/7 mythic with 3 raiddays

Thanks again for all the awesome discussions and information you provide in this forum, casuals like me do highly appreciate it.

If you struggle on mobile DPS for arcane, you can just run fire. The difference in DPS is very small as you approach heroic/mythic ilvls, and you don't have to compromise your fire/arcane set up. Fire will also let you capitalize on the parts of the fight that are 2 target better as well.

[Kams]

I'm a physicist and am used to approach a question by examining the hypothetical, simplified limits first. Hypothetically, if one has 0 points of Mastery, I don't see any point of having a conserve phase at all.

Mage Armor gives you passive mastery, raid buff, food, enchants, gems. If you're playing Arcane you are going to have Mastery. The reason you burn is because you have evocate to get your self back up. The burn phase is also not 100% perfect. Some times your mana is in a way where your last blast leaves you with close to 40% mana and others closer to 50%. Just go with the flow of it and follow the rotation your dps will be what it should be.

[Caimion]

There's also baseline Mastery, is there not? Even in a hypothetical world where you have 0 points in Mastery from gear/buffs/enchants/gems, you still have +16% damage based on mana from baseline.

[Frosted]

There's also baseline Mastery, is there not? Even in a hypothetical world where you have 0 points in Mastery from gear/buffs/enchants/gems, you still have +16% damage based on mana from baseline.

Yup, 8baseline mastery points, so 16% mastery for arcane as you say.

[Komma]

The following discussion are not backed up by sim/parse, which means I'll be grateful if someone can provide some.

Hrm. Okay. I was thinking that Tectus should have made a good example of how "fight means that your gear really doesn't matter much". In general, the "best stat" for a spec seldom leads the "worst stat" by more than a 20-30% margin. The majority of gear contribution to DPS is through having a stat, instead of having the right stat. This is also why higher itemlevel gear is usually better, regardless of itemization.

Here are some sims that looks at what happens if you were to simply use Frost gear for a number of different fights.
240 second single target (similar to Butcher):

Code: Select all

DPS Ranking: 116092 100.0% Raid 42125 36.3% SN_RoP_PC 39143 33.7% Mage_Frost_T17H 34823 30.0% LB_RoP_Kind

Arcane is ahead.

180 second 5 target AOE (sort of Tectus-ish):

Code: Select all

DPS Ranking: 239656 100.0% Raid 118607 49.5% LB_RoP_Kind 71125 29.7% Mage_Frost_T17H 49924 20.8% SN_RoP_PC

Fire is ahead by far.

420 second 2 target cleave (similar to Twins):

Code: Select all

DPS Ranking: 134302 100.0% Raid 46985 35.0% LB_RoP_Kind 45175 33.6% Mage_Frost_T17H 42143 31.4% SN_RoP_PC

Fire is slightly ahead of Frost.

As mentioned, the fight matters much more than your gear when it comes to whichever spec is stronger. With the limited gear options we had preparing for BRF, this is even more true.

EDIT: btw, I just realized that the two of us haven't conversed for months. Are you mad/bored at me for whatever reason that I haven't realized?

I don't have a reason to get mad at anyone who doesn't cause trouble in the forums. It's just that the number of posts makes it hard to respond to all of them.

Mage Armor gives you passive mastery, raid buff, food, enchants, gems. If you're playing Arcane you are going to have Mastery. The reason you burn is because you have evocate to get your self back up. The burn phase is also not 100% perfect. Some times your mana is in a way where your last blast leaves you with close to 40% mana and others closer to 50%. Just go with the flow of it and follow the rotation your dps will be what it should be.

Kams here gives a good explanation and description of how Arcane should be played. Don't overly fixate on exact mana percentages; only robomage needs exactly numbers to understand how to function. Experienced Arcane players understand which factors are most important to the rotation, and adapt on the fly. "Going with the flow" is taking into account these "goals", and making the best of whatever situation/resources you have on your hand.

[Gregory]

Do fire simulations count in 2p17 trick?