Arcane Mage - Stats, Talents, Trinkets and Racials

Alright, with the rising number of Mages in the guild, who can benefit from this, i think a continuation of the Haste v Mastery topic is in order.

Initially intended to include everything in the main post, but thats one long of a topic, so lets try to address it here. NB: Take the exact DPS values with a pinch of salt, as they are still tweaking the Pets' DPS, but the general idea and conclusions shouldn't change.

To start off, need to repeat that the whole reason why Haste is gaining momentum this patch is entirely because of the pets from T18, whose cast time is also affected by your haste rating. Haste has always been a strong stat for Arcane, but traditionally it was never a prio. To illustrate, untill now we've had to options:

The first is the traditional Mastery heavy build, gearing Mastery > Everything.
The second is a haste heavy build, effectively gearing Haste > Everything (up to a point, which we will discuss below).

To show the difference between the two and how T18 set bonuses affects this, lets look at the DPS ranking excerpts below:

Tier Bonus enabled:

DPS Ranking:

127346 100.0% Raid

64375 50.6% Mage_Arcane_T18N_haste

62972 49.4% Mage_Arcane_T18N_mastery

Tier Bonus disabled:
DPS Ranking:

102815 100.0% Raid

53154 51.7% Mage_Arcane_T18N_mastery

49662 48.3% Mage_Arcane_T18N_haste

It's pretty obvious the set bonuses are causing this haste phenomenon. I've checked the effects of the 2pc vs 4pc for their contributions to the increase in haste scaling, and the 2pc accounts for the VAST majority of it's increase in power. Over 90% of the increase to its value is attributed to the 2pc alone, the 4pc is an minor increase to haste in comparison.

So clearly, the pets are doing something and the reasoning behind it pretty obvious - the pets have breakpoints.

Similar to old DoT breakpoints, the pets have haste breakpoints in that at some haste level they will gain an additional attack. Because the pets are a significant damage increase, the result is that this breakpoint has a significant effect on the overall gearing scheme of Arcane. Here is a snapshot of the number of attacks each pet gets as haste increases. (Jaina uses a 3s base "attack" speed, the other three are all base 2s and are lumped together because of it).



So, luckily, there is a single haste rating (the one at the bottom, to the extent its not obvious) where all of the pets gain an additional attack! This is why the breakpoint is so significant - you essentially have four breakpoints collapsed together, resulting is one large spike in DPS because of it.

If it's true that this haste strength is just the breakpoint effect, we should see mastery re-gain strength and overtake haste as we increase item level. To check that, lets look at at a Normal, Heroic, and Mythic ilvl comparisons:

Normal ilvl gear comparison (same as above, but this time including the haste ratings)

DPS Ranking:

127346 100.0%  Raid

 64375  50.6%  Mage_Arcane_T18N_haste ( 26.09% haste )

 62972  49.4%  Mage_Arcane_T18N_mastery ( 17.53% haste )
Heroic ilvl comparison:

DPS Ranking:

151761 100.0%  Raid

 75971  50.1%  Mage_Arcane_T18H_haste ( 28.31% haste )

 75790  49.9%  Mage_Arcane_T18H_mastery ( 19.24% haste )
Mythic Ilvl comparison:

DPS Ranking:

184771 100.0%  Raid

 94139  50.9%  Mage_Arcane_T18M_mastery ( 21.23% haste )

 90632  49.1%  Mage_Arcane_T18M_haste ( 30.88% haste )
Note: the T18N, T18H and T18M gear profiles are just ilvl scaled 
Also, when we speak of Haste %, i mean Raid-buffed Haste.

So, from the above it is clear that with increasing ilvl, the mastery geared profile approaches and overtakes the haste geared profile, eventually becoming far ahead of it as it just surpasses the breakpoint. I think this is a pretty good case for the strength of haste being entirely derived from the 20% raid haste breakpoint. There are additional breakpoints past 20% though. Looking again at the spreadsheet I showed above, but at a higher haste range this time:




There are additional breakpoints, but they are well beyond what the haste arcane profile can obtain without heroism, and they are not stacked up nicely like the 20% haste breakpoint is - so even if it is achieved, the DPS spike will not be as strong as the 20% haste breakpoint and therefore less likely to shift the strength so heavily in hastes favor.

Confounding Factors

Above I think is a clearly laid out explanation of the 20% haste breakpoint (though feel free to list any questions or comments below), and why it's pretty important to gear haste 20% > Int > mastery > everything else. Unfortunately, World of Warcraft is not a nice game. Pets in World of Warcraft are subject to latency (latency beyond what we normally experience as players). They cannot (or do not by design) chain cast their spells or attacks. To keep it concise - Pets have delay. They have delay in acquiring their target, in swapping targets, and in queuing up a spell to follow the spell they are casting. Additionally, the magnitude of this delay is not consistent. Most importantly, this delay is random. (Or, seemingly of such an odd behavior to be called random, or random enough). sometimes they delay, sometimes they do not.


Example:


One summon of jaina may chain cast all of her spells back to back, getting a perfect 10s of uptime - therefore resulting in the above mentioned breakpoints. But what if jaina experiences some delay between casting? Say, she has some amount of delay (which is a random amount between each cast, as small as 200ms and upwards of 1s), which totals to 2s of "delay", or wasted uptime. If we want to model this by saying that essentially jaina is reduced from a 10s uptime pet to a 8s uptime pet (which chaincasts here), then the breakpoints shift! What if every pet had this happen, randomly, throughout the fight? What if there were delays that resulted in values between 7 and 10s of uptime? Now instead of a sharp, powerful breakpoint at 20%, you have multiple breakpoints that are randomly being accessed throughout the fight depending on how much delay the pets experience. 

A detailed analysis of those breakpoints will include some which are way above the 'reachable' 30% haste rating, so to avoid over-complication, lets just look at those who are reasonably achievable with HC/Mythic gear:



So we have two additional breakpoints that are 1) not covered by our initial 20% value, and 2) are not ridiculously out of our reach. If we refer back to our profiles above, the heroic geared profile reaches the 26% haste breakpoint that will cover all reasonable cases of delay for the base 2s pets (Arthas, Sylvanas, and Tyrandre) while also being equal in DPS to the mastery geared profile (mastery begins to overtake haste at around 26% Haste in reality, with the recent nerfs into account)! We should be able to use this "breakpoint" then as one of the target haste values to gear for. It will maximize the value of haste both in an idealized SimC universe where we are always at the 20% haste value, but also help us to encompass the RNG induced breakpoints that are incurred when dealing with the delay which comes along with pet attacks. I do not believe that the final breakpoint for jaina alone is worth the additional 200 haste/mastery tradeoff. As we can see with the mythic profiles, once you have gone beyond the breakpoint haste value quickly decays below mastery again. Assuming that the pets will eventually deal around equal amounts of damage, I think the 26% haste value is probably the most optimal haste value to shoot for in HFC.


Summary & Conclusion

The T18 2pc pets introduce haste breakpoints which affect the way arcane mages want to gear in 6.2. The idealized breakpoint is 20% raid buffed (or 1350 rating on gear). However, due to pet delays which have a noticeable impact on the amount of attacks pets will get off, 26% raid buffed (or 1800 rating on gear) may be the most optimal haste rating to gear for. This should be reasonably obtainable by mages in heroic Hellfire Citadel gear.

Nevertheless, we do need to account for the recent Pet DPS Nerf and actually consider how much of the optimal gear will we have the chance to obtain (realistically), so this leaves us with two gearing profiles, which are quite close to each other in terms of output:

1) Conservative Haste Gearing - Haste 20% (1300rating) > Int > Mastery > Haste > Mult > Crit > Vers
. If you cannot easily get to the 26% haste, then i would suggest sticking to this, until you actually can reach 26%. Also note that if you have latency issues yourself, then sticking to the 20% haste might be better for you. 

2) Aggressive Haste Gearing - Haste 20% (1350rating) > Int > Haste (~25-26%) > Mastery > Haste > Mult > Crit > Vers. I do like haste in general and if i can get to 26% haste easily, i would use this profile personally.


Tangent - If you agree, there will be different trinket priorities. Mastery values the mage trinket very high, while haste values the insignia. Both value the prophecy of fear high. But it think its fair to say that Insignia Mythic or HC is much more reachable than the Mage trinket, so practically, that should affect us much