[Discussion] Patch 6.2 PTR

[Imnick]

UM becomes more valuable when you have Archimonde's trinket, but keep in mind that he's an end-of-tier boss.

[Bashlow]

He drops 720 item lvl on heroic, so we will see a lot of class trinkets in the first ID

[Frosted]

I want to talk a bit about Arcane and gearing next patch since we're now extremely close to the patch coming out and they've gone ahead and done some tuning of our pets. This could all be blown away if they nerf the pets more, but it's close enough to 6.2 that it's time to just bite the bullet and start talking about what we're going to do if things stay close to how they are now.

To start with, there are currently two options for Arcane that you'll find discussed so far in this thread:

The first is a haste heavy build, effectively gearing Haste > Everything.
The second is the traditionally Mastery heavy building, gearing Mastery > Everything.

First, why is haste even an option? It's entirely due to the T18 2pc, the pets. 100% due to them. it is not the ring, the class trinket (trinket sim spoilers: you don't use the class trinket when going full haste).

To show this, here are the comparisons of T18NORMAL mode gear sets, with tier bonuses enabled vs disabled.

Tier Bonuses Enabled:

Code: Select all

Generating reports... DPS Ranking: 127346 100.0% Raid 64375 50.6% Mage_Arcane_T18N_haste 62972 49.4% Mage_Arcane_T18N_mastery

Tier Bonuses Disabled:

Code: Select all

Generating reports... 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, coincidentally, there is a single haste rating 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 save myself a bunch of time, I'm just going to do Normal, Heroic, and Mythic ilvl comparisons instead of a ilvl increase spectrum like I've done in the past. If there is sufficient doubt this is enough to justify my conclusions, I will do a more through look across the ilvl spectrum.

From above, the Normal ilvl comparisons (and their haste ratings)

Code: Select all

Generating reports... 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 comparisons:

Code: Select all

Generating reports... 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 comparisons:

Code: Select all

Generating reports... 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 profiles above are just ilvl scaled, the mythic ones are not using the legendary ring as they are in the default mythic profiles)

Above we see that as we are 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% 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, it is not 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, and why it's pretty important to gear haste 20% > 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. If someone would like this expanded upon, I can, but that would be an entire post in itself I think. 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. It would look something like this:



Here I've included the original breakpoints discussed above (the 20%, and the unreachable high 40%/50%) in green as a way to frame the additional breakpoints that emerge, which are in yellow.

We can see that some of these are above 30% haste, which is unreachable in haste geared mythic profiles, so we can throw those out as being so unreasonably high as to be non-important to us.



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! We should be able to use this "breakpoint" then as the target haste value 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 the most optimal haste value to shoot for in HFC.

Conclusions:

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.

tldr - 1800 haste > mastery > whatever.

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.

[Frosted]

I ran some very quick tier comparisons for people who want to know if they should pick up normal T18 or not. I basically just took the default T17 mythic profile and replaced it with the amount of tier shown here.

[Ianes]

So arcane is completely dependant inT18 set pieces, which seem straight up OP. Meanwhile, frost's set bonuses are virtually irrelevant. Also, which tier pieces did you replace?

[Frosted]

I think legs/hands for 2pc, and chest/shoulders for 4pc. It doesn't really matter though, the DPS change from using another piece would be minor.

[Zyrachiel]

Apologies if this was addressed somewhere already, but I don't have nearly enough time to browse AT as I wish I had.

How likely is it that the comfort of having the same gear for fire and arcane survives in 6.2? Will we still be able to stack mastery and use arcane for ST, and fire for MT?

[Dutchmagoz]

Apologies if this was addressed somewhere already, but I don't have nearly enough time to browse AT as I wish I had.

How likely is it that the comfort of having the same gear for fire and arcane survives in 6.2? Will we still be able to stack mastery and use arcane for ST, and fire for MT?

It won't be quite as efficient as it is now, but still possible. You just will only be able to use fire on true AoE/cleave fights, not fights with a bit of aoe/cleave but like half single target.

[Zyrachiel]

It won't be quite as efficient as it is now, but still possible. You just will only be able to use fire on true AoE/cleave fights, not fights with a bit of aoe/cleave but like half single target.

Thank you for answering, I understand.
If I may, what sort of setup are you personally going for in 6.2? Assuming there will be no further changes from the current PTR build.

[Killget]

I ran some very quick tier comparisons for people who want to know if they should pick up normal T18 or not. I basically just took the default T17 mythic profile and replaced it with the amount of tier shown here.

Hey Frosted,

Just to make sure, the frost/arcane/fire t17M, i presume as they are default mythicT17 they do not have T17 set bonuses included??

i.e. is the gain of 2pct18 including the loss of 4->2pcT17 for each of the specs? or would this be calculable?

[Frosted]

It's the transition from T17 to T18. So It's from T174pc -> T172pc_T182pc -> T18 2pc.

[Caimion]

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 save myself a bunch of time, I'm just going to do Normal, Heroic, and Mythic ilvl comparisons instead of a ilvl increase spectrum like I've done in the past. If there is sufficient doubt this is enough to justify my conclusions, I will do a more through look across the ilvl spectrum.

Are the Heroic and Mythic sims using the same talents for both sets of gear? Is it possible that the phenomenon DeaThShiNoBi observed, where different talents are suited to different gearsets, may change the 'breakpoint' conclusion? Meaning that at Normal ilevels, we gear haste due to the breakpoint phenomenon, but at higher ilevels we still gear for haste due to favourable scaling on UM_ROP_PC w/the class trinket - with the end result where haste is preferred across the ilevel spectrum?

[Frosted]

Talents don't shift anything around.

I have a follow up post prepared that goes further into the haste vs mastery stuff for Arcane next tier that should hopefully make the decision between the two make sense. However, as it's heavily influenced by pets, I'm going to hold off on posting it until this weekend incase we get more pet tuning this week. (haste heavy builds don't use the class trinket, btw).

[Caimion]

Talents don't shift anything around.

I have a follow up post prepared that goes further into the haste vs mastery stuff for Arcane next tier that should hopefully make the decision between the two make sense. However, as it's heavily influenced by pets, I'm going to hold off on posting it until this weekend incase we get more pet tuning this week. (haste heavy builds don't use the class trinket, btw).

Good to know, thanks! I figured you had more, since you've dropped the hint about not using the class trinket a few times now Saves me having to run my own sims. I'm rather sad about the class trinket, neither Fire or Arcane will get to use it

[GriffKentner]

Good to know, thanks! I figured you had more, since you've dropped the hint about not using the class trinket a few times now Saves me having to run my own sims. I'm rather sad about the class trinket, neither Fire or Arcane will get to use it

Funny thing is, Frost won't be using it either! Way to go Blizz. *golf clap*

[Lavok]

This might be a stupid question so shoot me if it is, but how does the class trinket modify the arcane spell priorities under AP precisely? Does it just become AB>all and is this the case also for all ilvl versions of the trinket? Lastly, is it worth getting the normal ilvl trinket over BRF mythic ones? I've seen rough comparisons for the transition between the tiers concerning set bonuses, but failed to find similar ones for the trinkets.

(All of that is discounting the above discussed possible haste build.)

[Equio]

(First of all sorry for any mistakes, english isn't my native language)

So I've found this loot comparison spreadsheets for warlocks. https://docs.google.com/spreadsheets/d/ ... =871207953
I was wondering if I could use that spreadsheet for my mage after modifying the stats priority. Would it work or is there a better alternative since I'm still confused about the top stats vs higher itemlevel items.

[Chev]

(First of all sorry for any mistakes, english isn't my native language)

So I've found this loot comparison spreadsheets for warlocks. https://docs.google.com/spreadsheets/d/ ... =871207953
I was wondering if I could use that spreadsheet for my mage after modifying the stats priority. Would it work or is there a better alternative since I'm still confused about the top stats vs higher itemlevel items.

This is very similar to what Berlina and I did at the start of 6.0 (now long out of date). One thing I have noticed with their calculation is they are not taking into account the spec attainment. E.g. Crit is increased for Fire Mages by 15%.

[Rabona]

Sorry if I'm accidentally making the math geeks vomit, but can't you just multiply the stat weights of crit(/ms/mastery) with 1.15 to account for that?

[Frosted]

(First of all sorry for any mistakes, english isn't my native language)

So I've found this loot comparison spreadsheets for warlocks. https://docs.google.com/spreadsheets/d/ ... =871207953
I was wondering if I could use that spreadsheet for my mage after modifying the stats priority. Would it work or is there a better alternative since I'm still confused about the top stats vs higher itemlevel items.

This is very similar to what Berlina and I did at the start of 6.0 (now long out of date). One thing I have noticed with their calculation is they are not taking into account the spec attainment. E.g. Crit is increased for Fire Mages by 15%.

You guys should update that gear sheet!