Berlinia's Arcane Mage Guide for T17 (Highmaul Update)

[gameorg]

Looks like arcane is on top, according to warcraftlogs of 99 percentile for heroic fights. It is not a big gap and its depending on boss fight aswell. But i assume that arcane can pull ahead the more arcane optimal gear is out, since everyone is running around with multistrike atm.

Does someone have an updated stat priorty chart?

[Chev]

Does someone have an updated stat priorty chart?

http://altered-time.com/forum/viewtopic.php?f=3&t=833" target="_blank

[gameorg]

I saw that one, but we cannot optain that gear yet... But i guess it provides a good indication where we should go.

[Chev]

yes you can. Both sets of data were been done with attainable gear.

[gameorg]

yes you can. Both sets of data were been done with attainable gear.

shame on me for not reading the first paragraph. Thanks for patience

[oomy]

Looking at the 99th percentile isn't very accurate lol

[Xenost]

Hi guys,
I have a question regarding arcane rotation, I use Simcraft Action priority list to min/max my rotation but I find something weird.
The talents used are SN/ROP/PC
According to simcraft, I should use AP if the CD of PC is over 15s, regardless of being in burn phase.
During the sim the problem doesn't appear because the CD is off right when burn phase start (due to perfect conditions ...) but the problem can appear in real conditions if I have to delay the burn phase due to movement/mechanics.

Maybe I miss something in Simcraft that may cover it and doesn't allow to AP if burn phase is about to start, but it seems like burn phase always start right away...

Thank you in advance.

Edit : What it seems like in synthesis :
-Never delay burn phase
-Delay AP for PC if PC.cd <15s
-Delay PC for AP (except for glyphed AP)

There definetly must be something I miss because of that :
init_crystal
prismatic_crystal,if=buff.arcane_charge.stack=4&cooldown.arcane_power.remains<0.5

That means that if AP and PC are off CD in conserv phase, it will use PC and go into Crystal_sequence, which i can't find in the tabs :s.

[tedronai]

I've a question on the burn phase. Back in Cata where we originally had burn and conserve phases, it was relatively easy to time because you just did AB spam the whole time until you evocated with clearcasting seeming to add a very small variation in burn phase length. But now, with free arcane missiles in the mix, the time it takes getting under 50% seems to vary significantly if you happen to get a good string of AM procs. Or in a more mundane example, you have to circumnavigate a fight mechanic that requires a bit of movement, keeping you from spamming AB and thus regenning mana.

So, how long should a burn phase last after the AP buff has expired? It seems the longer you hold off Evocation (when it's ready and off cd), the more your AP will come off cooldown before you're within time range to start another burn phase. Is it worth continuing to spam down your mana (and potentially get more AM process) and then wait to use AP later once Evo's CD is close enough to start another burn? Or is better to prematurely end a burn phase with more than 50% mana so that AP and Evo's timing can stay relative?

[Dokem]

I have the same doubts myself. WHat is the priority with burn phase and prismatic crystal usage, to decide when to use your cooldowns.

That's include supernova and he trinket shards of nothing.

SUpernova: on cooldown, but you want to save 2 or at least 1 for each PC.
AP: ideally have AP for each PC and during burn phases,
Shards of nothing: on PC.

I dont think you can have all that perfectly aligned. THings will change on each type of encounter and this is a priority list not a rotation. You need to favor PC, then BURN, then conserve, in the usage of your CDs and supernova.

correct me if im wrong. Im waiting what Barlinia has to say about post buff changes,

The top talent combo seems to be: SN/ROP/PC. + cold snap for the extra PoMs. On fights with too much movement I'll still use IF over ROP though, im not that pro to use ROP at the best in fights like Twins.

[gameorg]

Looking at the 99th percentile isn't very accurate lol

why not? Iam interested in max possible dps to compare, not the average one.

[Berlinia]

Looking at the 99th % shows what the spec can do. Blizzard tries to make specs playable for the 70% classes, but it's not their responsibility if a player can't play their class correctly.

[Chev]

With the Burn phase, you want to get to around ~50% before using Evocate to get you back up to ~100 but it's not the end of the world if you Evo before or after you hit 50%. If you get missiles while burning, you will want to exhaust them before using Evo. It is a DPS loss not to use missiles when they are available and you are at 4 stacks of Arcane Charge.



Here is the comparison for using missiles during the burn and not using missiles. (missiles are still used during the conserve phase)


With the Crystal, you want to use it as close to on CD as possible unless Arcane Power will come off CD while crystal is off CD or you are not at 4 stacks of Arcane Charge. When you have put the crystal down, hope the boss wont move for the next 12 seconds, pop any personal CD's that you have, refresh NT or use any SN charges that you have and then go into your burn phase as normal. To go into your burn phase, you want to be around 90% Mana when you put the crystal down so your mastery is not effected too much by Mana levels while the crystal is up and so you definitely don't have to waste time using Evo while they Crystal is up.

[Neuroburner]

I'd tried 3 crystal rotations, no burn, 50-60% burn and 10-20% burns. Self buff, no BL, 60% mastery(ilvl 642) 3AP/Crystal after 3rd AP I burned to evo + 1 nova after evocation 5 attempts for each rotation. First 50-60% burn and no burn was pretty same 16k dps, but 10-20% burn give me 18k. I tried it when after Mar'gok figth I got 18k dps with no burn rotation and second mage got 22 with 10-20% burn.

[Chev]

I have updated my last post to include <40% as the burn target and ran it at 25K iterations.

The sudden difference for the sub 30% burns is probably due to the sim's decision on when to initiate the burn. I will have a look at the APL to see if I can change those as well rather than just the target but I don't have the time just yet.

[Dokem]

I do more DPS burning to 20-30% mana, than burning to lets say 10% or burning to 50-60%. I'm not theorycrafter, but I can notice the difference. Anyone else notice that?

[Chev]

Excluding the 10 - 30% burn results as those are going to be effected by poorly optimised APL for that play style, you can see the mana level that you decide to burn to makes up about 1% of your final DPS.

Chances are, you have had other factors which have effected your DPS between fights. E.g. More Missile procs, less movement, less casts interrupted, the tank didn't move when put your Crystal down. It could also be that by not focusing on your mana levels allows you to focus more on other aspects of your rotation which have a bigger pay off. Any of those could make it seem like one play style seems better than another. Especially when comparing a handful of fights. This is why we use Simulation Craft. On average, each play style is tested 10,000 times or more. For you to do this manually, assuming you did the same fight 20 times a night, 7 days a week, it would take you 16 months to get the same level of confidence with your results.

To further explain this, this is what an average set of results looks like for a single simulation.

As you can see there is huge difference between absolute worst night of RoboMage's life when nothing went right (15k) and when his stars aligned and he managed to pull of 20k DPS (p.s. this screen shot was taken prior to the Fire buffs) but on average for the 10k fights, RoboMage did 17k DPS and this is what Simulation Craft returns as the result for what ever you were running the simulation for.

[Vog]

With the Burn phase, you want to get to around ~50% before using Evocate to get you back up to ~100 but it's not the end of the world if you Evo before or after you hit 50%. If you get missiles while burning, you will want to exhaust them before using Evo. It is a DPS loss not to use missiles when they are available and you are at 4 stacks of Arcane Charge.

Could you please share the way in which you made the comparisons between mana percentages?

When I try

Code: Select all

copy=below30 actions.burn+=/evocation,interrupt_if=mana.pct>92,if=time_to_die>10&mana.pct<30

at 1 iteration and then look at the mana graph I notice it's not working. I guess I could make a separate mage.simc for every mana percentage I want to test and call those, but that's quite tedious.

[Komma]

With the Burn phase, you want to get to around ~50% before using Evocate to get you back up to ~100 but it's not the end of the world if you Evo before or after you hit 50%. If you get missiles while burning, you will want to exhaust them before using Evo. It is a DPS loss not to use missiles when they are available and you are at 4 stacks of Arcane Charge.

Could you please share the way in which you made the comparisons between mana percentages?

When I try

Code: Select all

copy=below30 actions.burn+=/evocation,interrupt_if=mana.pct>92,if=time_to_die>10&mana.pct<30

at 1 iteration and then look at the mana graph I notice it's not working. I guess I could make a separate mage.simc for every mana percentage I want to test and call those, but that's quite tedious.

You need to make a new copy of the entire "actions.burn" section, starting with the "actions.burn=..." line. What you listed just adds one line to the existing subaction list.

[Vog]

Thanks Komma.

I've been running some sims at 250k iterations for various mana percentages to initiate evocation (ranging from 30% to 60% in steps of 5%) and for various durations (ranging from 60 to 600 seconds) and for various talent setups (untalented, Nova/IV/PC and NT/IV/AO). I'm consistently seeing that for the first evocation, the optimal percentage to initiate evocating is lower than for the subsequent ones.

For the first evocation, I get the best results between 35%-45% depending on talent setup
For subsequent evocations, I get the best results with initiating the evocate at ~55% regardless of talent setup.
The most puzzling thing is that these results hold up even when disabling talents, bloodlust, trinket procs, weapon procs, and AP.

Is there any way to distinguish between the first cast of evocation and subsequent casts in the APL (maybe through time_to_die)?

[Komma]

For the first evocation, I get the best results between 35%-45% depending on talent setup
For subsequent evocations, I get the best results with initiating the evocate at ~55% regardless of talent setup.
The most puzzling thing is that these results hold up even when disabling talents, bloodlust, trinket procs, weapon procs, and AP.

This is likely due to Bloodlust increasing Evocation's mana gain through Nether Attunement.

The answer to this would not be distinguishing between different evocations, but instead changing the target mana level based on current haste and the remaining duration of the haste buff.