Fire Mages are alive again (Updated for 6.2.3)

[Touchthis]

Hi,
How many targets are needed for Tharbek's Lucky Pebble (or any of the mythic dungeon mastery on-use trinkets) to outperform Desecrated Shadowmoon Insignia? (this is assuming sustained cleave on these targets)

The mythic dungeon mastery on use trinkets bring you more burst dmg, it outperform Shadowmoon Insignia on short fights too. I use these trinkets on Council Kormrok Iron Reaver Xhul and Iskar.

[Searix]

i'm assuming mastery on-use trinkets are also better for single target too?

Why does it only have to be cleave

[skiz]

@searix: It is an combustion scaling issue.
More targets => combustion does more % of your damage
Shorter fight => combustion does more % of your damage.
In both cases mastery scales way better than that meh int procc from DSI.

However, even on long (singeltarget) fights those mastery trinkets are equally with DSI. But getting high itemlevel ones is the problem. Atleast that is what I've seen from data.

[Searix]

I have mythic socketed dsi so question is very relevant here.

Why is it higher than on use mastery trinkets for long single target fights though? It only seems to occasionally line up with cooldowns, and when it's not it seems rather mediocre.

Just saying someone should check the math. I wouldn't be surprised if always having mastery use trinket to pair with cooldowns wasn't > dsi that only (1/3rd of the time?) pairs with cooldowns beyond first

[Curnivore]

^People often forget the conservation phases (of Fire or Arcane, maybe more so on Fire) carry a weight to the overall damage too even if they don't look good in the short term. But, I would not attempt to do the math and I would not suggest to a regular player to do any math. Run a simulation because to be perfectly frank with you, the Fire community has a lot of conjecture and limited napkin math that gives misleading answers and it's in much higher need of holistic simulations, provided the script is coded correctly (I heard it is recently).

[Frosted]

@searix: It is an combustion scaling issue.
More targets => combustion does more % of your damage
Shorter fight => combustion does more % of your damage.
In both cases mastery scales way better than that meh int procc from DSI.

However, even on long fights those mastery trinkets are equally with DSI. But getting high itemlevel ones is the problem. Atleast that is what I've seen from data.

Am I missing a post that compares the different trinkets across various conditions? (length, #targets, mainly)

[Searix]

@searix: It is an combustion scaling issue.
More targets => combustion does more % of your damage
Shorter fight => combustion does more % of your damage.
In both cases mastery scales way better than that meh int procc from DSI.

However, even on long fights those mastery trinkets are equally with DSI. But getting high itemlevel ones is the problem. Atleast that is what I've seen from data.

Am I missing a post that compares the different trinkets across various conditions? (length, #targets, mainly)

Kinda where I'm at - I'm not sure where or if skiz's info is in the thread. We need a new thread, or we need a clean-up of some sort there's a ton of good information (chinese vs. superchinese, crit cap, gear to use) etc throughout the thread but otherwise not much.

On the note of simulations - Simulationcraft seems to be a bit behind on the theorycraft. Between not pre-pyroing, then scorching until you have full stacks (which is best for 5sec rings and especially 0sec rings), and im not sure if using chinese, it seems to lag very far behind.

What i do know is various top parses do use the mastery trinkets

[Frosted]

Anything that requires running towards the target SimC will not be able to do. You cannot dynamically alter player/target distance during a sim. So it will always be behind there, for sure.

As for pre-pyro, I don't see why that isn't there (or cannot be done).

I haven't read through this entire threads in some time and unfortunately I'm out of town for about a week - after that if someone could summarize what needs updating in SimC I could get working on it. worst comes to pass I'll just bother wand in IRC for all the details of what needs to be done.

[Velerion]

If for any reason during a fight ring needs to be held, or trinket and ring just get really desynched,is the better option to wait til ring is popped or set up combustion when your trinket procs?

[skiz]

@searix: It is an combustion scaling issue.
More targets => combustion does more % of your damage
Shorter fight => combustion does more % of your damage.
In both cases mastery scales way better than that meh int procc from DSI.

However, even on long fights those mastery trinkets are equally with DSI. But getting high itemlevel ones is the problem. Atleast that is what I've seen from data.

Am I missing a post that compares the different trinkets across various conditions? (length, #targets, mainly)

No post from me about that issue as far as I remember. I just simmed the "worst" case with the highest potential itemlevel some of those trinkets can have some time ago. I.e. singel target, ~450 seconds. I don't even remember if I took http://www.wowhead.com/item=110019/xeri ... us=642:760" target="_blank or one of those on use trinkets.

Anyway, if searix wants to find out what kind of gear he wants to wear, he should sim himself. Amount of targets, fightlengths you're interested in and which kind of trinkets you have matter. Simcraft may be slightly behind theoretically, but you'll get answers within +/- 10% I guess. Especially, on short fights (<90 seconds) simcrafts results should be accurate.

Another way to look at it is guessing with logs, if that is more what you do want. Seeing some 705 on-use-mastery trinkets on archimonde rankings beeing competitive with DSI. Some ~695 on-use-mastery trinkets on council rankings.

P.S. I doubt it is worth to optimize combustion in simcraft when it is removed in legion. (In terms of updating the default APL - Changing code to fake / introduce running into simcraft - fixing fireball / pyroblast timing issues)

[Searix]

@searix: It is an combustion scaling issue.
More targets => combustion does more % of your damage
Shorter fight => combustion does more % of your damage.
In both cases mastery scales way better than that meh int procc from DSI.

However, even on long fights those mastery trinkets are equally with DSI. But getting high itemlevel ones is the problem. Atleast that is what I've seen from data.

Am I missing a post that compares the different trinkets across various conditions? (length, #targets, mainly)

No post from me about that issue as far as I remember. I just simmed the "worst" case with the highest potential itemlevel some of those trinkets can have some time ago. I.e. singel target, ~450 seconds. I don't even remember if I took http://www.wowhead.com/item=110019/xeri ... us=642:760" target="_blank or one of those on use trinkets.

Anyway, if searix wants to find out what kind of gear he wants to wear, he should sim himself. Amount of targets, fightlengths you're interested in and which kind of trinkets you have matter. Simcraft may be slightly behind theoretically, but you'll get answers within +/- 10% I guess. Especially, on short fights (<90 seconds) simcrafts results should be accurate.

Another way to look at it is guessing with logs, if that is more what you do want. Seeing some 705 on-use-mastery trinkets on archimonde rankings beeing competitive with DSI. Some ~695 on-use-mastery trinkets on council rankings.

P.S. I doubt it is worth to optimize combustion in simcraft when it is removed in legion. (In terms of updating the default APL - Changing code to fake / introduce running into simcraft - fixing fireball / pyroblast timing issues)

Biggest problem is default API doesn't use on-use trinkets.

Luckily it's actually pretty easy to sim. Using the default profiles (which handles non-lust bloodlust combustions terribly: Fireball -> Pyrox3???):
Mage_Fire_T18M : 135347 dps
Mage_Fire_T18M_Kyraks : 117702 dps

It has 49.77% mastery. Max rank kyraks gives 25.61% mastery according to wowhead, an increase of 51.46% of our combustion damage.

So we can simply increase our combustion (or any profile's) combustion damage by 51.46%. Combustion with the Kyrak's profile does 19586 dps, or rather it does 10,078 extra dps giving us a total of 127,780 dps.

It's not over for the Kyrak profile however, as we now also do more damage during Nithramus. Assuming we get 6 seconds of spread combustion, and 4? seconds of crystal combustion during Nithramus, and assuming it hits for ~200k during lust on the pull (i hit for 135k on my profile, so +50% is ~200k), that's ~10 ticks of 75k extra per tick combustion = 750k bonus damage during ring from combustion, then an extra 44% from ring = ~1.08 million extra damage we're not accounting for every 2 minutes (and a little less over non-lust, we'll say 600k for an average of ~760k per ring).

760k every 2 minutes is 6333, now giving us 134,113 dps, even with the default profile. Again however with 2 unoptimized combustion rotations (I do a minimum of 100k combustions without lust, with sandmans/council with the same gear as mythic profile), the 87924 average combustion the profile lists is pretty low.

TL;DR: Fairly certain single-target max level kyrak's is better with Chinese/Super-Chinese by ~10k on single target, a LOT more with any other adds.

[skiz]

Wow. default apl seems to be really broken. Still using a different simcraft engine here, but I would guess that those lines should help out:

Code: Select all

gear_crit_rating=2229 actions=stop_pyro_chain,if=prev_off_gcd.combustion actions+=/counterspell,if=target.debuff.casting.react #on use crit on trinket1 #actions+=/use_item,slot=trinket1,if=time<5 actions+=/call_action_list,name=movement,if=raid_event.movement.exists actions+=/rune_of_power,if=buff.rune_of_power.remains<cast_time actions+=/call_action_list,name=combust_sequence,if=pyro_chain actions+=/call_action_list,name=crystal_sequence,if=talent.prismatic_crystal.enabled&pet.prismatic_crystal.active actions+=/call_action_list,name=init_combust,if=!pyro_chain # Utilize level 90 active talents while avoiding pyro munching actions+=/rune_of_power,if=buff.rune_of_power.remains<action.fireball.execute_time+gcd.max&!(buff.heating_up.up&action.fireball.in_flight) actions+=/mirror_image,if=!(buff.heating_up.up&action.fireball.in_flight) actions+=/call_action_list,name=pre_combust,if=time<10&(talent.prismatic_crystal.enabled&!pet.prismatic_crystal.active) actions+=/call_action_list,name=aoe,if=active_enemies>10 actions+=/call_action_list,name=single_target #Build up 3 stacks with actions.pre_combust=/scorch #actions.pre_combust=/fireball copy="xeritacs_unhatched_egg_sac" trinket2=xeritacs_unhatched_egg_sac,id=110019,bonus_id=642/760,initial_cd=4

1. Ensure crit cap. Default profile should be slightly below. That crit value in there is a little bit to high for sandmans, but it works also with the on use trinket.
2. rewrote actions to start scorch spamming early on. I want that for initial CD shenanigans: equipping egg 6 seconds before combat triggers a 10 second ICD on it. So no early procc.
3. Copy and trinket change.

Default apl with above changes. I really just copied them under it. 450 second fighttime and singeltarget. Any optimized profile will probably perform better. (10k dps gain on default apl is because of critcap.)

(On-use is mastery on use. It is slightly behind that egg on shorter fights.)

80 seconds +/- 10% variance, 1 target: Sims below are without an upgraded egg, yip I forgot it.


150 seconds +/- 10% variance, 1 target:


150 seconds +/- 10% variance, 2 targets:


If you want to use on use mastery trinkets, try adding them here. This is just an example, you'll find that part in the apl, change it yourself.

Code: Select all

actions.combust_sequence+=/blood_fury actions.combust_sequence+=/berserking actions.combust_sequence+=/arcane_torrent actions.combust_sequence+=/use_item,slot=finger2 # below the 2 new lines actions.combust_sequence+=/use_item,slot=trinket1 actions.combust_sequence+=/use_item,slot=trinket2

P.S. have to mention that I accidently compared Egg to Iron reaver piston last time. Still mostly the same results.

Edit:
P.P.S.
Default apl has this line:

Code: Select all

actions.combust_sequence+=/fireball,if=crit_pct_current-1>(1000%13)&prev_gcd.pyroblast&buff.pyroblast.up&buff.heating_up.up&12-pet.prismatic_crystal.remains<action.fireball.execute_time+3*gcd.max&spell_haste<0.7

Which is simply during bloodlust: Fireball -> Pyro -> Fireball ->Pyro x3 -> Infernoblast. Kind of decent without running.

P.P.P.S.
Reverted crit change, so fire profiles aren't critcapped. Singeltarget 80 seconds, 10 % variance.

[Searix]

<3

If what I'm reading from your post is correct, single target Fire with on-use should be higher single target than Arcane, which makes sense imo

[Komma]

Code: Select all

gear_crit_rating=2229

I mean...you did sort of just add 700 crit rating with that statement alone. Even with the default APL, it's a ~7% increase to overall damage.

Code: Select all

180638 51.7% extra_crit 168632 48.3% Mage_Fire_T18M

180 seconds +-20%, fixed time, Patchwerk

The default profile is intentionally under crit cap. It's about the best balance of crit/mastery as you can get with the default HFC mythic items. Warforged items are not permitted under default profile rules, even more so when we're talking about fully upgraded 735 super-warforged crit trinkets from mythic dungeons. And for anyone using a similar profile, an few gem slots or warforged pieces would easily reach required crit levels.

I really question the value of using Scorch at all within the APL. Scorch is a 1.0 cast time compared to what, 1.4 for Fireball? The increase in proc triggers is tiny. In practice, there also seems to be a slightly higher proc rate on Sandman's when compared to simulations (See: reports of Felmouth Frenzy triggering it). I'd like to see controlled experiments run on that to identify any gains.

If I were to put any time into the current Fire APL, I'd be most interested in crit-cap FB -> Pyro -> Pyro -> IB -> Pyro -> Pyro combos. It'd be easiest to implement and most likely to yield interesting results. For now though, most efforts will be on Legion developments.

[skiz]

Code: Select all

gear_crit_rating=2229

I mean...you did sort of just add 700 crit rating with that statement alone. Even with the default APL, it's a ~7% increase to overall damage.

Code: Select all

180638 51.7% extra_crit 168632 48.3% Mage_Fire_T18M

180 seconds +-20%, fixed time, Patchwerk

To avoid misunderstandings of my above post: the default APL in that post also benefits from that gear_crit change. Basically, alll lines under the "copy=" thingy are just in the new mage-APL named "xeritacs_unhatched_egg_sac". Everything above that copy line also affects the default APL in my post. I should have renamed it. I do know komma knows it, but someone probably may have problems understanding that.

If I were to put any time into the current Fire APL, I'd be most interested in crit-cap FB -> Pyro -> Pyro -> IB -> Pyro -> Pyro combos. It'd be easiest to implement and most likely to yield interesting results. For now though, most efforts will be on Legion developments.

I would add:
- Scorch for proccs before burn. (with sandmans)
- Burn with sandmans or with PC.
- non pyrocamping "conserve" with ~47+ % crit.

If what I'm reading from your post is correct, single target Fire with on-use should be higher single target than Arcane, which makes sense imo

Just saying that average singeltarget burst from fire is higher than from arcane. Singel target fire vs arcane is a gear and fight time issue. And I'm pretty sure that fire is ahead more often than arcane on singel target. Keep in mind that arcanes burst is more spiky: fire's burst is done over 30 seconds and arcane's over 16. So if you've a fight time near 16 seconds arcane is probably the winner.
https://www.warcraftlogs.com/statistics ... espan=1000" target="_blank Just disable every other spec, and assume everyone plays perfectly? Hard to say yes to such things. It is more like: Is arcane the best specc for singel target? No. Is fire the best specc for singel target ? No. Is frost the best specc for singel target? No.

[Zayele]

My apologies if this was discussed earlier in the thread, but under what conditions can you completely stop DPS in order to prevent Sandman's from activating? On fights where Nithramus is used 3 to 4 times, especially, the ICD gets completely desynchronized with the ring. Is it always the best choice to do nothing in such a scenario and wait until ~5s are left for Nithramus to come off cooldown?

[Searix]

Yes

It's ok to desync it for a few seconds but not having Sandmans kills your burst

[prateems]

Superchinese: Fireball->Pyro->Fireball->Pyro->Pyro->Infernoblast->Pyro->Pyro->Combustion->Ib while running in

Yeah, thats the bad part of giving names to rotations. Someone improves it and calls it super. Don't blame me for that name. It is simply chinese rotation with a higher ignite. You'll probably miss 1 global of blastwave. Adding another fireball->pyro in front of this one is useless. Very sensitive to haste and running. You want 700-900 haste optimal, but even with lower haste this one is awesome. Start running at the start of your Infernoblast. Here is a picture how it should look. If ignite and pyroblast have switched positions that pyroblast hits before ignite - then you started to run to early. (Yellow marked events)

What set pieces do we rely on to get 700 ish haste? Seems a bit difficult to do this given mythic gear, but perhaps I've just tried the wrong gear combinations so far.

[Devidose]

What set pieces do we rely on to get 700 ish haste? Seems a bit difficult to do this given mythic gear, but perhaps I've just tried the wrong gear combinations so far.

Mythic tier chest and shoulders are the only pieces that give haste, but together with mythic Edict [all non warforged] you still have 904 [339 (chest) + 168 (shoulders) + 397 (edict)].

[laguna]

I have seen a lot of recent logs on council either doing 4 pyro+IB or 4 pyro instead of 5 pyro. Are the first 2 better options on council?