[5.2] PTR Standardized DPS Testing - Come Help!

[zomgDPS]

Aye.. time to get back to the purpose of this thread:

Did some heavy testing on the latest build over the past night and today, I have some conclusions that I am pretty confident about, but first, the reports:

Reports
Build: Arcane | RoP | Mastery
Test: 1
Result: ~67k
Strategy: Nothing special, ABx4 AMs ABr

Build: Arcane | Inv | Mastery
Test: 1
Result: ~62k
Strategy: Tried both strategies. First one a stright up 'burn' (i.e. AB spam with AMs as they come), also tried the 'mixed mode' strategy of my OP, that being, actually 'conserving' in the first ~30-35 secs of the Invoc buff, then burning away and repeating. Both strats were pretty close in the multiple runs of each I tried.

Build: Arcane | Inv | Haste
Test: 1
Result: ~57k
Strategy: Burns. ABs with AMs, never use ABrs

Build: Arcane | RoP | Mastery
Test: 2
Result: ~81k+
Strategy: Nothing special, ABx4 AMs ABr. Used AP twice (unglyphed)

Build: Arcane | Inv | Haste
Test: 2
Result: ~76k+
Strategy: Tried both 'burn' as well as 'mixed mode'. Both were close, which makes sense given the AT phase for both is identical.


Conclusions/Analysis
I want you guys to pay close attention to Invocation Haste.
It is being trounced by Mastery RoP at every corner. Mastery RoP is producing better test 2 and test 1 results. I have a strong suspicion that Haste Arcane is dying with 5.2

Heck, on a whim I even tried Mastery Invocation runs, and as the results show, they were producing better results compared to Haste Invocation.


Also, the DPS nerfs can be felt in full force now. There is no doubt about it. There is a MASSIVE nerf to Arcane's top end burst (especially during AT). All the way from peak output as well as maximum damage during AT has been nerfed hard. I peak much much higher with Frost (I will get you some of my frost test results shortly).


With Throne of Thunder looking friendlier to RoP, I am struggling very hard to see why someone would ever take Invocation as Arcane. If you guys can produce different results please do post them.
The irony is, Blizz tried so hard to change Invocation Arcane, but it may end up being the more useless Arcane flavor come 5.2

---------- Post added 2013-02-22 at 02:49 PM ----------

Strategy: ye olde scorch rotation but with glyphed FFB as a filler | Keep mana > 70% hardcast AB to not let stacks drop off in case of no procs | use AB to not cap mana in case of good procs.

Ah good. I wanted to test this out too. It was on my list. It seems you are getting respectable results with it so far.

I will try to get some numbers on this strat as well. "FFB weaving" if we want to call it

[Azalis]

Reports

Build: Arcane | RoP | Haste)
Test: 1
Result: ~70k
Strategy: Arcane Barrage when no missiles, at 3 or 4 arcane charges, and below 90% mana. Basically I just replaced scorch with arcane barrage. If I start to fall behind on mana I dump at 1 or 2 stacks to catch up. I used pure haste gems in yellow sockets.

Build: Arcane | RoP | Haste
Test: 2
Result: ~80k
Strategy: Same as above

Build: Arcane | RoP | Haste
Test: 3
Result: ~84k
Strategy: Same as above

---------- Post added 2013-02-23 at 08:26 PM ----------

Reports

Build: Frost | Invocation | Haste)
Test: 1
Result: ~69k
Strategy: Full int in red sockets, int+haste in yellow, haste+hit in green

Build: Frost | Invocation | Haste
Test: 2
Result: ~87k
Strategy: Same as above

Build: Frost | Invocation | Haste
Test: 3
Result: ~82k
Strategy: Same as above

[Shangalar]

As of this moment, you can find me on the PTR on the Brill EU server under the name Shangtesting.

[Deleted]

Report:
Build: Fire | Invocation | Crit Build
Test's: 3
Result: 96-98k
Strategy: Patchwork (10min). 25.02.2012
Full Crit sockets (yellow), Crit-Int (red), Int-Hit (blue)
use LvL15 #1 Talent
use LvL75 #2 Talent
use LvL90 #1 Talent
I start with Selbuffed Int, MA, 275 Bufffood, Int Flask (i read later about this Thread so sry :/ )
Typical base Rotation FB, IF, Pyro, LB.
I got with Berserk (Troll), Mirrors, AT and Presence of Mind all time a minimum of 4 Pyros for my Combustion, and between the 3 min Cooldown of berserk and Mirrors, 3 Pyros (FIX) für the next Combustion. I Rotate on this. I test 3 Times all time the same amount of DMG. Gear: 507 (all 5 Set of T15)

[Shangalar]

How did you get to 507 gear? I have 502/503 with the 5-set and everything else Tyrannical from the vendor.

[Swizzle]

How did you get to 507 gear? I have 502/503 with the 5-set and everything else Tyrannical from the vendor.

You can buy the 522 trinket from the same vendor that sells tier. That may push your total to ~506-508 iirc.

[Shangalar]

It pushed me to 503 lol... I have it in my inventory. Two pvp trinkets give me 502, 1 pvp and the 522 trinket set me at 503.

[Swizzle]

It pushed me to 503 lol... I have it in my inventory. Two pvp trinkets give me 502, 1 pvp and the 522 trinket set me at 503.

Really? Are you sure you have all the right gear?

[Shangalar]

Every piece on me is Tyrannical 493 except the 5 T15 set items which are 522.

[Deleted]

How did you get to 507 gear? I have 502/503 with the 5-set and everything else Tyrannical from the vendor.

i played with my Char but i havn't uses the 522 Trinket.

[Shangalar]

Which probably means that you have a better/upgraded weapon and/or offhand?

[Methusula]

i played with my Char but i havn't uses the 522 Trinket.

You need to read the rules of the test and follow them. Your test if not run properly, means nothing.

[Deleted]

yeah just read later about this, so i can'T go for the rules...but i will do few hours

[Shangalar]

So I’ve finished doing some extensive testing and here are the results divided into... a lot of things.

First some basic info:
- gear: 5-set T15 522, everything else Tyrannical 493, trinkets were PvP Mastery trinkets without any procs or useful PvE abilities to minimize proc influence
- for Test 3 I used a flask, food and potions during the fight, along with Time Warp
- Arcane Macro: Brilliant Mana Gem, Arcane Power, Presence of Mind, Alter Time, Arcane Missiles
- unless specified differently, all testing was mostly ABx4, Missiles, Barrage (with bombs ofc)
- I forgot to enchant the pants
- since we’re doing this without any buffs, the NT ticks are all screwed up compared to what they’d be on live

Chapter 1 – Arcane (with Nether Tempest)

1. Mastery
Stats: 21943 SP (24137 with Brilliance), 3778 Haste, 10183 Mastery (13183 with Mage Armor)
a) Rune of Power
Test 1: 66-69k
Test 1 with Fire Blast weaving: 62-65k
Test 1 with indiscriminate use of Missiles: 57k
Comments: it’s better to “waste” potential Missiles procs than to use them before max charges *unless* you have both Missiles before the 3rd Charge
Test 2: 75-76k
Test 3: 76-79k
Comments: with RoP, AB 1&2 are mana positive, during Power only AB1 is positive
Comments: with RoP it’s not possible to do anything except clear with Barrage, possible that without the 4-set bonus we can throw in an extra Blast every now and then but with the bonus the opportunities for it are very rare
b) Invocation
Test 1: 56k
Test 2: 71k
Comments: not working

2. Intellect (I tried gemming around Intellect instead of Mastery)
- Stats: SP 25154, Mastery 8423 (-1700 Mastery and +1000 SP compared to the Mastery build, buffs included)
a) Rune of Power
Test 1: 62-64k
Test 2: 70k
Comments: Mastery is stronger than Spell Power for RoP (nothing new here)
b) Invocation
Test 1: 56k
Comments: just no.
Comments: 3 charges are enough to go from 10-15% mana to full with Evocation

3. Semi-Haste
- Stats: SP 24137, Haste 5083, Mastery 11927 (buffs included)
a) Rune of Power
Test 1: 64-67k
Test 2: 72-75k
Test 3: 85-87k
b) Invocation
Test 1: 55-57k
Test 2: 68-70k
Test 3: 70-72k

4. Semi-Haste NT tick
- Stats: 5508 Haste (5334 is without the buff what 3056 is with the buff), 11466 Mastery
a) Rune of Power
Test 1: 65-67k
Test 2: 75-78k
Test 3: 85-87k
b) Invocation
Test 1: 51-54k (no point continuing)

5. Full Haste with Frost Armor
- Stats: Haste 9016 (next NT tick without raid haste is 8860), Mastery 4877
a) Rune of Power
Test 1: 62-63k
Test 2: 69-71k
Test 3: 83-85k

This is as far as I got today. Still need to test Haste+Invocation and then moving on to Fire.

[Ilyakimina]

Arcane without Scorch-weaving has become completely static. You need RoP to make it more or less work, and you need to stay put. Even if with Ice Floes you can move once for a few yards every 45 seconds, you need to place another RoP on the floor, vs evocating every minute.

Taking into account the fact that you just can't maintain both specs playable at the same time (you need to regem/reforge and probably will have different gear priorities), we can conclude the following:

On the tests that you are performing, Arcane should be ABOVE Fire in a static hit-the-dummy fight for it to be worth it. 4-5% at least. That way, you would underperform Fire in fights with a lot of movement and cleave fights, and outperform it in single or multitarget fights when you can stand and be a dps-cannon.

Your tests so far have not produced that results; on the contrary, Arcane seems to be 5-7% below Fire on single target and no movement fights (correct me if I'm wrong please). One week away from 5.2, I don't think anything can be done now: unless your guild or yourself don't care about the dmg output, we will all go Fire.

[Minimemage]

I know the following won't be useful in the spirit of this testing thread, but I wanted to share my own testing results with my character. My mage is here.

You must understand that the most important information for me - is how my actually geared toon will perform in each spec come 5.2. Obviously much less useful for everyone else. I copied my character as-is, and tailored the gearing and gemming and enchants to the best of my ability for each spec using whatever I had banked or was wearing. No bought tier items or otherwise since they will not be available when I start T15.

My tests are approximately 3 minute sessions with flask and 300 food. I go until AT comes back up and note the final DPS number.

Arcane:
Mastery gems, Mastery reforging, Mage Armor, RoP, stay > 85% mana however necessary, AM at higher stacks if possible, don't sit on 2 AM.

Average of ~94200 dps on boss dummy across testing.

Frost:
Haste gems, Haste reforging, Frost Armor, RoP

Average of ~96180 dps on boss dummy with last night's PTR patch across testing.

Fire:
Crit gems, Crit reforging, Molten, RoP. I got to about 9000 crit rating with the gear I had available.

Average of ~81300 dps on boss dummy across testing. I beleive I need much more testing here since RNG is such a large factor.

TLDR: My take home lesson here is that because I haven't been obtaining the right crit gear all along in T14, fire is not an option for me starting out on T15. Frost will probably be my choice for the first time in a long time, and I'll switch to Fire when I can approach the "right" amount of crit rating (11-12k).

[Ilyakimina]

Btw, thanks to all of you guys for your testings. A very special thanks to you guys that have stuck to the procedure specified in the OP.

[Soulstrike]

my gear set is in my armory semi haste build with mastery gemming using mage armor no flask or food was used or time warp during rotation

stats: 26941 spell power 15.1% haste 11349 mastery(with mage armor) hit capped ofc

ilvl504

arcane

test 1:77k dps
test 2: 81k dps
test 3 : 79k dps

rotation: pretty simple AB until 4stack and use arcane missile only at 4stack even if you have 2 AM <4stack of AC, then refresh stack with Abarr then repeat
i have noticed that my mana net is gain when i use AB x4>2 AM>100% mana so i try to fish another AM with AB i will be at 88% mana when i fire my Abarr at this point this only works when you have 2stacks of Am procs ready at 4stacks otherwise you will lose mana and lose dps

at the start i try to get 2AM AT combo get going by refreshing my AC at 2stack with Abarr and pray for good AM procs :P
ABx2>barr>ABx4>(2AM proc) use AP& AT combo begins

i did the same testing on live and numbers dont look pretty compare to ptr >_>

test1:98k dps
test2:100k dps
test3:97k dps

so yea it seems i lost 20k dps comparing from ptr and live but hopefully arcane's scaling will improve my dps overall later in tier(probably go fire with all those crit gears tbh)

edit: @minimemage looks like our ilvl is about 5-6 different whats your dps like on live atm??
does arcane scale that amazing or did i do something wrong?? will have another test going tonight(maybe it was just lag screwing me up)

[SpaceDuck]

@minimemage (so popular today huh! :P ) when you say you went with haste gems for your frost test, was that pure haste in yellow sockets and int+haste/int+hit (or other) hybrids in everything else? if so, have you tried with pure int in reds and int+haste hyrbid everywhere else (even in the yellow sockets)?

[Minimemage]

Hello!

@Soulstrike
Here's a log from last night's raid as arcane. I would say I was about average performance, I died a few times during the night. Empress was decent, about 120k.
http://www.worldoflogs.com/reports/r...qfr/dashboard/

I don't max haste or mastery on live, I balance both after an extra NT tick and run Frost armor. All my sims have pointed me in this direction and it results in good performance in raids. However on PTR I've tested the three types (Haste max, Mastery max, and balanced), and mastery is a clear winner as it doesn't take long to get to 4 stack and we're often clearing. This is anecdotal based on my personal testing, there's a reason you should set SimC to run 10000 iterations to give you good results. Unfortunately our tools aren't ready yet so doing this is the best I have.

@Spaceduck
Yes, exactly as you mention, 320's in yellows, and int+haste, hit+haste for appropriate colors. Until SimC plugs in the mage spell changes I can't generate my stat weights as usual to know if int:haste is > 2:1 and therefore worth gemming for. But it would be interesting to test and see.

Edit: I also wanted to clarify that I've take a page from Elpadlol's book (armory) on the PTR and tried fire again with only 2xT14 and other offset crit pieces. Its much more promising (above 90k) but dry spells still cause me a lot of frustration. I don't think it will take many 522 upgrades to get there, but might be too unreliable for progression raiding until I have a crit rating closer to what he's running.