[5.2] PTR Standardized DPS Testing - Come Help!

[dementor]

Your experiment is flawed from the beginning. Even if everyone has the same race, the same gear and the same professions, the dps one person does can not be compared to the dps another person does due to: reaction times, players skill, connection speed, etc.

That being said I copied my own character, and if what I find is invaluable to you because of that, so be it. My character on live is a 507 mastery arcane mage. I changed out to get 5 piece T15.

Just using our brain cells we can tell that mastery has no purpose in an invocation arcane build. The reason mastery was so powerful was because players were above 85% mana on live which is not possible on the PTR. Hence the only option for our primary secondary stat would be haste. This leaves us with two questions. Is that value of haste such that 2 haste>1 int? I decided that since almost every class that has haste as its primary secondary stat has 1int>2haste (think frost or spriest) that I would follow suit. Second question is which is the worst secondary stat, crit or mastery. I had decided that since my gear was from live, which virtually has 0 crit on it, I wouldn't be able to do a valid haste/crit build, so I went with Int>>haste>>mastery>crit. As such red gems went int, yellow went int/haste, and blue went int/hit. I also went with frost armor, not mage armor.

Once that is figured out, there are basically 3 different arcane styles you can play between invocation and rune of power.

The first style is burn invocation (evocation>NT>AM (when arcane charge>=4)>AB until mana is gone then repeat with evocation.

The second style is burn/slightly negative invocation where you use the burn evocation play style when cool downs are up, otherwise you slowly drain your mana as to only evocate when invocation buff is about to wear off (AB to 2 or 3 stacks, use whatever AM procs you have then Abarr).

The third style is burn/conserve rune of power where you burn almost all you mana during cooldowns, then use a mana positive rotation so that you are back to full mana by the time your cooldowns are back (AB to 2 or 3 stacks, use whatever AM procs you have then Abarr).

Play style 1 and play style 3 led to almost the same numbers, play style 2 was 5kish lacking.

I would like to test our fire, but again my gear has very little crit that I don't think I could make a valid comparison.

I think did 2 more tests as frost, both using invocation, one with frost bomb, the other with nether tempest. Both frost bomb and nether tempest beat out any of the arcane playstyles by approx 10k, with frost bomb being a little better.

Edit if you want exact numbers, no cleaving
1) 93k dps
2) 91k dps
3) 95k dps
4) 110k dps (frost bomb)
5) 105k dps (nether tempest)

[zomgDPS]

...Just using our brain cells...

Let me know when you get to this part, until then, if you want a soapbox for your own ramblings, I suggest you start a blog.


Otherwise, since the entire concept of this thread seems to have zipped right past you (or maybe through you? given the void left by the lack of those brain cells you mentioned?), I suggest you either stick to the tests or find somewhere else to roam, since I sure as hell do not have time to rip your nonsense apart.

Dementor indeed...

Let me be clear, since it seems some folks here still haven't grasped what this thread is.


This is not a "lets see whose epeen is bigger by producing bigger numbers" thread (trust me, if it were, I would be slapping pretty much all of you right now with my ginormous digital cock).
This is not a "who does the most DPS" thread either.
Neither is this a "we need to accomodate every single variable in the game and get accurate to 9 decimal places DPS rankings" thread. That thread will come at a much later date, when the tools needed to make those measurements are online.


This is a "PTR 5.2 has siginifcantly shifted a lot of core mage play-styles, we need to experiment and see which playstyles are still viable, and which are doing well, and which do not work" thread.
For that purpose, since we are trying to at least get in the right zip code, the variables from the first post will suffice.

Now, you can come here and make bold statements like "Well this and that obviously don't work", but Dementor, in your 36 posts on these forums, you really haven't shown much prowess in the ability to make rational arguments (yea.. I looked), let alone be able to state with any kind of authority your divinations of how mages will play in 5.2.

So yea.. sure.. make your statements, but unless you provide us with some kind of proof, evidence, fact or logic, don't expect to be taken any more seriously that I have taken you right now.

That being said, given that you truly believe in your own badassery. Please, do come to the PTR, gear up a premade toon like I explained in the first post and do some tests, then come back here and post your results.
Better yet, come meet me, in SW, and I'll show you just what a Mastery Invoker Arcanist can do. After all, I'm more than ready to put my money where my mouth is.

Are you?

---------- Post added 2013-02-15 at 05:51 PM ----------

That being said I copied my own character, and if what I find is invaluable to you because of that, so be it....

Edit if you want exact numbers, no cleaving
1) 93k dps
2) 91k dps
3) 95k dps
4) 110k dps (frost bomb)
5) 105k dps (nether tempest)

These results are meaningless, since you obviously failed at following the very basic instructions in the first post.

Again, this is not a "let me just do whatever I want so that I can post big numbers" thread. If you do not follow the rules, you aren't helping.

Changing all the variables in order to post bigger numbers is helping absolutely nothing (not even your ego - since that is being crushed as we speak).

So refrain. Stick to the tests (they really are super simple) or lurk.

[dementor]

You need to get off your high horse sir. It has nothing to do with how high or high low the numbers are. If you can get other images of how big other peoples dicks are out of your head and read my post you would understand.

The 5 numbers listed could have very well been
1) 20k
2) 19.8k
3) 20.7k
4) 24k
5) 22.9k

It was about searching for general trends of which play style was most effective. Again you can't compare player A to player B even if they are doing the same rotation, in the same gear, with the same professions. THAT EXPERIMENT IS FLAWED. But you CAN compare generally trends BETWEEN players.

If you can get over how big of an epeen you think I have, and actually look what I contributed to the conversation, you may be able to go try the same play styles I did, and report back if you found the same TRENDS I found.

A quote from you: PTR 5.2 has siginifcantly shifted a lot of core mage play-styles, we need to experiment and see which playstyles are still viable, and which are doing well, and which do not work

There were only two play styles that the PTR changed. Mastery scorch arcane, and invocation haste arcane. Fire is still the same, frost (minus using pet for fof) is still the same as far as rotations go.

I proposed 3 different ways to play arcane mage, tested them, reported my findings, then editted in exact numbers not because I wanted to say I have the biggest dick in all the lands, but to make my findings less arbitrary.

Maybe if you used some of your braincells, you could test things in the most efficient way possible. I have a int>hast>mastery mage. My results are going to be clearer than someone that has half their items with crit on it. If you found someone else to gear different gearing, they could go through the playstyles and figure out of they found the same trends as me or if they did not.

At the end of the day, isn't this about what rotation is going to be the best? What will be the best way to reforge and gem your gear? You can find the answers to those questions without using your strict criteria.

Heisenberg, BS in Biochemistry (probably never had to do experiments in that major eh).

[zomgDPS]

Again you can't compare player A to player B even if they are doing the same rotation, in the same gear, with the same professions

For the purposes of these tests, you can, because, as I have stated around 14 times in this thread, this is not a "precise" study. We are not trying to find out DPS measurements to the nth decimal place.


We are working with BALLPARK figures. Is that not sinking in?

If you do the same rotation as me, in the same gear, your BALLPARK results will be the same as mine. Funnily enough, the entire previous page of results form this very thread prove that point.

People are getting in the 56k-64k BALLPARK with a variety of similar rotations and gear. That is precisely the kind of data we need.

And that is my problem with you. There is a page of results backing up what I am saying on one hand.. and you beating your chest demanding demented attention on the other.

I'm done with you. Move on.

---------- Post added 2013-02-15 at 06:27 PM ----------

Heisenberg, BS in Biochemistry (probably never had to do experiments in that major eh).

Awww.. this part was cute

[Mastamage]

Just using our brain cells we can tell that mastery has no purpose in an invocation arcane build. The reason mastery was so powerful was because players were above 85% mana on live which is not possible on the PTR. Hence the only option for our primary secondary stat would be haste.

This makes me laugh because I have proven in testing that Mastery - Invocation build does in fact work. GG w/ just posting numbers btw and not give a shit about the reason for this thread.

[Joshythemage]

Build: Frost | Invocation | Hit>Haste>Crit (to cap)>Mastery, same gear as before
Test: 2
Result: 75-78k
Strategy: Patchwerk Style, all out/all CD's (including Time Warp and trinket) used but still only self buffed.

Test 1 with Recount set to "Overall Damage" making absolutely sure my orb wasn't cleaving and Test 2 was with Skada. Still pretty close to what I got before.

[zomgDPS]

Just a heads up dudelettes.

The final two segments of Throne of Thunder will be available for the next 72 hours (starting right now), for open testing, since those sections have seen less testing than the front half of the zone. We've made and are continuing to make adjustments to the first two wings based on the last week's LFR testing, and will open those portions up again with the next PTR build.

Please share any and all feedback about the boss and non-boss encounters in the sticky threads for each wing.

Sauce.

It seems they have opened up LFR for testing again.

If possible, get in there this weekend and test out the specs if you can. Post back here with your results but please be descriptive of what was going on (e.g. Tested boss X, used build Y (same as before with the (spec) | (talent) | (stat weight) thing).

If possible, keep an offspec that is of a different spec (e.g. have an arcane and a frost offspec) so that you can switch between and get some comparisons.

The more specs we can get results for the better. If you can do multiple specs on the same fight, that would be REAL helpful.


I know I'll be in there this weekend (LFR), so don't be shy and say hi! xD

[dementor]

@ Mastamage, any play style works... you don't have to test for that. Which play style is the best, that you have to test for.

But since you guys are unable to comprehend anything other than your terrible experiment, and I like to be helpful to the mage community, I will play your game.

I choose experiment number 3 (patchwerk) for first 4 attempts (then I got sick of waiting for lust) because the other two don't give a good idea of a specs output, plus most fights are in the ballpark of 5 minutes (since we like talking about ballparks). For example if you asked a BM hunter to do test 1 he would look like garbage. If you asked him to do test 2 he would look amazing. If you asked him to do test 3 it would even out.

But before I post my findings, let us do some thinking with our brain cells again, which is apparently hard for some. In rare cases a rotation is found that wasn't intended in the game... in the case of arcane mages that was scorch weaving to keep 6 stacks. Since the design goal of arcane mages was mana management/burst through controlling arcane stacks, this obviously wasn't something that blizzard would cater to.

Arcane Mage Mastery- Increases all spell damage done by up to x% based on the amount of mana the mage has unspent.

Scorch is the reason that mastery stacking worked for arcane mages with rune of power. Scorching while in the rune could offset the mana cost of arcane blast keeping the mage at a high (preferably higher than 85%) percentage mana while maintaining 6 arcane charge stacks.

There was another equally valid way to play arcane in 5.1 which is haste/invocation (see Modoshi, Shake Miracle, US-Illidan) where the play style revolved around keeping up NT, using AM procs, but otherwise AB until oom where you would evocate and continue on 6 stack camping.

Using my brain cells I can see that using mastery, which only benefits from staying high on mana, wouldn't work under the invocation model since the invocation model uses more than the first 15% of your mana bar. But 2 of my tests below will prove that this strategy wont work on live and it will continue to not work in 5.2 while using premade gear, with premade professions and no racials.

Now, blizzard made a change in 5.2 so that it is impossible to camp full arcane charges. They changed the mastery/rop combo but taking away scorch and lowing the mana return of rop. They changed the haste/invocation model so that evocating removes your arcane charge stacks. But the basics remain the same.

Mastery favors rop where you keep your mana high by dropping your stacks often.

Haste favors invocation where you burn you mana and quickly get it back with evocation.

Apparently these concepts go along with neurosurgery.

Enough talk lets get to what I found.

1
Build: Arcane, Frost Armor, Invocation, Int>2haste>crit>mastery
Test: 1 (absolutely no cooldowns)
Result: 64.3k dps over 6 minutes
Strategy: Keep up nether tempest, never use AM procs unless fully stacked, AB until oom, evocate

2
Build: Arcane, Mage Armor, RoP, Int>2mastery>haste>crit
Test: 1 (absolutely no cooldowns)
Result: 68.5k dps over 6 minutes
Strategy: Keep up nether tempest, stay above 70% mana, use AM procs at highest arcane charge (usually 2 or 3), AB to dump extra mana, Abar at 2-4 stacks depending on mana.

3
Build: Arcane, Frost Armor, Invocation, Int>2haste>crit>mastery
Test: 3
Result: 82.1k dps
Strategy: same as first test

4
Build: Frost, Frost Armor, Invocation, Int>2haste>crit>mastery
Test: 3
Result: 80.6k dps
Strategy: Same as frost on live. Used frost bomb

5
Build: Frost, Frost Armor, Invocation, Int>2haste>crit>mastery
Test: 3
Result: 76.3k dps
Strategy: Same as on live except I used nether tempest. Even having BF procs as #1 priority I missed 5-6 due to back to back procs.


Then to tickle myself, I decided to go against what I thought was right and do a haste/rop trial
6
Build:Arcane, Frost Armor, RoP, Int>2haste>crit>mastery
Test: 3
Result: 76k dps
Strategy: Burn during cooldowns, use a mana positive rotation so that I could burn it all again during the next set of cooldowns.


@zomgDPS, hope your happy now.

---------- Post added 2013-02-16 at 05:03 AM ----------

Test #2, Mage Armor. It may work just fine, but it wont be the best and the best seems to be the point of the thread. The only part I criticize is that everyone has to be in the same gear. zomgDPS keeps talking about ballparks. You don't think you could get a ballpark understanding of which combination of everything works from someone that character copied instead of using a premade template. Then go on to call everyone that would rather not set up a whole new toon to run some tests and report their findings to the mage community a bunch of retards that don't belong in a thread talking about what will be the best for 5.2?

Do the trends from my character copy closely resemble my findings from my premade toon... not exactly, but pretty close.

---------- Post added 2013-02-16 at 05:08 AM ----------

Instead of ridiculing me, try it out for yourself and see if you get similar findings. Ive done 10+ tests now, most people have posted 1 or 2. At the end of the day, 5.2 is going to come, and I will continue to be a ranking mage. You're right I don't need to be posting in these threads, but I like theory crafting and helping other mages. But if that is unwelcome, which it seems it has been since my first post on this thread, I'll gladly let you guys hash it out on your own and post on your "help my dps" threads 5.2.

[Polarthief]

I was using the the dummies in SoTM, so when I would use Inferno Blast, it would spread to the dummies around it. The highest peak came from when I had Jade Spirit and Synapse Springs rolling during AT which netted me 4 Crit Pyroblast for ~250k each which then resulted in a nice Combustion due to the 2 piece from T15, which was then spread again via Inferno Blast. I don't think Horde have any isolated dummies, maybe in TB, so I'll look around again and do a solo parse, but yeah, the peaks are from incidental cleaving.

Ahh okay. Alliance has one at Stormwind, which is what I did to conduct my test.

Read more noob!! xD

You can buy a full set of epics that are better than anything you have on live right now, including the new tier gear, on the PTR.

I didn't know 493 was a higher number than 500-517... (and I already said I picked up the tier gear) Not to mention, if the Tyrannical gear WAS better than my gear, then why did I pull an additional 20-30k over others?

But since you decided to be so rude, next time, I'll just keep the data to myself. I figured more data is always better, seeing as how when we raid, we won't be using PvP items.

PS: Reading's overrated.

[Estride]

*snip*

I've came up with similar results on the disparity between frost and arcane in your first post. Would you know what could have made frost drop below arcane in your second round of tests?

[dementor]

If you mean the drop from 110k to 81k, its the garbage gear you must wear in order to post in this thread.

But your more likely talking about Test 3 and Test 4. The answer to that is unknown. If I remember correctly I was getting lucky on AM procs, which is a huge boost in dps. There is a reason simcraft runs thousands of iterations in order to come up with an averaged dps. Running 1 5 minute dps test doesn't tell you too much, it has to be averaged out over many times. From what I've read in other forums arcane and frost end up being about the same with fire coming back to the forefront. You can read about that in the other 5.2 forum.

Also it comes down to itemization. The gear I have from character copy go well with frost since I have a frost set from doing challenge modes. It's hard to tell which spec or rotation comes out ahead when you have a hobo set of gear that isn't itemized toward any one spec. I understand reforge and regemming, but the static stats you have to work with make a big difference as well.

[Polarthief]

If you mean the drop from 110k to 81k, its the garbage gear you must wear in order to post in this thread.

You could always be like me and say "fuck that". Just post if you use your own character by posting stats.

Also, everyone should be pointing out if there's another dummy nearby for cleave damage because all 3 speccs have some. (Frost has Ice Lance glyph, Fire has DoTs, and Arcane has ABarr. All 3 also have the Mage Bomb cleaves) If unintended, I recommend doing a test to see straight up single target dps (Stormwind for Alliance, Silvermoon[?] for Horde)

[Shangalar]

It is my desire that nobody ever in this thread questions the methodology of the experiment we are doing.

As zomgDPS said himself and as *everyone* agrees, it is far from perfect and serves a different purpose than getting numbers. If someone else has a better or more effective test in mind, please send a private message to zomgDPS and talk it out or post it in the main 5.2 thread and we'll see if we can use it. Until then, the only thing I want to see here are test results based on the outlined methodology and a discussion about them.

Thanks dementor for deciding to join in and not derailing the thread anymore.

[Polarthief]

It is my desire that nobody ever in this thread questions the methodology of the experiment we are doing.

As zomgDPS said himself and as *everyone* agrees, it is far from perfect and serves a different purpose than getting numbers. If someone else has a better or more effective test in mind, please send a private message to zomgDPS and talk it out or post it in the main 5.2 thread and we'll see if we can use it. Until then, the only thing I want to see here are test results based on the outlined methodology and a discussion about them.

Thanks dementor for deciding to join in and not derailing the thread anymore.

*Shrug* I guess I don't see the point in using PvP gear and blue trinkets to test our raid dps, but that's just me.

I'll dip out then I guess because I get discouraged seeing constant results under 70k, which may or may not be cleave damage because I'm willing to bet people are using the Shrine dummies.

[Shangalar]

I wouldn't be surprised if this is the case. I asked a mage in my guild to participate in this experiment and his first reaction was "if it helps killing arcane, sure".

The lucky thing is, as we have mentioned a billion times, we are only testing rotations and not numbers. The gear is because if you have an item level of 525, Crit + Incanter's Ward will do more damage than a 495 Haste/Invocation build and it hurts what we are trying to accomplish.

[Polarthief]

The lucky thing is, as we have mentioned a billion times, we are only testing rotations and not numbers.

So in other words, any tests besides Arcane is worthless because Fire and Frost haven't changed in rotations in the slightest. Gotcha.

*Totally ducking out of pointless-to-me-thread then*

[Shangalar]

You are welcome to use the same gear to test Fire and Frost to see how they compare to Arcane in similar gear and similar conditions (test propositions). The original intent is then to see if any of our specs need to be buffed before the patch.

But yeah, if you want to for example test if Fire works well with 50% Crit at item level 530, then it is quite pointless to participate in this thread.

It is also possible to test your dps output during raid testing (adjusted item level), but then skill plays too much of a factor for definitive results when you compare yourself to other players, along with luck.

[Polarthief]

You are welcome to use the same gear to test Fire and Frost to see how they compare to Arcane in similar gear and similar conditions (test propositions). The original intent is then to see if any of our specs need to be buffed before the patch.

But yeah, if you want to for example test if Fire works well with 50% Crit at item level 530, then it is quite pointless to participate in this thread.

It is also possible to test your dps output during raid testing (adjusted item level), but then skill plays too much of a factor for definitive results when you compare yourself to other players, along with luck.

The luck part being if you have any 5% Haste buffers in the raid *chuckle*

[dementor]

I wouldn't be surprised if this is the case. I asked a mage in my guild to participate in this experiment and his first reaction was "if it helps killing arcane, sure".

The lucky thing is, as we have mentioned a billion times, we are only testing rotations and not numbers. The gear is because if you have an item level of 525, Crit + Incanter's Ward will do more damage than a 495 Haste/Invocation build and it hurts what we are trying to accomplish.

The point I was making, which you just repeated, is that numbers don't matter just the general trends. If you look at the trends on which playstyles were more favorable for my character copy toon, then you compare to what play styles worked for my premade character, they were rather similar. And I bet they would be for anyones character copy toon vs premade.

zomgDPS was the only person derailing the thread by flinging unnecessary insults, which I am not sure why he was not infracted. Ill take my knowledge and insight of the mage class to a different thread as well.

[Deleted]

GC:"number tweaking hasn't started yet"