Jump to content


Photo

Arcane in Cataclysm: Updated for 4.01


  • This topic is locked This topic is locked
265 replies to this topic

#41 Logix

Logix

    Piston Honda

  • Members
  • 136 posts

Posted 14 October 2010 - 08:39 PM

Ghostcrawler On Patch 4.0.1 Class Hotfixes - World of Warcraft News - Curse

Initial adjustment/nerfbat applied. Mage and frost touched, but not as much as Fire. I imagine AB was the primary issue - interesting to see how this affects the burn duration/crit numbers.


I have just completed my investigation.

It seems both the base damage and coefficients for all 3 arcane spells was reduced by approximately 8%.

The new arcane coefficients are:

Arcane Blast: 0.9723
Arcane Barrage: 0.7037
Arcane Missiles (talented, over 5): 1.0765

I will be making these changes and will hopefully have an updated version of MMS (including a level 80 toggle) by the end of the day. Stay tuned.

That being said, from the math I have done so far, these changes have no discernible impact on the optimal strategy for Arcane. It's basically just a global damage reduction.

#42 Wiztrix

Wiztrix

    Glass Joe

  • Members
  • 1 posts

Posted 14 October 2010 - 08:52 PM

The return of Polymorph -- It's rumored that the new pulls will be challenging once again, such that CC will be needed (when's the last time we sheeped anything other than a Cow or a WG player?)

Given that, a consideration may be to poly-glyph.

Just a thought.

Trix

#43 Daethar

Daethar

    Glass Joe

  • Members
  • 1 posts

Posted 14 October 2010 - 09:09 PM

Are we sure on that formula for Mana Adept? This seems to be that as long as we have more than 0 mana, and more than 0 mastery, we will always be doing over 100% damage (100% being before training mastery.)

#44 Logix

Logix

    Piston Honda

  • Members
  • 136 posts

Posted 14 October 2010 - 09:20 PM

Are we sure on that formula for Mana Adept? This seems to be that as long as we have more than 0 mana, and more than 0 mastery, we will always be doing over 100% damage (100% being before training mastery.)

Correct.

You will always be benefiting from Mana Adept (barring a strange situation where your mana is artificially being kept at 0%) even though the benefit will be immensely small near 0% mana. Your mastery will never fall to 0, since you will always have 8 base mastery points even with no gear.

The only way to prevent gaining some theoretical benefit from Mana Adept, is to not train the Mastery skill.

#45 binksy

binksy

    Glass Joe

  • Members
  • 3 posts

Posted 14 October 2010 - 10:36 PM

With arcane now needing 17% hit and having burn phases every 2 minutes I've been actually considering as a viable trinket. I would reforge the hit to mastery, and would have a very nice 716 SP addition to each burn phase. I currently use 277 and , but also have access to the Talisman and 245 Reign.

I don't actually need the hit from the trinket, but if I had it I would just reforge hit off some other pieces. I'm mostly attracted by the 716 sp for each burn phase. I also know people were considering which doesn't give as much sp for burn phases, but has a nice int boost instead. Do you think the int would outweigh the extra sp and crit/haste we could reforge to from gaining hit on the quill?

I was also laboring under the old views of haste > crit for arcane but it looks like its no longer the case. I reforged all my gear to mastery and for most pieces it was the crit I reforged. I switched to haste when I was at 33% crit unbuffed, but do you guys think we should just reforge haste across the board and stack crit as well as int/crit gems in yellow? Also the idea of using molten armor for burn phases is very interesting, especially since we can glyph it as well as mage armor with the current bug.

#46 Superdad

Superdad

    Glass Joe

  • Members
  • 1 posts

Posted 15 October 2010 - 12:39 AM

I'm also curious about whether it's better to just gem int accross the board and forget about set bonuses. It would synnergize with improved mana gem, mage armor, and longer burn phase.

Similarly, do you think Ember Skyflare Diamond meta should replace CSD? (ember is 21 int and 2% max mana)

#47 Logix

Logix

    Piston Honda

  • Members
  • 136 posts

Posted 15 October 2010 - 04:54 AM

I'm also curious about whether it's better to just gem int accross the board and forget about set bonuses. It would synnergize with improved mana gem, mage armor, and longer burn phase.

From everything I can tell so far, Int is by far the best choice to gem for, assuming you are hit capped.

That being said, I have a new version of MMS ready to go which will allow everyone to get a much better feel for what the stat equivalence is with your specific gear inputs (as well as full support for level 80 mages). I was about 5 minutes away from pushing the update live when I noticed a new beta build was being deployed.

Upon doing a quick test it seems like this new build does not contain the spell coefficient changes that are on the live servers right now. The new build also came with some new premades in what seems to be entry level raid gear.

Due to this, I am holding off bringing the new version of MMS online until we can gain some clarity into what the new coefficients actually are. This will also give me some time to do a full round of bug tests to see if any of the major Arcane bugs still exist (AM range bug, MoE bug, etc) as well as update the default profiles for the level 85 MMS arcane mage with the stats from the entry level raid gear, which I think will give us a much better indication of what state arcane will be in when the Cataclysm raiding season begins.

When MMS does get its update, hopefully we will be able to get a much clearer picture of what our stat equivalences will be (as well as how much of an impact MoE has to arcane).

edit: Corrections.

#48 Pook

Pook

    Von Kaiser

  • Members
  • 28 posts

Posted 15 October 2010 - 05:04 AM

The addon Mage Nuggets has a live display of the percentage damage boost from Mana Adept, I've found it quite useful to actually be able to see what effect reduced mana has.

#49 Kenfussion

Kenfussion

    Glass Joe

  • Members
  • 1 posts

Posted 15 October 2010 - 07:56 PM

Is there a cap on mastery? Or at 100% mana you get 12% additional spell damage due to the Mana Adept Mastery and then and additional 1.5% for each point in mastery (at 100% mana) through gear plus the default 8 mastery points? Or is the 8 points we get by default making the 12% spell damage and each additional point above the 8 will add and additional 1.5%? (Meaning at 10 points we would have 15% additional spell damage).

#50 Kariska

Kariska

    Glass Joe

  • Members
  • 3 posts

Posted 15 October 2010 - 08:34 PM

Is there a cap on mastery? Or at 100% mana you get 12% additional spell damage due to the Mana Adept Mastery and then and additional 1.5% for each point in mastery (at 100% mana) through gear plus the default 8 mastery points? Or is the 8 points we get by default making the 12% spell damage and each additional point above the 8 will add and additional 1.5%? (Meaning at 10 points we would have 15% additional spell damage).


What you said last.

I made a simple spread shit to calc it easily:
Google Docs Mage Mastery Calc

#51 Narub

Narub

    Glass Joe

  • Members
  • 15 posts

Posted 15 October 2010 - 09:19 PM

Hell there,

So for this week I tried fire mostly with the exception of Blood Queen.

My question now is aimed towards the mages that did a proper full run of ICC and tried arcane the whole way what they did with their gemming. I would assume you ofcourse prioritise getting hit capped, after that get some extra haste. But after this I'm kinda stuck to change my crit to mastery or just leave it be.

Has any mage tested a full mastery set compared to crit for example with good positive feedback?

I hope this fits in this topic, if it didn't I'm sorry.

#52 Sandarj

Sandarj

    Glass Joe

  • Members
  • 6 posts

Posted 15 October 2010 - 10:41 PM

Some input/evidence on the Armor Dilemma (mage vs molten in burn phase): during our run Wednesday I tried switching to Molten for Festergut's burn phase and keeping Mage for Rotface. I didn't really feel like the GCD burnt to throw Molten on justified the one or two extra crits I got out (esp since they were towards the end of the cycle when I evocated). Overall throughput was better using GCDs on standard CDs, with about 300 more DPS recorded.

#53 sadeiko

sadeiko

    Glass Joe

  • Members
  • 4 posts

Posted 16 October 2010 - 12:23 AM

On trinkets I was wondering and have done a few tests on dummies resulting in a a bit more than 1k damage boost over by using edit: this was tested from only doing a ab4x pop'n'burn once, then waiting for cds and testing again. dps from Talisman was around 10300-10500 and dps with Ice Crystal being 11500-11800

Although better options seem to be out there from icc and on, however for the future of arcane mages, might these "front loaded" trinkets proved more power than back loaded ones(that ramp up and deal most damage towards the end of their cycle such as ToR? This takes full advantage of mastery of course, as well as helps to make sure the burn phase is as powerful as possible.

#54 Logix

Logix

    Piston Honda

  • Members
  • 136 posts

Posted 16 October 2010 - 12:38 AM

A new beta build will be going up shortly.

Arcane
* Arcane Specialization now increases damage of Arcane spells by 15%, down from 25%.

Fire
* Fire Specialization now increases damage of Fire spells by 10%, down from 25%.

Frost
* Frost Specialization now increases damage of all Frost spells by 15%.

It is, as of yet, unconfirmed if these changes are in addition to, or replace, the base damage and coefficient reductions that were pushed live and to the beta servers (albeit for a short time only).

A minor observation from these changes.

The relative power of Arcane's Flame orb vs non-Arcane specs flame orb just went up. Given that the premade, entry level raid Arcane mages can somewhat easily attain 25% damage bonus from mana adept, This means that Arcane's flame orb will be doing more basic damage then Fire's above ~52% mana and more damage then frost's above ~60% mana.
At 52% mana, the damage bonus from Mana Adept will be 13.03%. Fire's flame orb can do a maximum of 13% extra basic damage (10% from the specialization, 3% from the fire power talent). Naturally, this conclusion does not calibrate for the additional benefits both Frost and Fire glean from Flame orb, i.e. the explosion for fire and the increased FoF proc chance for frost.

Either way, given the rather absurdly good coefficient and extremely high DPCT of flame orb, it may end up being a staple part of Arcane's 'clicky macro of doom' when initiating a burn cycle.

#55 Forell

Forell

    Glass Joe

  • Members
  • 4 posts

Posted 16 October 2010 - 01:31 AM

On trinkets I was wondering and have done a few tests on dummies resulting in a a bit more than 1k damage boost over by using edit: this was tested from only doing a ab4x pop'n'burn once, then waiting for cds and testing again. dps from Talisman was around 10300-10500 and dps with Ice Crystal being 11500-11800

Although better options seem to be out there from icc and on, however for the future of arcane mages, might these "front loaded" trinkets proved more power than back loaded ones(that ramp up and deal most damage towards the end of their cycle such as ToR? This takes full advantage of mastery of course, as well as helps to make sure the burn phase is as powerful as possible.


Anecdotally it feels like Crit is more helpful for the burn phase than SP. I want to take advantage of the 200% crits more than I want my base to be a bit higher. Obviously it needs to be theorycrafted, but it seems like that's the way Arcane is heading post 4.0.1.

Similarly with haste. If 2ABAM/ABr is the rotation for the conserve phase in Cata due to proportionally smaller mana pools (I am led to believe ours are artificially high at current, hence why we can go for AB3AM/ABr without much issue), then aiming for 2 second casts on AB could well be the ideal as it means a smooth rotation through AB2ABr consistently if AM doesn't proc. On live with 1.9 second casts (and 200-300ms latency) there is difficulty chaining that rotation effectively, which results is a stutter in the casting, which ends up to be very frustrating.

#56 Obeast

Obeast

    Glass Joe

  • Members
  • 16 posts

Posted 16 October 2010 - 07:51 PM

I may have missed it during while skimming the top post (thank you for your work!), but it seems to me that the mathematical result that, sans cooldowns, arcane mages should use their cycles in order of increasing MPS/DPS deserves a mention somewhere in the "other optimizations" section.

Even if it turns out to be optimal to use your high-DPS cycles near 100% to stack the mana gem bonus and other cooldowns with your peak mastery bonus, which seems quite plausible to me, once those bonuses wear off it will be optimal to switch back to your lower MPS/DPS cycles even if you still have more of your "burn phase" left.

I.e., you'd go (again, taking as given that it's optimal to burn near 100%) Gem to 100% -> other cooldowns -> Burn cycle until cooldowns + gem wear off -> lower DPS/MPS cycle(s), in ascending order if more than one -> Burn until evo comes up. Seems like something worth noting.

#57 Magdalaine

Magdalaine

    Glass Joe

  • Members
  • 2 posts

Posted 16 October 2010 - 07:55 PM

All I can say is that in a 25 man icc, with hero/raid buffs/my own cooldowns popped, I was critting at 80k+. I have the Reign of the Unliving trinket and dfo. Later in the raid, I became mind controlled at the Blood Queen and killed the entire 25 man raid. Thus I have earned the nickname Magzilla.

The info here is excellent. Thanks guys for all you put into it.

#58 lrdcollins

lrdcollins

    Glass Joe

  • Members
  • 4 posts

Posted 16 October 2010 - 08:37 PM

All I can say is that in a 25 man icc, with hero/raid buffs/my own cooldowns popped, I was critting at 80k+. I have the Reign of the Unliving trinket and dfo. Later in the raid, I became mind controlled at the Blood Queen and killed the entire 25 man raid. Thus I have earned the nickname Magzilla.

The info here is excellent. Thanks guys for all you put into it.


I had one crit for 99k, lol. Anyway It was discussed somewhere earlier about gemming. Has anyone done any math yet about whether we should just be gemming straight intellect or for set bonuses. The Int Haste gems for sure I think can go, and as far as spirit goes, is it worth it now with no more molten armor. Also I know its been discussed of and on, but any concrete haste and crit numbers we should be looking for. Based on what I've seen in raids to this point Crit>haste even for arcane.

Side note: I only really started using this site 5 months ago, and so far have been a leech more than a contributor. I would be very interested in getting involved in the theory crafting aspect and testing the numbers, but I would have no clue where to begin. If help is needed I am more than willing but would need some guidance to start.

#59 Logix

Logix

    Piston Honda

  • Members
  • 136 posts

Posted 17 October 2010 - 01:22 AM

I may have missed it during while skimming the top post (thank you for your work!), but it seems to me that the mathematical result that, sans cooldowns, arcane mages should use their cycles in order of increasing MPS/DPS deserves a mention somewhere in the "other optimizations" section.

Indeed. It was less a case that I left it out and more a case that I was trying to find out exactly where and when I should place this information. I will be sure to add it (with proper links to the math) in my next iteration of the original post (which should be very soon).


Speaking of iterations. I completed my round of testing on both live and beta. It seems that the reduction of the damage bonus of the specializations bonuses are in fact not in addition to the base damage and coefficient reductions from a few nights ago. Currently on beta, the spell base and coefficients have been returned to their previous values and instead replaced with the overall specialization bonus reduction.

On live, however, the specialization bonus reduction is not present, hence, the base damage and coefficient reductions are still in place.

All of this, and more, is now fully supported in the latest version of MMS, which I have just pushed live.



A new version of Mana Made Simple is now online!
(http://manamadesimple.elementfx.com/)

The new version brings some much needed improvements and updates.

Full support for level 80! Use the drop-down in the character setup to switch MMS between WoTLk and Cataclysm modes.Added Master of Elements talent. Now you can toggle mana returns from MoE on and off, to see what kind of impact it has and whether it is worthwhile for your strategy.Added better outputs from the Cycle builder. You can now accurately see which particular cycle is being used when showing sim casting outputs.Crit, Haste and Mastery inputs now changed to combat ratings instead of actual percentages. This will really help when trying to figure out which stat to reforge into what. A little ratings converter tool has been added to help you figure out how much rating you have.Tons of tweaks to the cycle builder and core engine. Hopefully everything will remain stable as the sim increases in complexity.All damage calculations are up to date with the proper coefficients as they exist on live (4.01) and the most current version of the beta (build 13195).
A few helpful notes on using MMS.
I seem to be getting quite a few questions concerning the Cycle Builder tool. Most of them stem from people forgetting about the priority system in place. A quick reminder. The Cycle Builder tool allows MMS to use multiple cycles during a particular strategy. It will use these cycles according to the priority that they are inputted in. To ensure a particular cycle does not get overridden by another, increase its priority.


That being said, it would not be a complete update if MMS does not help us answer some of the outstanding questions about Arcane right now. This particular update attempts to answer two of our questions so far. These are:
Is Master of Elements even useful? If so, by how much?What are our combat ratings priorities? Which is better, haste, crit or mastery
The following is the first round of analysis into answering these questions, hopefully, the results will be useful to someone.


How useful is Master of Elements to Arcane?
As you may know by now, MoE returns mana on ABr crits and on AB crits. However, on AB crits, it returns mana based off of the mana cost of an unstacked AB irresspective of which stack of AB critted. It is, as of yet, unconfirmed if this is a bug or by design. However, since this has been MoE's behavior for a while now, we are proceeding with the investigation assuming that it is not a bug.

Due to this, the following tests were setup in MMS.

Test 1.0: Testing MoE's usefulness during a "Conserve" phase.
Setup: To isolate MoE's contribution, a lot of the mana based RNG was disabled. AM procs, Clearcasting procs were supressed. The test was run for 2000 iterations over a 110 second casting window, casting AB2 ABrs, in order to formally simulate an adequate casting rotation. Gear was of the new entry level raid mages (default settings in MMS).

Results:
With MoE activated:
Spoiler

Average DPS: 6,680.2

With MoE disabled
Spoiler

Average DPS: 6,628.5


Analysis:
Using default stats, we see that MoE contributed to approximately 51.7 extra DPS at the test gear level. Which is an increase of ~0.77% DPS increase. This is not very much at all. In order to see if this increase scales. I ran a second test using a much higher crit rate.

Test 1.1: Testing MoE's usefulness during a "Conserve" phase with much higher crit rate.
Setup: Exactly the same as test 1, but crit rate bumped up to ~30%.
Results:
With MoE
Spoiler

Average DPS: 7,135.1

Without MoE
Spoiler

Average DPS: 7,061.4


Analysis:
Bumping up the crit rate provided ~73.7 extra DPS, which is ~1% extra damage.

So what we can see is that as crit increases, the value of MoE obviously increases but it still remains around the 1% increase in DPS mark.

MoE is a 2 point talent. Is a 1% increase in DPS during a "Conserve" phase enough to warrant a 2 point talent investment? Lets first see how well MoE does during a burn phase.

Test 2: Testing MoE's contribution to a "Burn" phase.
Setup: Reset the character. The test is a 15 second burn of AB spam. All other settings are identical to test 1.1
Results:
With MoE
Spoiler

Average DPS: 9527.4

Without MoE
Spoiler

Average DPS: 9519.0


Analysis:
MoE contributed to an extra 8.4 DPS, which is almost negligible (less then even 0.1% DPS gain).

This result should not really come as a surprise, since we are not casting ABrs during a burn phase, hence, MoE returns are very poor.
We can conclude that MoE is not very useful at all during Burn phases.


Final conclusion for this preliminary analysis of MoE
So, 2 talent points for an ~1% DPS increase during your "conserve" phase? Under other circumstances, I would totally ignore MoE, however, we realize that there really isn't anything else we could take instead that is a pure DPS boost. What I do know now is that I may end up taking Burning soul over MoE now. But either way, I think I am at least fairly certain that MoE is in no way a game-changer for Arcane and can easily be skipped if a fight calls for points elsewhere.

What is a bit worrying is that even for a tier 1 talent, MoE is very weak for arcane in general. Frost and Fire both get some pretty solid tier 1 choices in Arcane (3% haste and Clearcasting, both of which easily deliver more than 1% dps). If nothing else, perhaps we now have some evidence to try and get MoE buffed? Or at least fixed so that it returns mana for stacked ABs properly?

Let us see what we can do.



Up next. Using MMS for some preliminary investigations into stat equivalences. Stay tuned!

#60 Kavan

Kavan

    Bald Bull

  • Members
  • 1018 posts

Posted 17 October 2010 - 02:43 AM

Can you clarify something for me in your test? It appears all tests run for a fixed amount of time, but you're only varying mana regen from MoE. You're saying that you only used a fixed rotation for the whole duration. This implies to me that the only difference you're calculating is from mana adept bonus dropping as mana drops faster. In a real situation however you would use that extra mana to switch the breakpoint between cycles, so at least in theory you could utilize more out of the talent than what your test is suggesting.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users