SpartAdept, have you by any chance built a 359 version of your Fire set?

Silveriwind, the second line here makes the first superfluous:


actions+=/use_item,name=shard_of_woe,if=cooldown.evocation.remains<26
actions+=/use_item,name=shard_of_woe,if=cooldown.evocation.remains<86


By any chance did you mean >86?

Yes, this sould be >86. I wanted to make sure that the trinket during the conserve phase is only used when it will be ready for the next burn phase.

Shouldn't it be >70 in this case? Since duration is 10s and cooldown is 60s.

Liv, I followed the link, and looked at your sample cast sequences and the DPS chart. The simulator doesn't seem to show DPS dropping to zero for the six seconds while a mage is evocating. But for sure, it should be zero during that time. Because your cast sequence shows that prior to this, no flame orb was cast. Generally, it was AB spam to mana = 26%.

I am not sure if it was actually zero, the the DPS chart smoothed it out, or there is an error somewhere in there. But in a long fight, where your sequence has two evocates, then there should be two such 6 seconds of zero DPS.

Looked at the sample sequence a bit more. I don't understand this part. It say that M stands for arcane_blast,if=buff.arcane_blast.stack<4&!buff.bloodlust.react. I assume this means when blood lust is active? But M appears throughout the entire sequence of the over 420 second fight, but it seems to me you can only bloodlust or time warp once during a boss fight, since its shared cooldown, and the cooldown is 10 minutes.

I am also not sure about flame orb. If you are simulating a tank and spank fight, with zero movement. Why would the spell sequence include flame orb? You are spending a 1.5 second cooldown to cast it. That's 1.5 seconds you could be using to cast AB instead. If we really can hit 25k DPS in top gear, then that's 37.5 k damage. Does flame orb even do 37.5 seconds damage in total over its entire life? It hits for little over 1000 every time, and lasts only 15 seconds.

In a true tank and spank fight, an arcane mage won't bother with flame orb, not if he is averaging 25k DPS.

I made some basic calculations. Each Arcane blast stack increases damage by just 10% only. Even a max 4 stack, will increase damage from your base by just 46% if you compound it by 10% each stack to 4 stacks.

Mastery can be stacked to infinite amounts. I am in non raid gear and my mastery is already increases my base damage by 28%. If we are talking about an arcane mage who is raid buffed, plus max top line gear, are we looking at 40% at least?

And an arcane mage doing a AB rotation to keep at near 100% mana is not having zero AB stacks. In practise, say if he goes up to 4 stack before he clears it with AM, then he averages a 2 stack. We do need to through in the weaker damage of AM into the mix, and the slower cast of course.

But at the same time, the decrease in cast time is just 0.1 seconds each stack. So, even at max 4 stack, Arcane blast goes from a base cast of 2.35 seconds to 1.95 seconds. An average 2 stack for a mage running an AB rotation gives 2.15 seconds. So, the difference in the two is just a 10%% faster cast rate.

Even translated into raw damage, thats just 10% more damage. But if you get yourself to low or zero mana trying to spam AB, then you are losing as much as 40% on ALL damage.

Don't forget another thing. an AB crit with max mana pool does a ton of a lot more damage than an AB crit with close to zero mana pool. At max mana, my mage hits for 14.5k basic arcane blast, crits can easily hit over 40k even on a training dummy.

But when I spam AB to low mana, my hits are 11k, and my crits are a little over 20k. So, one AB crit, at max mana, even unstacked, is twice the damage of a 4 stack AB crit, when you are at zero or low mana.

And top gear will bring higher levels of crit, adn higher levels of mastery. I think the simulator doesn't model mastery enough. I also think our raiding arcane mages are too willing to embrace the concept of "AB spam is the highest damage" which is a throw back from the wrath days.

I believe that at high levels of gear, mastery becomes so important that for long fights, staying at 100% mana is far more important than being able to spam AB for a quick burn, only to spend another 6 seconds doing zero DPS to evocate back again.

I used to think that AB spam as often as possible was the way to go. But as I played more and more in the cataclysm environment, things just didnt apepar to be the case. My max damage is when I am at full mana. Burning to low mana never ever felt like I was doing mega damage.

The sweet spot actually seems to be the first time you use a gem, and you are at a 4 AB stack, and you have Arcane power going, and you crit. That is when you see those eye poppingly large AB crits. This could be very early in the fight. But it doesnt last. At most after that, maybe 2 casts at this sweet spot, then you have to clear away the AB stacks with an AM. Otherwise, you can't stay above 90% mana because each cast is eating near 10k mana. And as your mana pulmets, your damage drops accordingly.

Once you have burned off half of your mana, and you are still spamming AB, it doesnt matter anymore, your damage is lower, even with time warp going.

Liv, can take a much closer look at mastery?

Also, try a rotation spell sequence in the simulator, that includes gems, volcanic potions instead of mana potions, but with a rotation that never runs out of mana and aims to stay at near 100% mana all the time. I think you might be surprised at the result, especially if we are talking about top geared arcane setup.

Consider volcanic potions vs mythical mana potion. Volcanic potion adds 1200 spell damage for 25 seconds. This is further inflated with mastery, haste, crit, and other raid buffs. Mythical mana potion equals 10000 mana. At a ABx4 stack, that's equal to one single arcane blast. I believe Volcanic potions adds more damage in total over 25 seconds compared to the one single extra max buff arcane blast you get out of 1 mythical mana potion.

Sorry for the super long post.

Vontre, I was playing with your magegraf, great tool! May I ask a question. Under spell detail for arcane blast, the final DPS you show for arcane blast is based on a base arcane blast, with all buffs, or an arcane blast which is benefitting from ABx4 stack? Or some aggregated average over an entire fight?

BTW, with my gear (which is far far from the best), just with hit rating changed to get hit cap, a volcanic potion gives more dps than a mythical mana potion on Vontre's simulator.

It's just a base Arcane Blast with no stacks. All the spells in the detail list are calculated with the buffs provided and aggregate trinket effects - you can see a complete list of modifiers by clicking on a spell.

Alvira, I can answer some of your questions.

The DPS during evocation is not zero on the chart because it is an average over thousands of iterations. Since evocation doesn't happen exactly at the same time on average you don't hit 0 exactly. Try a simulation with 1 iteration and you'll see it will reach 0.

About flame orb. This spell has a very high DPET. For arcane, it deals on average 42k damage and this makes it the hardest hitting spell. So using it on cooldown is very important.

Alvira, feel free to experiment with the simulator; if you come up with changes to the profile or action list that produces better results, please post them so the rest of us can confirm your results. If that happens, appropriate changes can be made.

Remember, though, this is not "my" simulator. I provide input on and minor changes to the Mage module where I can, but mostly I work on the HTML output generation. At this point, most of the profiles and action lists are much more the work of other people in this thread besides me.

This line:

arcane_blast,if=buff.arcane_blast.stack<4&!buff.bloodlust.react

...means when bloodlust is not active (the ! checks for not-condition instead of condition).

Another minor "bug" I've noticed:

Improved Mana Gem - Spell - World of Warcraft has an actual duration of 15 seconds instead of the 10 seconds mentioned in the tooltip, both on live and PTR. The tooltip here is just plain wrong here and is still wrong on the latest PTR build. I think it's best to hardcode it for now.

And here's another very minor improvement for Shard of Woe usage:

On another note, I've checked sc_mage and the debug output a bit for any obvious bugs that could explain the dps gap between arcane and the other specs, but so far I've not been able to find any.

I've been following this thread with much interest. I have been working on a theorycrafting tool that is another good resource for this discussion. This is a long post. I'll give the TLDR version first:
1.) I propose there is a slightly more optimal Fire Mage BiS gear set than the one in SpartAdept's post. This is the set: Mr. Robot - Fire Mage - World of Warcraft
2.) I also propose that the stat breakdowns that we are using for Fire Mages are not optimal. Mastery and Haste need to be valued slightly higher, but still the same relative to each other.

I did my analysis on the Fire Mage 372 BiS gear setup as an example. I used SimulationCraft build 403-29 on the Live version of the game for my DPS tests. I know the patch is going to change some things as far as DPS numbers and BiS setups, but, for now I figured testing on Live is still the best for discussion purposes.

SpartAdept quoted some stat breakdowns for fire in his post with the BiS Fire Mage setup:
Cataclysm Mage Simulators and Formulators

The Fire Mage compendium quotes slightly different numbers:
Cataclysm Fire Mage Compendium

If we find the total EP value of the posted BiS Fire Mage set we get:
EP Value of posted set, including enchants
Using the values in SpartAdept's post: 38673.07
Using the values in Fire Mage compendium: 37121.72

If we now plug these same stat weights into the Mr. Robot site, we can find a set of gear that achieves slightly higher EP values for each:
SpartAdept's post weights: 38785.54 (Call this Set A.)
Link: Mr. Robot - Fire Mage - World of Warcraft
Fire Mage Compendium weights: 37245.41 (Call this Set B.)
Link: Mr. Robot - Fire Mage - World of Warcraft

Notice these two sets are slightly different. The boots, bracers, and reforges are different. Both sets are of a higher EP value than the set in SpartAdept's post, though.

Ok, they have a higher EP value, but are they better sets of gear? I ran SimC on the original setup from SpartAdept's post as the base point and found a DPS of 21890. That agrees with Lhivera's output on the OP for a basic fight. I used the default Fire Mage rotation in the version of SimC stated above on a Patchwerk style fight. I then ran SimC on these other two sets. Set A had almost identical DPS, actually simulating out to 40 DPS lower. Set B was even lower at 21788.

To me, this is an indicator that the stat breakdowns are not doing what we want them to do. A higher EP value should always result in higher DPS. So, I compared the posted BiS setup to the more-optimal EP value setups. I was able to make a couple of observations:
1.) Simply changing the trinket from Bell of Enraging Resonance to Theralion's Mirror always results in a DPS gain with SimC. This is expected, since Theralion's Mirror has a higher EP value than Bell in both cases.
2.) The posted set has more haste and mastery, but less intellect/spell power than the "more optimal EP" sets. This might mean that the stat breakdowns are not valuing haste and mastery enough with respect to intellect/spell power.

Sure enough, if I bump up the stat weights on haste and mastery slightly, I am able to use the Mr. Robot site to quickly find an optimal set with the similar gear setup as the set posted by SpartAdept, except for the belt and trinket. The Mr. Robot site finds a slightly more optimal reforging combination also. The resulting set is this (I normalized the stat weights to set intellect to a value of 100, just because I like whole numbers):
Mr. Robot - Fire Mage - World of Warcraft

It simulates out to ~22045 DPS versus ~21890 with the posted set. So, as expected, tweaking the set to obtain a slightly higher EP value also yields a DPS gain. I would propose this as the new "BiS" Fire Mage setup for now.

I wouldn't be surprised if someone clever could find an even slightly better Fire Mage set than this. The Mr. Robot site uses an approximate algorithm to find its solution. More info on the Mr. Robot site is here: Team Robot Gear, Gem, Enchant, and Reforge Optimizer - read "the fine print" for info about what we think the accuracy is.

An additional complexity for fire when trying to use EP values even if you try to do some dynamic adjustments on them is that there are several discontinuities that might not even show in EP values depending on the granularity of differentials. So it's not really that the EP values are wrong, you just can't use them directly to get total dps (at least not in any easy way). If you do try to use EP values a slightly more correct way would be to know the stats that were used to obtain the EP values and then when you evaluate your gear compute the differentials from gear to the stats used for EP values and then compute linear combination on differentials only.