The overall value of the intellect can't be higher than 50 spell power. Being extremely generous with the proc uptime calculations and timer stacking, I can only get it to 180 spell power equivalent, still less than IDS. I suppose the mana could be worth something maybe..?
128 x 1.1 (BoK) x1.15 (Arcane Mind) = 162 int (slightly more if you're a gnome)
141 Int is 24 spellpower from Mind Mastery as well as 0.97% crit from the intellect, which works out to 45 crit rating.
Editing the item in rawr to have 24 Spellpower /45 crit instead of 128 int still has it as very high value for arcane - better than the Illustration of the Dragon Soul and Embrace of the Spider, but slightly worse than Scale of Fates and Flare of the Heavens.
Incidentally, I get the same DPS value for the edited version and the original version with "unlimited mana" selected. I wonder - would it be possible for rawr to display when an item is being highly valued because of the mana (or regen) it provides?
then why does it only say to am when there's 0.4243977/0.865788 sec left on buff?
The + at the end means "or more."
Edit: Oops, I'm apparently wrong about this, see a couple posts down.
Last edited by nathanbp : 08/27/09 at 3:47 PM.
Originally Posted by Crowl
If you have to control a robot dinosaur that fires lazers and there's a time when you shouldn't be shooting those lazers then the encounter is clearly flawed beyond hope of fixing.
141 Int is 21 spellpower from Mind Mastery as well as 0.85% crit from the intellect, which works out to 39 crit rating.
Editing the item in rawr to have 21 Spellpower /39 crit instead of 128 int still has it as very high value for arcane - better than the Illustration of the Dragon Soul and Embrace of the Spider, but slightly worse than Scale of Fates and Flare of the Heavens.
Incidentally, I get the same DPS value for the edited version and the original version with "unlimited mana" selected. I wonder - would it be possible for rawr to display when an item is being highly valued because of the mana (or regen) it provides?
This is why I find it confusing... the "+" just means that the proc is visible. It is entirely unrelated to the number, as far as I can tell.
Again, from earlier on in this thread:
Cycle Code Legend: 0 = AB, 1 = ABar, 2 = AM
State Descriptions: ABx,ABary,MBz+-
x = number of AB stacks
y = remaining cooldown on Arcane Barrage
z = remaining time on Missile Barrage
+ = Missile Barrage proc visible
- = Missile Barrage proc not visible
This is why I find it confusing... the "+" just means that the proc is visible. It is entirely unrelated to the number, as far as I can tell.
Again, from earlier on in this thread:
Cycle Code Legend: 0 = AB, 1 = ABar, 2 = AM
State Descriptions: ABx,ABary,MBz+-
x = number of AB stacks
y = remaining cooldown on Arcane Barrage
z = remaining time on Missile Barrage
+ = Missile Barrage proc visible
- = Missile Barrage proc not visible
That's what I get for making assumptions. I'd still think it means that amount remaining or more and the very low numbers may be because it looks for lowest based on a possible 4T8 proc, even if calculating without 4T8?
Originally Posted by Crowl
If you have to control a robot dinosaur that fires lazers and there's a time when you shouldn't be shooting those lazers then the encounter is clearly flawed beyond hope of fixing.
I have always had a hard time reading what the Arcane Optimal Cycles means. I wish there was a better way to explain them or at least a means of gathering there explaination. Every time I've read what strings of characters means I have tried to figure it out on my own and have always been wrong. Does anyone have a better way of understanding rawr output?
In response to adding Raid Utility to the Arcane spec. What kind of "unbalance" aspects could be seen if you gave Arcane Barrage a haste buff along the lines of the Scorch buff (the 5% crit to spells).
I could see this adding back in the Arcane Barrage into the rotation for raid utiity, while maybe lowering dps a little due to having to cast it, but scorch does pretty much the same thing for those lucky "Scorch Slaves".
Call it Arcanist's Gravitation or something along that line causes all melee and magical attacks to "gravitate" to the target increasing haste by 5%. To offset the fact that it costs more mana to apply this debuff (18% vs 8%), the damage could be left alone and the debuff could last 45 seconds as opposed to 30. A single application instant cast debuff to start a boss fight with would bring some utility to the spec. Not sure how this might affect PvP, but just a thought that might be passed on to developers.
In response to adding Raid Utility to the Arcane spec. What kind of "unbalance" aspects could be seen if you gave Arcane Barrage a haste buff along the lines of the Scorch buff (the 5% crit to spells).
I could see this adding back in the Arcane Barrage into the rotation for raid utiity, while maybe lowering dps a little due to having to cast it, but scorch does pretty much the same thing for those lucky "Scorch Slaves".
Call it Arcanist's Gravitation or something along that line causes all melee and magical attacks to "gravitate" to the target increasing haste by 5%. To offset the fact that it costs more mana to apply this debuff (18% vs 8%), the damage could be left alone and the debuff could last 45 seconds as opposed to 30. A single application instant cast debuff to start a boss fight with would bring some utility to the spec. Not sure how this might affect PvP, but just a thought that might be passed on to developers.
Adding an odd 5% raid-wide haste buff seems a little weird to put on Arcane Barrage.
I read your post wrong the first time, and thought you were talking about giving Arcane Barrage a haste buff to the player for casting it. Similar to Druid's Nature's Grace talent built into the spell "Grants 20% casting speed for the next 3 seconds" or something similar. It could be a very interesting way to take the spec, get Arcane Barrage into the rotation, and add a skill level to Arcane.
In response to adding Raid Utility to the Arcane spec. What kind of "unbalance" aspects could be seen if you gave Arcane Barrage a haste buff along the lines of the Scorch buff (the 5% crit to spells).
I could see this adding back in the Arcane Barrage into the rotation for raid utiity, while maybe lowering dps a little due to having to cast it, but scorch does pretty much the same thing for those lucky "Scorch Slaves".
Call it Arcanist's Gravitation or something along that line causes all melee and magical attacks to "gravitate" to the target increasing haste by 5%. To offset the fact that it costs more mana to apply this debuff (18% vs 8%), the damage could be left alone and the debuff could last 45 seconds as opposed to 30. A single application instant cast debuff to start a boss fight with would bring some utility to the spec. Not sure how this might affect PvP, but just a thought that might be passed on to developers.
Unfortunately haste only works as a player buff/debuff, in its current state it would be impossible to work it as a boss debuff along the lines of scorch or misery
Not much of a change. The rotations are basically the same, it only means that we'll have to evocate more than we previously predicted. It also means that AB spam is almost certainly out of the picture unless under extreme situations and we're basically limited to balancing ABSpam4MBAM and evocation.
With this data, let's have a look at the Mana-to-Damage conversion rates.
1) 142 dps, 525 mps => 0.27 DPM. Which is the terrible AB spam conversion.
2) 93 dps, 38 mps => 2.44 DPM from using MBAM early before it falls off.
3) 53 dps, 19 mps => 2.79 DPM from using 3/4MBAM or if it falls off.
4) 251 dps, 69 mps => 3.64 DPM from using AB4AM, with 3MBAM or if it falls off.
Not using Evocation is roughly gaining 10k dps while losing about 4k mps. So, roughly a 2.5 DPM conversion, which is 10 times better than using pure AB spam, just that you have a rough comparison. The other cycles below seem valid for everyday use though.
Now, let's have a look at how much mana is granted by 1 Intellect in a 5 minute fight, ignoring multipliers for now.
1) 15 Mana, or 0.05 MPS
2) 18 Mana from Evocating twice, or 0.06 MPS.
3) 0.02 MPS from increased spirit regen with 3/3 Meditation. (Roughly, assuming 1.5k Int, 50 MPS while casting.)
4) 0.03 MPS from Replenishment.
So, assuming we have to use Evocation twice to keep our AB4MBAM cycle running, 1 Intellect grants us 0.16 MPS, or 0.2 MPS with BoK/AMind, which is at least 0.49 DPS from the mana alone - unless you're at the edge of AB Spam.
1 Intellect also grants 0.2 SP (0.29 DPS) and 0.0076% crit (0.35 crit rating, 0.31 DPS).
Of course, the above are general estimates. But you see, 1 Int from gear/gems grants roughly 0.5 DPS from Mana and 0.6 DPS from Crit/SP if you're not at the brink of AB Spam and use Evocation. If the extra damage allows you to cut back Evocation, the effectiveness of Mana slightly drops, but the quality of the result remains. [Talisman of Resurgence] et al. are so high in the lists because the mana from Intellect is nearly half its DPS value.
If you replace the Intellect by the SP/Crit counterparts, these items look far less interesting, exactly because of the extra mana. Which, in all honestly, we've been trained to fully ignore in the last 4/5 years
then why does it only say to am when there's 0.4243977/0.865788 sec left on buff?
I'll try to explain a bit more in detail how optimal cycle calculations work so that it'll hopefully make more sense.
The cycle calculations are made based on spell selection logic which tells what spell to cast based on the current state. There are two levels of states. Casting state is the current status of buffs and cooldowns. There are a lot more possible states than what you get shown in the output of computing optimal cycles. There are states for all possible durations left on missile barrage that can happen when choosing to cast a spell.
Casting states are then combined into control states. A control state is a collection of casting states where you'll necessarily make the same decision. This can be for example because you can't distinguish between proc actually being on or not until after the fact. Several casting states can also be grouped together because of performance concerns. This is the case for current implementation of arcane cycles. All states that only differ in missile barrage buff duration are grouped together in one control state. The states you get listed are representative states from each control state. What this means in practice is that you should ignore the duration of missile barrage buff when interpreting the cycles.
Ideally the current control states would be split at least in two, one control state for buff running almost out and one for the rest. Main problem is that the current algorithm is exponential in number of control states and it already takes a very long time to compute, so that isn't exactly feasible until I implement a better algorithm. Essentially it boils down to an optimization problem with bilinear-fractional constraints.
In response to adding Raid Utility to the Arcane spec. What kind of "unbalance" aspects could be seen if you gave Arcane Barrage a haste buff along the lines of the Scorch buff (the 5% crit to spells).
I could see this adding back in the Arcane Barrage into the rotation for raid utiity, while maybe lowering dps a little due to having to cast it, but scorch does pretty much the same thing for those lucky "Scorch Slaves".
Call it Arcanist's Gravitation or something along that line causes all melee and magical attacks to "gravitate" to the target increasing haste by 5%. To offset the fact that it costs more mana to apply this debuff (18% vs 8%), the damage could be left alone and the debuff could last 45 seconds as opposed to 30. A single application instant cast debuff to start a boss fight with would bring some utility to the spec. Not sure how this might affect PvP, but just a thought that might be passed on to developers.
New utility design moves away from forcing lower dps to apply raid buffs. Scorch is just out-dated. Newer buffs work more along the lines of the Moonkin 13% spelldamage, ie they get kept up as part of the normal rotation.
I think it has been suggested before to give arcane mages a wrath of air totem effect for raid utility. Currently only a shaman of any spec can provide that buff, at the cost of dropping another totem such as windfury. Not really a problem in 25 man raids, but happens often in 5 or 10 man groups, with zero or one shaman. There are mage-type mobs with a spellstealable "aura of arcane haste" that provide a party-wide haste buff. They could add something into the Netherwind Presence talent where your arcane missiles or barrage provide a raidwide haste buff or self buff that provides an aura, for x amount of seconds, which doesn't stack with wrath of air. The 6% personal haste would be left intact as the "selfish benefit" for when you have enough shaman in group, and this wouldn't require any real rebalancing of arcane damage because we're not gaining anything new we wouldn't have in a raid setting, just some needed utility.
With this data, let's have a look at the Mana-to-Damage conversion rates.
1) 142 dps, 525 mps => 0.27 DPM. Which is the terrible AB spam conversion.
2) 93 dps, 38 mps => 2.44 DPM from using MBAM early before it falls off.
3) 53 dps, 19 mps => 2.79 DPM from using 3/4MBAM or if it falls off.
4) 251 dps, 69 mps => 3.64 DPM from using AB4AM, with 3MBAM or if it falls off.
Not using Evocation is roughly gaining 10k dps while losing about 4k mps. So, roughly a 2.5 DPM conversion, which is 10 times better than using pure AB spam, just that you have a rough comparison. The other cycles below seem valid for everyday use though.
Now, let's have a look at how much mana is granted by 1 Intellect in a 5 minute fight, ignoring multipliers for now.
1) 15 Mana, or 0.05 MPS
2) 18 Mana from Evocating twice, or 0.06 MPS.
3) 0.02 MPS from increased spirit regen with 3/3 Meditation. (Roughly, assuming 1.5k Int, 50 MPS while casting.)
4) 0.03 MPS from Replenishment.
So, assuming we have to use Evocation twice to keep our AB4MBAM cycle running, 1 Intellect grants us 0.16 MPS, or 0.2 MPS with BoK/AMind, which is at least 0.49 DPS from the mana alone - unless you're at the edge of AB Spam.
1 Intellect also grants 0.2 SP (0.29 DPS) and 0.0076% crit (0.35 crit rating, 0.31 DPS).
Of course, the above are general estimates. But you see, 1 Int from gear/gems grants roughly 0.5 DPS from Mana and 0.6 DPS from Crit/SP if you're not at the brink of AB Spam and use Evocation. If the extra damage allows you to cut back Evocation, the effectiveness of Mana slightly drops, but the quality of the result remains. [Talisman of Resurgence] et al. are so high in the lists because the mana from Intellect is nearly half its DPS value.
If you replace the Intellect by the SP/Crit counterparts, these items look far less interesting, exactly because of the extra mana. Which, in all honestly, we've been trained to fully ignore in the last 4/5 years
First off, where are your DPS/MPS numbers coming from? I don't see them anywhere in the post you quoted... 0.67 DPM for AB spam? I would burn almost 15k mana for a single 10k AB crit?
Your comment on us being trained to ignore mana, I think is key - we have been totally ignoring mana constraints, and, for non-arcane specs, that is exactly what we must do to maximize DPS.
There are, however, additional limiting factors. Normally, mana can only be considered for direct conversion to damage when damage output is limited by available mana - the converse of the original statement. If you run OOM, the additional mana can be directly used and turned into damage. For most classes and specs, Damage is directly a function of time and DPS - mana is only a constraint in that it limits the time you have to DPS.
However, arcane is not the "normal" case. Mana can indirectly be turned into damage by either allowing you to forgo evocation(s) or by allowing you to "upshift" to a higher DPS/MPS rotation (careful, a higher DPS but lower DPM rotation may result in lower overall damage if it is not sustainable). Your Arcane damage function now resembles a multi-state function where the time spent in each state is determined by the available mana.
@Kavan: Thank-you for the clarification. I was under the (mistaken) illusion that these number-strings represented rotations instead of control-state instructions. Seeing them as instructions makes much more sense.
Unfortunately haste only works as a player buff/debuff, in its current state it would be impossible to work it as a boss debuff along the lines of scorch or misery
Could it possibly apply a debuff so that, when the boss is hit by players, the players recieve a buff? For example, the arcane mage applies the debuff. Everyone who damages the boss while this debuff is up gets a buff, 5% haste on your next cast/attack. Could that be used to get around the current mechanics, do you think?
First off, where are your DPS/MPS numbers coming from? I don't see them anywhere in the post you quoted... 0.67 DPM for AB spam? I would burn almost 15k mana for a single 10k AB crit?
Your comment on us being trained to ignore mana, I think is key - we have been totally ignoring mana constraints, and, for non-arcane specs, that is exactly what we must do to maximize DPS.
The numbers are the differences between the cycles that I quoted from Kavan's post. The 1st number is the difference between cycle 1 and cycle 2 - i.e. the DPS and MPS difference between ABSpam and AB4MBAM - the 2nd number is the difference between cycle 2 and cycle 3, et cetera. So, it tells you how much more DPS you gain from switching up a cycle, and how much more mana you spend when doing so, and the quotient of those numbers is the DPM conversion.
It's actually 0.27 DPM for AB Spam, so it's even worse than you assumed. You have to spend 35k mana for a 10k damage gain. That's because AB Spam is extremely expensive while only offering a very slight damage increase because AB4MBAM is extremely good DPS already while being vastly more efficient.
When it comes to the importance of mana, I think nothing sums it up better than Sancus' signature.
Could it possibly apply a debuff so that, when the boss is hit by players, the players recieve a buff? For example, the arcane mage applies the debuff. Everyone who damages the boss while this debuff is up gets a buff, 5% haste on your next cast/attack. Could that be used to get around the current mechanics, do you think?
The simplest way seems to apply it in a similar way as Replenishment. The typical "10 players on your group get the buff for X seconds on your cast" seems plausible and effective.
How come my Embrace of the Spider trinket become the 2. best trinket for my arcane spec when i turn on Unlimited mana? It says its better than Illustration of the Dragonsoul, Reign of the Dead and Eye of the Broodmother (I suspect the Reign of the Dead is bugged). The only better trinket is Flare of the Heavens.
I have Scales of Fate in the other slot.
When i turn Unlimited mana off, it is still the 5. best.
Could it possibly apply a debuff so that, when the boss is hit by players, the players recieve a buff? For example, the arcane mage applies the debuff. Everyone who damages the boss while this debuff is up gets a buff, 5% haste on your next cast/attack. Could that be used to get around the current mechanics, do you think?
Why are people thinking of the most complicated possible ways to apply this? It could be something as simple as the Moonkin Aura buff, except granting haste instead of crit.
Wether intentional or not, Missile Barrage no longer reduces the channeling time of Arcane Missiles however it does still reduce the mana cost by 100%.
This is in complete contradiction to what the tooltip of Missile Barrage states on the PTR.
The lag/delay on the Missile Barrage proc (elaboration by me on page 105) remains intact.
Wether intentional or not, Missile Barrage no longer reduces the channeling time of Arcane Missiles however it does still reduce the mana cost by 100%.
I think we can assume unintentional since it would entirely break the spec.
I think we can assume unintentional since it would entirely break the spec.
Unless it's part of a set of changes intended to bring Arcane Barrage back into rotation, with AM being for efficiency and Barrage being for DPS. But that'd naturally require quite a bit more adjustment to make it work.
At Veridian Dynamics, we can even make radishes so spicy, people can't eat them. But we're not, because people can't eat them.
The numbers are the differences between the cycles that I quoted from Kavan's post. The 1st number is the difference between cycle 1 and cycle 2 - i.e. the DPS and MPS difference between ABSpam and AB4MBAM - the 2nd number is the difference between cycle 2 and cycle 3, et cetera. So, it tells you how much more DPS you gain from switching up a cycle, and how much more mana you spend when doing so, and the quotient of those numbers is the DPM conversion.
It's actually 0.27 DPM for AB Spam, so it's even worse than you assumed. You have to spend 35k mana for a 10k damage gain. That's because AB Spam is extremely expensive while only offering a very slight damage increase because AB4MBAM is extremely good DPS already while being vastly more efficient.
When it comes to the importance of mana, I think nothing sums it up better than Sancus' signature.
I see - I was under the mistaken impression that these were absolute instead of relative numbers. Spending 35k mana for a 10k damage increase over ABSPAM4MBAM does make sense. I sort of wish there was an intermediate option between AB Spam and ABSpam4MBAM, because "upshifting" to AB Spam is, as you said, so horribly inefficient it is almost never worth it.
A better way to bring Arcane Barrage back into rotation would be to make Missile Barrage affect all three spells that consume the Arcane Blast buff instead of just Arcane Missiles. For example,
Missile Barrage (3 talent points)
When you consume the Arcane Blast buff, the consuming spell gets the following additional benefit
Arcane Missiles: Hastes channel time by 8/16/25% per stack of Arcane Blast.
Arcane Barrage: Affects one additional nearby target for each stack of Arcane Blast. Additional Arcane Barrages deal 25/50/75% damage.
Arcane Explosion: Allows 1/2/3 stacks of Arcane Blast to be preserved when casting.
These are just examples. In this situation, Arcane Missiles would be reliably and progressively hasted to its current benefit as you build the stack, and would be the preferred single-target spell (while becoming weaker for PvP.) This would also allow for downranked rotations ( < 4 stacks) to be predictably and progressively mana efficient, giving partial haste.
Arcane Barrage would be the preferred clearing spell not just for moving, but also for when more than one target is in your line of sight, as commonly occurs in some boss fights. Thus its utility would increase from where it is now, without being overpowered for PvP. Plus it makes sense for a spell called "barrage" to be able to, you know, barrage an area when charged up with Arcane Blast.
And Arcane Explosion would get a much-needed buff to compete with blizzard and PoM-flamestrike, as you'd be able to cast it with the Arcane Blast buff a few times before having to refresh Arcane Blast.
Using Missile Barrage in that manner (as a tool to benefit multiple spells, which we can choose between to take advantage of in different ways) would start bringing back some more decisions into our rotation, without necessarily making it overpowered in PVP and other situations.
At the moment theres one decision for almost every occasion, "Got Missile Barrage up? Use Arcane Missiles." Since AB spam is likely to be more out of the question after the recent PTR changes.