Caster Scaling in 5.1 & Beyond
by, 07-02-2013 at 09:05 PM (11928 Views)
We donít have an updated version of SimulationCraft built yet for the latest round of 5.2 changes, and in the case of Arcane the default rotation will need an overhaul. This doesnít mean we canít look at how things are in 5.1 and make some inferences in how we could expect the eventual numbers pass in 5.2 to go, based on current 5.1 stat weights.
The methodology for collecting the data was to run two caster only raid profiles through SimulationCraft, generating stat weights for the T14N and T14H profiles: Balance, Arcane, Fire, Frost (Mage), Shadow, Elemental, Affliction, Demonology & Destruction. From here we can enter them all into a spreadsheet, and start to crunch some numbers.
One of the first things I did was to average the weights for Intellect, Hit, Crit, Haste and Mastery, with the ratings doubled to reflect how they are itemised on gems. The results didnít quite match what I had expected, so I dropped out Hit as it doesnít really have any contribution to increasing damage once the cap has been reached. The new figures looked a bit better, roughly matching the distribution of DPS across all classes.
From these numbers, I can see that Fire is going to be a bit of an outlier in this, as it has a higher average stat weight than Affliction, but lower overall DPS. This is in part due to the fact that Fire scales much better with Crit than any of the other classes, by a good 35-48%, while also scaling reasonably well with both Haste & Mastery. Every other class has two clear favourites of these three stats (although I should point out that the Elemental T14H profile is in a slightly odd spot of having haste as the worst stat, but this is because haste is the only stat that has a negative feedback effect on itself and this profile has just enough to trigger it, likewise with Shadow).
The interesting results started coming when I compared the DPS figures against the individual stat weights. For the secondary stats there is a fair bit of variation but with Spell Power a pattern emerged. By plotting the weight for Spell Power against the resulting DPS, a fairly linear line emerged for both data sets. Using Spell Power as a basis for this rather than Intellect was so that specs like Fire, which have a much higher Crit Rating weight, werenít sitting as outliers.
Stop, Graph Time
The above graph shows the DPS of each spec from the SimCraft output, compared with the Spell Power weights. Normal profiles are green, Heroic are red. The dot order from left to right is as follows:
Normal: Elemental, Shadow, Balance, Destruction, Demonology, Fire, Affliction, Frost, Arcane
Heroic: Elemental, Shadow, Balance, Destruction, Fire, Demonology, Frost, Affliction, Arcane
The actual values vary from the trend line by +/- 3% (the graph makes the variance larger because Iíve zoomed it in, otherwise you wouldnít be able to see much).
The Obvious Conclusion Is...
Assuming that SimulationCraft is relatively accurate, we can conclude that the performance of any DPS caster spec is directly tied to how valuable Spell Power is. Increase the value of Spell Power, and DPS goes up correspondingly. Secondary stats will certainly affect DPS too, but they have a much lower impact on the whole picture, but itís worth noting that due to the positive feedback loops in stat value interactions, boosting the value of Spell Power will also boost the value of secondary stats too.
You can also see from the graph that the Heroic distribution is wider than the Normal one. This is because the increase in Spell Power value is between 0.19 dps (Elemental) and 0.43 dps (Affliction). Arcane is still out ahead due to itís 0.32 increase.
But what does this mean?
There are three spec groups in this: those that are in the middle, those out on top, and those falling behind. Using this data we can see that Arcane, Affliction and to a lesser extent Frost are all performing much better than all the other specs looked at, both in raw DPS and Spell Power Scaling. This gap looks like it will only continue to increase if changes arenít made. At the other end, Elemental and Shadow are falling behind a bit, and itís possible that this gap will also increase going forward.
I know that all of this is looking at 5.1 information, but with 5.2 in a state of flux at the moment itís not practical to make conclusions from it. However, we can consider what weíve learnt here with the 5.2 changes announced so far, and see that outside a fair few talent changes & re-balancing, the only spec that may end up changing its scaling is Arcane.
Hopefully the upcoming numbers pass on the PTR increases the performance of Elemental & Shadow into that central pack. It would only take a 5-8% increase in the value of Spell Power to make the required difference. On the other hand, Arcane, Affliction & possibly Frost may need a 10-15% nerf to bring them back into the pack
Total Trackbacks 0