When you say manually editing the cooldown, what do you mean by that? If you're changing the source that should work I believe. I think all you need to change is in Solver.cs modify EvocationCooldown = (240.0f - 60.0f * MageTalents.ArcaneFlows); in InitializeEffectCooldowns.
I find rawr to be working very well on my fire spec mage, however when I load from armory in arcane spec, the program fails to recognize any reforging or enchants on any of my gear. Is there a reason for this that I am missing? Or a bug?
Blizzard just significantly changed the format of character files on battle.net, so our parser broke in a few places. Expect to manually fix a few things like enchants/reforges/etc until we get it updated (soon).
Ok, I'm a decent fire mage and a semi smart person. But this is my first post on here so I feel like it's probably still a dumb question.
I recently acquired some new gear so I went to Rawr to optimize it, and it now wants me to reforge all of my crit to haste. I know that at some points, haste exceeds crit when you can reach another "tier" of haste, each tier being when it will contribute to your dot ticks. This makes sense to me. Otherwise it seems like with so many of our abilities being based off crit, that crit is the superior secondary stat under normal circumstances.
Anyways, when I was completely reforged for crit, I was well over the +2 combust tick (812 haste), and the next "tier" of any form for haste is the +3 combust tick/+3 other dots when herod tick (both occur at 1996 haste). When I fully reforged to haste in the manner that Rawr suggests, I am still a bit under this soft cap (sitting at about 1880 I believe).
So I am confused. Is haste really superior to crit in this plain ole plain ole situation? I'm not getting any dot ticks.... so it's just fireball cast speed really. Doesn't seem like it outways crit to me. Could someone please explain to me why Rawr thinks haste is better for me? Is my thinking of crit = amg wrong? Am I using Rawr wrong? Is Rawr wrong? A little math would be great too if someone has a good explanation.
I'm having an interesting thing happen with Rawr 4.1.5. I am raiding Frost, and I understand that the loose stat thresholds to meet are Hit cap > soft crit cap (33.33 or 33.34, or 33.4, or whatever) > Haste > Mastery. However, when I attempt to use the Optimization feature of Rawr, it devalues crit compared to haste and does not attempt to hit the soft crit cap. Am I missing something here? Shouldn't Rawr be optimizing around ensuring that DF and BF always crit?
Rawr doesn't optimize for any thresholds explicitely. It optimizes the total damage done which depends on the underlying model. In the last version I improved the model to include the FoF generation from Freeze and FFO. The FFO part is very heuristic at the moment and I plan to replace that with another model, so it's possible that it's having some undesired effects. If you want to help with this the best you can do is provide some spell breakdown comparisons between Rawr and Simc on the same gear set or before and after optimization.
I've noticed that Rawr (as well as SimC) is coded to provide Ignite damage from Flame Orb critical strikes, but in my own testing, this is not the case. I have examined the combat log and found no indication that Ignite will proc from Flame Orb. Does anyone have any evidence that would prove otherwise?
I originally stumbled upon this by running an arcane spec with two points in Ignite. Swapping to fire (and three points in Ignite) results in the same effect (or lack thereof, rather).
Edit: The following is found in the fire mage compendium:
Although Flame Orb does steady damage per second, it's not considered a DOT. Therefore, it does not directly interact with Flashburn or Combustion. However, the damage Flame Orb does can critically hit, leaving an Ignite, which can interact with those mechanics.
Either this needs changing as well, or I am simply wrong.
Is it possible that it is overvalued? Maybe a bit. At the same time SimC is most likely undervalueing it because it doesn't do a good job of cooldown stacking. Then you add in fights with damage multipliers and on use becomes even better than what Rawr shows in the default settings.
As for the reason why it might be overvalued a bit it's because the cooldown stacking constraints are slightly more relaxed than what can be accomplished in real fight due to performance reasons. If you're interested in what kind of impact it has I'd suggest playing with advanced solver settings and cooldown editor.
Kavan, when I asked this question some time back I was told that the reason was that Rawr wasn't (yet) modelling int procs correctly and that the Darkmoon Card Volcano was almost certainly - as suggested by SimluationCraft - a better trinket than the Soulcasket. I've been monitoring the forum to check if the situation has changed, but have heard nothing more on the subject since then. Is Rawr now modelling Int proccs better? You clearly believe the Soul Casket (and for that matter the Jewlcrafting trinket) is a better choice than the Card.
I'm not aware of any significant errors in int proc modeling in Rawr. I'm not sure what is the source of that assertion. It specifically accounts for spell power, crit and mana regen from increased int as well as the effect on mana adept. If needed I can provide some more technical details on how int procs (or procs in general) are modeled, but it's all available in the source.
One thing that has helped me feel more comfortable when I'm wondering if haste is really that good, is to recognize is that having more haste directly contributes to getting more crits in a fight (by allowing more casts through the fight), and therefore also creates additional crit-based procs.
Dendore - you mentioned haste tiers - are these still relevant? I thought the game engine had switched to a more linear approach whereby whatever duration needed is added to accommodate the last fractional tick.
re: flame orb crits not generating ignites, I just confirmed this on target dummies. Although as we know, target dummies are not always valid indicators of what happens on other targets.
This has been something more or less recurring when optimizing my mage with the arcane spec using rawr, but I reckon this last result was a bit over the edge. Here's the situation:
I'd been using the optimize feature of rawr to try and find better gear choices for my arcane mage, with thoroughness set always to maximum but using normal clicks. This always ended up finding a gearset with slightly less DPS than the starting one; strangely enough, no matter how many times I tried doing this, it would never return the same gearset, even though the starting one and the optimization options were always the same.
With this in mind, I decided to set throughness to maximum and use the ctrl-click option to get the absolute highest thoroughness. Running calculations over 10 hours on my Windows 7 with Dual Core 2 Duo CPU, it finally found a solution... it went from the initial 25083,01 DPS to 22856,97 DPS, the biggest drop I've witnessed when using rawr.
I reiterate those choices were for an arcane spec.
I don't really know if this is happening due to me messing up with any optimization options, but I don't think that's the case and, since I don't have much experience with programming, that's why I decided to post this here. If it's of any help, here's a link to my character's armory: . If anyone here would like any other information you feel is useful to try and reproduce the problem/investigate the matter, feel free to ask.
When I encounter problems like this with Rawr-Mage-Arcane, I usually try to tick off/on the "Arcane Light" advanced option, which seems to affect algorithm stability. In particular, "Build Upgrade List" only seems to work for me with "Arcane Light" off.