Charred Twilight Scale isn't set up correctly in Rawr nor in Simcraft. It is actually on a 50s icd and a proc chance on damage done (any damage, dots included). Although it's uptime * spellpower on most fights is slightly lower than DFO at least for a fire mage.
The mage model hasn't changed for some time now, so the only reason for differences between Rawr2 and Rawr3 that I can imagine is in some gear having wrong stats in one of them. The only way to tell what is going on would be for you to post the character xml where you see the issue.
Actually Rawr2 and Rawr3 have different stats for the CTS trinket. Found the issue.. just wish I knew which was the best way to input them to get an accurate gauge on how the trinket compares to H DFO.
There is no planned official support for Rawr4 at level 80 currently. That said the dev version does work for Arcane currently if you can compile it, but there is no precompiled release at the moment. I'm not sure when I'll be adding Fire/Frost support in. From what I've been following theorycrafting on it there's still quite some unknowns at least for fire so my tendency is to wait for more data. But if someone can help setting up a mathematical model for the cycles that would speed up the process.
I'm wondering if and when there is a planned update for rawr which will include reforging? I understand that the numbers may still be a bit of a guessing game for most specs but I'm just curious for a progress update.
Some notes about the current state of the mage model in Rawr4:
- Don't use it to compare talent specs, each spec may have different systematic biases due to modeling approximations
- Fire spec seems to be a lot lower than what simc is getting, still investigating
- Frost isn't modeling pet freeze
- Mirror Image glyph modeling is questionable
- By Spell output only lists direct damage component of combustion, dot parts are attributed to each source spell currently
- Fire isn't modeling aoe situations, but you have a setting to specify Pyromaniac uptime
- If you're not using arcane, you can go to advanced options and disable mana segmentation to improve performance
- Flame Orb is modeled as averaged by default, but you have option to treat it as cooldown at slight performance cost
- For comparison charts to work you need to place at least one talent point so it knows which specialization to use
- Clearcasting is still modeled as plain 10% proc
- Boss Handler settings aren't used at the moment
There are several tricky things regarding modelling this. For a start, there's the "lag" involved with not always being able to click the targetting reticle as soon as it appears. Secondly, although you want to be sure you have 1 FoF charge for Deep Freeze, it might be desirable to Freeze up to X seconds early if you have BF proc. Thirdly, if there's a natural FoF proc X seconds before DF comes of cooldown, do you wait and use that for DF instead of IL? so that you can Freeze afterwards for 2x IL/BF, or is that a dps loss because of above-mentioned targetting-click-lag?
I noticed something that felt strange when using Rawr today.
When i went to the "relative stat value" pannel, for a fire mage in heroic blue / crafted epics here are the values i did find :
That surprised me quite a lot since I thought that, as is said in the fire mage compendium :
Mastery and Haste are both considerably weaker than Crit. Mastery is slightly better than haste.
What especially surprised me was that Mastery was valued 3 times weaker than haste and 3.5 times weaker than crit when the results found in Lhivera's Simulationcraft directory gives values closer to each other in a similar situation
Mastery being a lot closer behind crit in particular.
I just took the time to manually setup my character in the latest Rawr, but the mage model only shows 'survivability' for me. The 'dps' number in the lower left corner stays at 0. Is there a problem in the mage model or am I doing something wrong? Other models shows 'threat', 'dps' and/or 'survivabilty'.