Thanks for the feedback, and I will check that out.

On the items - yep, latency field does nothing now, forgot to add that to the notes - started to add it, but then realized I wasn't totally sure how the last change (admittedly awhile ago) to spell casting queueing ("not a spell queue") would actually be modeled. I am sure there are some posts on it buried in the spreadsheet thread, just haven't had a time to look.

T6 - May not be in the distribution, if you search for upgrades they should populate. Will see about getting them included in next release if Astralyn hasn't already done so.

ISB Uptime - I explicitly said that it was modeled only for a single user, assuming they got all charges. Yes, there are better models, no, I haven't implemented them yet.

T6 set bonus - well, if no T6 gear is in there probably not going to be a set bonus either? But regardless, no set bonuses are modeled at all now.

Heroism, trinkets, meta's, etc. Currently no on use or chance to proc effects are modeled. Not all meta gems are correctly modeled as well. (Only effects on the "understood" stats are taken into effect -e.g. spell damage, combat ratings, etc.)

Re: Other gems - if there is an issue there then that is a problem with the model, and I will definitely check that out - thanks!

Also didn't see it mentioned, but any source of mana return other than lifetap isn't currently accounted for.

Honestly I think there is a misconception on your part (which may have resulted from me not being more explicit about the current beta). No, this is not nearly as fully featured (with respect to modelling) as the spreadsheet is. No, it does not do everything I would like it to, but I think that's true of most things here. Basically I would like to get feedback (good, bad whatever), and I do appreciate your list. If there are specific issues with the models, the more detailed the information is the faster it will get fixed, as I have limited time so tend to go for the low hanging fruit. So certainly keep the comments coming, just wanted to be clear that I am in no way advocating that this currently superceeds the spreadsheet - this is a *beta* piece of software (at least the warlock module is def. a beta piece).


Yep, no reason at all that can't be done - I'll probably follow the Rawr.cat model that was mentioned later and simply have the values display as DPS * 100.


Thanks, I'll get that changed asap (and say thanks that that seems to be the only misspelling ).

Interesting... Does that mean that no .NET apps at all work on Wine? I can't imagine so, since Mono for Wine works, but I don't see how any .NET app could avoid all generic font use, without explicitly declaring a font for every single control in the app (and thereby pissing off their users who change their system font).

tried to copy gdiplus.dll from windows xp and set wine to use it, but either I'm doing something wrong(probably) or it doesnt help, as I get same error.

Aye, in that bug report you linked, there's a linked bug report for what happens when you do exactly that, which is also bugged.

I am missing the influence of four parts Moonkin T5.

The Set-bonus is a great improvement and it increases our DPS significant. Though it would be great to have the option to activate it, like four Parts Tirisfal.

Of course four Parts T5 would change the optimal Rotation to IS,3xSF or MF,3xSF instead of Starfire spam (which Rawr is showing me at the moment).

Thanks in advance.

I'm terrible at this sort of relational math, and was secretly hoping you had already intended to something else but just had a bug in the formulas :B

In the meantime with the current release, to compare gear, sorting by damage(red bar) will give us the appropriate ranking for the gear. This is because the red bar already factors in the extra casting time(denoted by the blue bar) when calculating the approximate DPS, so the blue bar is just kind of there in reference. I guess you already noticed this and removed it though.

So I spent a little time brainstorming ways to display an accurate damage ranking list. I currently haven't the slightest idea of how to implement this but here are the basics:
1) First calculate the best cast cycle like you currently do. Save the values of all the items.
2) Lock this cast cycle and temporarily change the regen to something huge like 1000mp/s. Redo item calcs.
2a) Items with high spirit/mp5/int on them will see a reduced number. Take this difference, and turn that part of the bar blue. This is benefit the player is getting with the regen stats on this item.
2b) Items with low spirit/mp5/int on them will see an increased number. Take this difference, and tack this portion onto the end of the bar(color it green maybe?). This portion signifies the damage potential you can gain on the item if you eliminate the mana problem with gear or consumables/buffs etc.

So you will have something like(numbers are totally made up):
Belt of Ooming: 100 damage
[-----Red(50)-----][-----Green(50)-----]

Belt of Booming: 50 damage 50 spirit
[--Red(25)--][------Blue(60)-------]

We can see currently the belt of booming is giving us better overall damage, but we can also see that if I get enough regen upgrades in other areas, I'll get more overall damage from the belt of ooming.

Sorting by red+blue gives you a current damage ranking
Sorting by red+green gives you a potential damage ranking.

I really hope that made sense.


One small bug:
Improved Divine Spirit currently gives 0 benefit in Rawr. Should be 10% of spirit in damage.

And one request:
Curse of shadows under the "buffs" list.

Set bonuses are currently not included. Neither are trinket procs or meta gem procs, both of which would change the DPS and/or mana regen figures. I'm trying to think of ways to fit trinkets and meta gems into my current implementation, but set bonuses are going to need quite a bit of work between everybody on the project, I feel.

[e] To answer the above poster, I decided to just stick with a purely damage-only figure for now. What you're suggesting is a novel approach and technically feasible, but I want to get the model refined first. I may come back to it later. As for Improved DS, I hadn't noticed; that would explain some things.

Noticed on Rawr.mage:

It seems to be modeling Arcane Blast in the spell rotation for Deep Frost (which is honestly news to me for a deep frost spec), and setting Tirisfal pieces (notably the hat and shoulders) as my best gearing options over items like [Cowl of the Illidari High Lord] and [Hatefury Mantle]. Does this mod just have a jones for Arcane gearing and spell rotations?

Also the trinket gearing seems to be wonky. Modeling the same 10/0/51 spec at 132 hit, if I select Trinket Slot 1 it gives me Skull of Guldan, Hex Shrunken Head, Darkmoon Card: Crusade, Quagmirran's Eye, and Icon of the Silver Crescent as my best options in descending rank. Quag's Eye does not TC out to be the better trinket than Icon in my other TC modules, such as Vontre's spreadsheet.

Selecting Darkmoon Card for Slot 1 and then going to Trinket Slot 2 should theoretically give me Quagmirran's Eye as my next best option to DMC if I'm to understand this right, but instead it ranks Icon of the Silver Crescent above Quag's Eye, which is how it should be. Selecting Icon/DMC combo gives more overall DPS in the Stats column than the Quag/DMC combo, so that appears to be working properly, just if you select Trinket Slot 1 does it seem to be giving unusual choices for optimization.

Fire spec (2/48/11) is strange as well. It models Serpent-Coil Braid as my best option even over Icon when I'm over hitcap at 166 hit, but ONLY in Trinket Slot 1. Selecting Slot 2 tells me to use Icon over Serpent Braid. How does the trinket modeling work, 'cause it's confusing the heck out of me.

Rustyshrapnel, can't the difference in which trinkets are suggested for you depend on which trinkets you already have in those slots and thereby are replacing. For example, when you talk of the fire spec 2/48/11 and being over the hitcap, which trinkets do you already have in your trinket slots when you look at suggestions for slot 1 & 2?

If you look at the [Serpent-Coil Braid] in particular, in my own tests with Rawr it will move around quite a lot from high up on the list to quite low depending on minor changes of the options that affect my mana pool/mana regen, which is the way it should be in my opinion. Both [Serpent-Coil Braid] and [Quagmirran's Eye] should be pretty significantly affected by how much mana you have (first giving you more mana back, second spending it faster).

this is because it looks items up on the armory (as far as i know...) when you enter an item ID. Items can be manually entered, if theres just one or two you're keen on though...

Your stats certainly are changing. When you're looking at trinkets for Slot 1, it's not counting DMC toward your stats. When you're looking at Slot 2, it's not counting Icon toward your stats.

As Bloody Sorcerer said, that's because Rawr looks up items from the Armory, not Wowhead. If you can convince Wowhead or Wowdb to have their item data available via xml (and by item data, I don't mean tooltip), I'd happy add an option to look up via them instead of the Armory.

Regarding Arcane Blast in a deep frost spec, it's still going to be useful to spam blast near the end of a fight to dump your remaining mana, no matter the spec. 2T5 helps with that significantly, which is why T5 pieces that would give you 2T5 are rated highly. Or perhaps you've got 3 on, and it's rating the last 2 highly due to the benefit of 4T5 that it would give you. In general, if you see items of a certain set ranked especially high, it's probably because one more piece of that set would give you a very powerful set bonus.


I'll see about defaulting the armory lookup to your current character's server, good idea.


dopefish, let me know if you have any questions about adding the set bonuses; should be quite easy to do so.