When you program the module is there a way to bypass some gear when optimizing or building an upgrade list? I use Rawr for a balance druid also and it does those two functions much faster than for my elemental shaman. In fact it can take a long time to build an upgrade list for the shaman.

Yeah I'm working on a faster solution with cache and whatnot. Maybe I'll add an option to choose between "thorough" and "quick" calculations.

I have absolutely no idea what it's like to code an individual model, but is there a way to have it check say 30 in each slot (15 above, 15 below)? I see it modeling items that aren't BiS at level 70, much less level 80.

Lava Burst receives hugely inflated values (in the 45 000 region) for critical strikes on v2.1.9. The value is displayed immediately when I import a profile and adding raidbuffs correctly increases the critical damage value. Probably an error at calculating the coefficient on LvB? If you import e.g. SoNoFu's character, you see:

"Lava Burst: 6634-7359 / 47431-53619*"

which doesn't look right. And if you fill out all the raidbuffs you could possibly dream off, your estimated DPS for a 3-minute fight goes through the roof.

Did you read the first page? Specialy the first few posts upto post #16

The problem was adressed and he has a fix for it ready, think we just have to wait till next 'release'.

Compiling a 2-day old build doesn't adress it, probably the fixed version hasn't been checked-in yet.

edit: Wrong thread.

Mmmm, would you say SoNoFu's character has mostly Best In Slot items?
The World of Warcraft Armory

Current build of the Estimator (not checked in yet) says so, except for the trinkets and one of the rings.

BiS based on the current (2.1.9) build:

Head - [Valorous Earthshatter Helm]
Neck - [Wyrmrest Necklace of Power]
Shoulders - [Valorous Earthshatter Shoulderpads]
Back - [Pennant Cloak]
Chest - [Valorous Earthshatter Hauberk]
Wrists - [Bands of Mutual Respect]
Hands - [Valorous Earthshatter Gloves]
Waist - [Cincture of Polarity]
Legs - [Leggings of Voracious Shadows]
Feet - [Footsteps of Malygos]
Ring - [Signet of the Malevolent]
Ring - [Signet of Manifested Pain]
Trinket - [Dying Curse]
Trinket - [Illustration of the Dragon Soul]
Main Hand - [Torch of Holy Fire]
Off Hand - [Voice of Reason]
Totem - [Totem of Hex]

I don't want to turn this into a "BiS" thread, I'm just curious if that looks right. I was under the impression that [Benefactor's Gauntlets] and the other four pieces of T7 were BiS for those particular slots. I understand Rawr is balancing the items due to budget and hit. Also, being a JC will probably make a difference in some of the choices.

Anyway, keep up the good work. Can't wait to see the new build.

It depends a bit on the parameters obviously but the new build gives similar results. I believe Astrylian will put out a new version this weekend, so you have plenty of time to play with it. Do note that the simulator is slow and a new version of the simulator that should be faster is still a work in progress. The question is: is it worth the trouble? The goal is not an accurate simulation of a fight, but a prediction of what gear is better; if the gear prediction is accurate though the predicted DPS is off the scale, that's not a problem :-)

I agree. I do believe the formulas should be as accurate as you can get, but leave the rotations to the spreadsheets. I would spend a little extra time on "proc" items such as trinkets to make sure they are modeled correctly. Who knows, Blizzard may start adding on-proc rings and necks again.

From previous BIS discussions based off of the SEIC spreadsheet, that BIS set is at least close, but Leggings of the Wanton Spellcaster has been said to be preferable to Cincture of Polarity as a source of +hit. I don't know why Rawr would find differently. Anything that increases confidence in the validity of the Rawr conclusions is welcome.

Thanks for working on this!

Several updates

First of all, I removed the option to simulate. Though it was giving more accuracy, it was too slow for the Rawr standards.

I also implemented Water Shield and Glyph of Water Mastery. Beyond whatever you may have expected, I do not actually have a high level shaman, hence I "forgot" the basics.

Elemental Weapons has been implemented as well.

Glyph of Flame Shock (or rather, the option of not having it) has been implemented. I'm wondering if I'm doing something wrong here. I assume you will always want to have Flame Shock on your target when you use Lava Burst and that you always want to cast a (critical) Lava Burst whenever it is available. Hence, if your Flame Shock is not glyphed you will use it immediately after you use Lava Burst (because then you will get most out of the dot, it makes little sense to cast it later and FS also costs more mana than LB, so it's not bad to use clearcasting on Flame Shock) which means you will spend less time casting Lightning Bolt. Much less. Per minute you cast ~4 more flame shocks and ~4 less lightning bolts.

Using Best In Slot gear, I get
7,6 lava burst per minute
7,6 vs 3,9 flame shock per minute (3,7 more without the glyph)
43,3 vs 47 (3,6 less with the glyph)

The amount of dps lost is so huge that you HAVE to use Flame Shock unless you want to severely gimp yourself. Is this observation correct or am I making a mistake somewhere?

Also, [Leggings of the Wanton Spellcaster] is easily best in slot in 2.2.0b1+

BiS in my current build, with about every buff you can get

Head - [Valorous Earthshatter Helm]
Neck - [Wyrmrest Necklace of Power]
Shoulders - [Mantle of the Corrupted]
Back - [Pennant Cloak]
Chest - [Fallout Impervious Tunic] and [Valorous Earthshatter Hauberk]
Wrists - [Bands of Mutual Respect] and [Unsullied Cuffs]
Hands - [Benefactor's Gauntlets]
Waist - [Cincture of Polarity]
Legs - [Leggings of the Wanton Spellcaster]
Feet - [Footsteps of Malygos]
Ring - [Band of Channeled Magic]
Ring - [Signet of Manifested Pain]
Trinket - [Dying Curse]
Trinket - [Illustration of the Dragon Soul]
Main Hand - [Torch of Holy Fire]
Off Hand - [Voice of Reason]
Totem - [Totem of Hex]

A combination that also works: [Leash of Heedless Magic] + Runed gem + [Boots of Impetuous Ideals]

Just fyi, btw, I don't want this to be a BiS thread either, but it might be interesting if a list like this puts people on a track to find bugs, which obviously helps Rawr development.

Based on those selections, the best T7 bonus you could have would be the 2 piece. Is the 4 piece bonus not worth more?

I ran mine with the new build and the results were quite different than previously. [Band of Channeled Magic] I had not considered before. I'm not saying that's incorrect, just different.

One last thing. I was not satisfied with the gem choices that it made. However, I'm a little confused on how to set up custom gems templates. I read the info page at the Rawr homepage, but it still seems a bit unclear. Any info there or check another thread?

Custom gem templates. When an item with sockets is considered, every template is applied. So if you select 5 templates, you will get 5 items. Obviously, duplicates are removed. What gems would you rather see? I can change the default templates of course. I will at least add "Rigid" and "Veiled" to the combinations. Because I was working on the BiS set with every buff on (stupid, stupid, stupid) I missed them. I've added Veiled, Rigid and Lambent gems

Regarding T7-4 set bonus, it is calculated, it's simply not worth a great deal and T7-2 is worth even less (nothing, if you don't go out of mana). As soon as I remove the Misery (3% spell hit) buff, T7 becomes more valuable simply from +hit.

CritDamage = NormalDamage * (1 + CritMod)
CritMod = 0.545 * (1 + LavaFlows + ElementalFury + LavaBurstBonus)
LavaFlows = .24 (24%)
ElementalFury = 1 (100%)
LavaBurstBonus = .1 (10%)
.545 = (1.50 * 1.03) from meta gem

Alternatives:
1. CritMod = 0.545 * (1 + LavaFlows + ElementalFury) + LavaBurstBonus
2. CritMod = 0.545 * (1 + LavaFlows + ElementalFury) * (1 + LavaBurstBonus)
3. CritMod = 0.545 * (1 + ElementalFury) * (1 + LavaFlows) * (1 + LavaBurstBonus)

The CritMod formula we currently use is based on the SEIC spreadsheet and is pretty similar to what happens in similar cases in other Rawr modules.

T7 4 pieces is worth about 28 spell power. If I change the implementation to alternative 2, it's still not good enough to choose 4 pieces of T7. One of the reasons seems to be that haste rating is pretty good, better than spell power. I can't find any mistakes in the code, so it seems to be true...

I figured it out. My lazy gene must have come out, because it was actually quite simple.

I am a JC, so I only give it a few choices:
JC Only - [Runed Dragon's Eye]
Red - [Runed Scarlet Ruby]
Yellow - [Potent Monarch Topaz]/[Rigid Autumn's Glow]
Blue - [Glowing Twilight Opal]
Prismatic - [Runed Scarlet Ruby]

However, after looking at my SEIC, I'm wondering if I should add some haste gems into those choices. Any suggestions?

Ignoring JC gems for the moment.

Red: [Runed Scarlet Ruby]
Yellow: [Rigid Autumn's Glow] or [Veiled Monarch Topaz] if you are bellow the hit cap, [Reckless Monarch Topaz] if you are over the hit cap and want the socket bonus, [Runed Scarlet Ruby] otherwise
Blue: [Vivid Forest Emerald] if bellow hit cap (and more preferable to use this first vs yellow/orange hit gems), [Glowing Twilight Opal] to get meta/socket bonuses, [Runed Scarlet Ruby] otherwise.

You should never gear for crit.

Yes. That was supposed to be a [Veiled Monarch Topaz], not a [Potent Monarch Topaz]. Sorry for the confusion.

How major of a difference are you looking at between the estimator and the simulator? As it relates to gear selection, not DPS simulation.

I was hoping to be able to use the simulator to detect certain errors, for example, that a rotation in practice won't be as perfect. E.g. maybe you can't refresh Flame Shock perfectly and maybe Lava Burst can't be cast immediately whenever it's off cooldown. With an estimation, this is simple guesswork. Right now, it's the "perfect" modus: things are calculated for a perfect Lava Burst cast and a perfect Flame Shock recast and a perfect fill of the time in between by Lightning Bolt.

Then on the other hand, perhaps it doesn't matter, because in general the gear choices would be the same. However, there might be cases that adding extra haste really doesn't help much, because you will only have to wait sometimes or you will be less efficient.

That's when you start allowing for things like that in your model. Of the three rotational options I have in SEIC, only one assumes everything is perfect, and the other two are either perfect + X, or just straight X.

RAWR VERSION: 2.2.0b3

2. Improved Shields talent is not effecting mana regen with water shield for overall DPS/time calculation for OOM.