Look at this positive way - it frees 2 talent points and provides AGI for you to boost your DPS

Yah but Yo! that makes very little sense - if an item is an upgrade it should be an upgrade regardless of the rest of your kit. How can something be good now, but after 2 upgrades to other slots, be worse off?

That is why I made a post on page 58 running through 600+ data points with main conclusion being " there is a solid 54.5% chance that actual AEP is different from the suggested by more than 30%" (soory for self-quote).

No. It is not constant, otherwise new static AEP value would be born.
" there is a solid 54.5% chance that actual AEP is different from the suggested by more than 30%" (soory for self-quote 2).

2 minor upgrades will be not enough for the guy in the example in order to change AEP from 5.5 (that is what local AEP is for him) to 1.3. He happened to be way off from stat combo that values hit rating as 1.4 AP, namely he has too much AP and not enough Crit/Hit.
It does not mean that he has to change his equipment towards the combo where there is 1.4 ratio to improve his DPS, however boosting his DPS by choosing +hit instead of +ap will drift his local AEP from 5.5 towards 1.4 and at some point beyond that.

Look at this another way - how valuable additional +hit rating is when person reaches hit cap? Is it 1.4 ap? Or is it more like 0.0? And it is not changing suddenly like 1.4, 1.4, 1.4 1.4 oops 0. It is drifting towards 0 the more + hit rating you have given AP remains the same (still with a big downfall at the end). Hope this helps.

Of course it does, Malan.

[Cataclysm Gauntlets]
77 AP
1.056 % Crit
1.52 % Hit
(374 HP + 396 Mana and ~ 0.3% Spellcrit)

[Fel Leather Gloves]
62 AP
1.08 % Crit
1.07 % Hit

[Gloves of Dexterous Manipulation]
77 AP
1.52 % Crit
0 % Hit

and compared to eg [Cyclone Gauntlets] (those aren't even on the list!)
59 AP
0.924 Crit
1.20 % Hit
(286 HP + 396 Mana + ~ 0.3% Spellcrit and 5mp5)

Conclusion: There is no huge range between all of them (beside set boni and a bigger mana- and healthpool).

To the "Yo! Discussion",
You are using 12% hit instead of 22%, that's a large difference. It's around hitting only 8 of 9 out of 10 attacks and should result in heavily lower ur/flurry uptime.
I still believe, that 22% hit is a little too highly set for a good aep comparison, it's to close to the hitcap (that's around 25.5 to 27.5 if i remember correctly) while you can "easily" obtain 1000 more ap and/or 10% more crit.

Malacort
What weapons are your rogues wielding? With S2/T3+ weapons, doesn't matter if they are sword/mace/dagger your rogues should get better results from Wf with a combat specc. Only a multilate rogue could really prefer grace of air over windfury for reaching ~ 45-50% crit raidbuffed with the ashtongue trinket buff (that result in a lot more cp from seal fate and ruthlessness/relentless strikes talents) and deadly poison allways up.

Sure, that is why I got expected different AEP value from what would be with more hit.
And that is exactly why "average AEP" provided poor test guy with wrong suggestion for choosing gem. And that is why it should not be used by him. And you feel right that it is not worth reaching hit cap while you can "easily" obtain 1000 more ap and/or 10% more crit because local AEP (true AEP for the specific stat combo) will favor AP and Crit much more than "average AEP". So, following what average AEP suggests is evil. I said it before, did not I?

This and Yo!'s last post are things I agree with. I think of it as building a square, in many respcts. If your objective is to creat the maximum area, and your choices are to lengthen side x or side y, generally you're going to increase the area more by increasing which ever one is already shorter.

Now, we're not dealing with a square, but the attributes (hit/crit/haste/ap) can be seen analagously to one for the purposes of qualitative but not quantitative comparison. If you're really low on hit compared to the other dimension, you're going to get more bang for the buck increasing hit. This is a lot like what stigmata said in a recent post about his gear.

My problem with Yo!'s initial posts were that they made it seem like you got nothing out of assigning the average AEP values. Personally, I agree that AEP changes as you gear up differently. However, my original reason for writing the sim and finding results was to confirm that hit was the be all and end all of enhance shaman statistics - it was the conventional wisdom, but I hadn't seen it proved anywhere, so I wanted to see why. It was quite the shock to find it so underwhelming at certain gear levels.

I still think that it has valid uses - and for a starting shaman with average gear, it gives you a direction to aim for. Once you get to the boundary cases where you've used it to the point that you have 0 hit rating and 1800 AEP, I think it's certainly very worthwhile to re run the sims (or just comprehend it internally) to the point where you have to realize that you're going to get much more of a return on hit than you initially did.

back to the rectangle. Here's how I see shaman decision making 4 months ago based on the gear available then: It's H wide by AP tall.

When H and AP = 10, the area is 100.

It turns out that for $5, you can increase H by 2 or AP by 5. At that point in the general shaman world, people were choosing to increase H by 2 units instead of AP by 5.

Fast forward to the current day, when that rectangle is 8H wide by 20AP tall, for an area of 160 square units.

Now for the same $5, increasing AP to 25 gives you an area of 200. If you were to increase the H dimension by 2, however, you'd have a total area of 200 also - they've normalized.

AEP is good for shaman over a very wide variety of gear. However, when you get to the extremes that AEP will lead you, you will indeed start making sub-optimal gear choices. This is confounded more by the fact that the higher AP goes with raid buffs (people are now raiding with 3k+ AP often with buffs) that hit is going to become more and more important (as will crit, for that matter).

I don't disagree with Yo's statements taken by themselves. What I have disagreed with is that for the vast majority of shaman, the current AEP ratings would misguide them. Gear up through kara, when using AEP, won't let you neglect hit, so shaman have been ok. Accordingly, when I ran the sims, note that AP stopped at around 2500 and crit around 31%. Now that times have changed, and we can see where gear is, a new version should probably be done (esp with haste gear, which was unforseen by me, which would more strongly value +hit). It's impossible to use a constant set of values that is good for everyone, but I think that the ones that have been used to date have helped a lot of shaman from sinking into a trap of "Max +hit first, and then worry about your other stats".

As I've said repeatedly, if tornhoof can get a closed form of parameterized attributes working, that would be my preferred solution.

Yo! - I hope you realize that if I've sounded harsh, it's not because I disagree with the notions you have set forth. I just don't want people thinking that as they grow from dungeons to heroics to SSC/TK, they have been making bad choices. I think the AEP are ok up through that level. Now that we are beyond that, there probably does need to be a rejiggering of values (and preferably in a parameterized way).

they do use s2/t3+; we have 1 mace spec rogue using bt weapons, 2 sword rogues with s2/hyjal/bt weapons. They were telling me to use goa over wf since their poisons did more dmg than wf was giving them.

Sure. But you neglect to address the thrust of my argument: static AEP values are useful at "reasonable" gear combinations. Sure, if you just have 2800 AP and 0 hit rating, that won't be as good as 2600 AP and 100 hit rating even though the AEP values say otherwise. But there's no set of gear which would allow you to obtain 2800 AP and 0 hit. So why pretend that there is ever going to be a shaman running around with 2400 AP and 47 hit when trying to invalidate the use of AEP figures?
I think most of us understand and concede this point, but it doesn't follow that AEP is therefore not useful for the vast majority of enhancement shamans -- or even the majority of enhancement shamans with a min-max raiding mindset who read this thread.

I will finally end by noting that I still haven't seen any number for the standard deviation of simulated tests at a given AP/crit/hit number -- unless it's lower than 9%, AEP is actually a better measure (given that the standard deviation of AP across all of Disquette's tests was 9%.)

Disquette - it is a honor to talk the same theorycrafting language with you
Squre example was something that I wanted to post next!
Total damage = (1 + haste%) * (raw damage from weapon + damage from AP*upkeeptimeofunleashedrage(crit)) * (total chance to land an attack including Crit with x2 weight) * 7/5 * upkeeptimeoffurry(crit) * etc.. is a greatly simplified analytical model of Shaman's damage. Saddly, it can not be used instead of step-by-step simulations because the mathematical expectance Mf(x,y) is different from f(Mx,My).
But it is a good illustration that damage can be viewed as multidimensional cube or square if only 2 arguments are being compared.
So lets do it now

Lets assume Damage =f(Ap,Hit)= Ap*Hit*K(unknown static coefficient K that absorbs all other attributes).
If ap = 100, hit = 1 damage will be 100, 1 point of additional hit will increase total damage by 100, 1 point of additional ap will increase total damage by 1. Guess what - these are called derivatives, and the pair of these derivatives forms gradient (100, 1), wich is also known as AEP weights, 1 hit is worth 100 points and 1 ap is worth 1 point
If ap = 50, hit = 50 damage will be 2500, 1 point of additional hit will increase total damage by 50, 1 point of additional ap will increase total damage by 50. Weights are: 1 hit is worth 50 points and 1 ap is worth 50 points which equals to 1:1.
If one starts nude with 1 hit and 0 ap and has a budget of 100 Drgagon kill points that he can spent somehow 1 DKP per one stat point - he will achieve max damage output by spanding 49 points towards Hit, 50 points towards AP and the remaining point towards any stat. This will result in 2500 damage. If, for some reason this person thinks that AP is always better he will spent all 100 points on AP scoring 100 damage.

Squre example was something that I wanted to post next!
Total damage = (1 + haste%) * (raw damage from weapon + damage from AP*upkeeptimeofunleashedrage(crit)) * (total chance to land an attack including Crit with x2 weight) * 7/5 * upkeeptimeoffurry(crit) * etc.. is a greatly simplified analytical model of Shaman's damage. Saddly, it can not be used instead of step-by-step simulations because the mathematical expectance Mf(x,y) is different from f(Mx,My).
But it is a good illustration that damage can be viewed as multidimensional cube or square if only 2 arguments are being compared.
So lets do it now

Lets assume Damage =f(Ap,Hit)= Ap*Hit*K(unknown static coefficient K that absorbs all other attributes).
If ap = 100, hit = 1 damage will be 100, 1 point of additional hit will increase total damage by 100, 1 point of additional ap will increase total damage by 1. Guess what - these are called derivatives, and the pair of these derivatives forms gradient (100, 1), wich is also known as AEP weights, 1 hit is worth 100 points and 1 ap is worth 1 point
If ap = 50, hit = 50 damage will be 2500, 1 point of additional hit will increase total damage by 50, 1 point of additional ap will increase total damage by 50. Weights are: 1 hit is worth 50 points and 1 ap is worth 50 points which equals to 1:1.
If one starts nude with 1 hit and 0 ap and has a budget of 100 Drgagon kill points that he can spent somehow 1 DKP per one stat point - he will achieve max damage output by spanding 49 points towards Hit, 50 points towards AP and the remaining point towards any stat. This will result in 2500 damage. If, for some reason this person thinks that AP is always better he will spent all 100 points on AP scoring 100 damage.

Squre example was something that I wanted to post next!
Total damage = (1 + haste%) * (raw damage from weapon + damage from AP*upkeeptimeofunleashedrage(crit)) * (total chance to land an attack including Crit with x2 weight) * 7/5 * upkeeptimeoffurry(crit) * etc.. is a greatly simplified analytical model of Shaman's damage. Saddly, it can not be used instead of step-by-step simulations because the mathematical expectance Mf(x,y) is different from f(Mx,My).
But it is a good illustration that damage can be viewed as multidimensional cube or square if only 2 arguments are being compared.
So lets do it now

Lets assume Damage =f(Ap,Hit)= Ap*Hit*K(unknown static coefficient K that absorbs all other attributes).
If ap = 100, hit = 1 damage will be 100, 1 point of additional hit will increase total damage by 100, 1 point of additional ap will increase total damage by 1. Guess what - these are called derivatives, and the pair of these derivatives forms gradient (100, 1), wich is also known as AEP weights, 1 hit is worth 100 points and 1 ap is worth 1 point
If ap = 50, hit = 50 damage will be 2500, 1 point of additional hit will increase total damage by 50, 1 point of additional ap will increase total damage by 50. Weights are: 1 hit is worth 50 points and 1 ap is worth 50 points which equals to 1:1.
If one starts nude with 1 hit and 0 ap and has a budget of 100 Drgagon kill points that he can spent somehow 1 DKP per one stat point - he will achieve max damage output by spanding 49 points towards Hit, 50 points towards AP and the remaining point towards any stat. This will result in 2500 damage. If, for some reason this person thinks that AP is always better he will spent all 100 points on AP scoring 100 damage.

Dude, just edit your profile and stick in an armory link.

This is dumb. Just say what you have to say without turning your post into a mountain of words.

What is "reasonable" gear combinations for you? Please take a look at big_test.xls - all of the combos recorded there were taken in-game. There are no 2800 AP 0 hit combos in there. AP runs from 800 to 2500 in 100 ap steps, hit changes from 12% to 25% in 50 hit rating steps, crit changes from 18% to 31.5%


If you will accept combos recorded in big_test.xls as "reasonable" than AEP is not useful for the vast majority of enhancement shamans that have such combos in game... Average AEP may provide worst suggestions to the most non-average geared players such as Elitist Jerks raiding BT . Especially with availability of +haste items. It will be actually possible to count exactly how many shamans will not benefit from "average" AEP if projects such as Armory Musings... will provide database for free download.
Each test at a given AP/crit/hit number was run for 1 hour, there were no repeated tests for the same AP/crit/hit combo. Disquette knows more about it as he is the one responsible for that test

Like I said, I wasn't on a comp that could play wow and im not spec'd enhance atm so posting my armory wouldn't have done anything but linked my resto gear. I'll log out in my enhance gear when I log tonight though.

Hi guys, first of all i'd like to say thanks for the glut of helpful information here. I switched to enhancement about 2 weeks ago and am currently trying to gear up to be useful in raids, your compilation has helped immensely.

On to my question and i hope it hasn't already been asked (i've read as much as humanly possible, but didn't spot this specifically anywhere.

I'm currently using [Reflex Blades] Main hand, and am struggling to find anything offhand thats slower. My current off hand isn't showing on the armory, but its a 2.60 green axe (65.6dps if i remember right) and i'll be picking up the High Warlord axe this weekend, but even that is only 2.60.

Given the problems with fast OH, do i want to just ditch reflex blades altogether and grab another high warlord axe? its .10 speed diff but from what i'm reading it'll still gimp me badly.

How do the strike mechanics work for weapons with matching speeds? is having 2 2.60 speed weapons a problem for windfury or does the main hand weapon always take priority?

I actually don't understand the whole AEP discussion at all, we have several more or less valid models of shaman enh dps. Each and every model has a certain weighting towards one or more stats, my model is more fond of hitrating and hasterating than of critrating.

My current model rates the attributes at a 2.2k AP,22% hit, 25% crit gear for:
1CR = 1.18HR
1CR = 1.045Str (incl. BoK)
1CR = 0.794 HasteRating

Actually I don't know why anyone should care about AEP values for unbuffed non raid envionments, there is no reason to use AEP if you never have BoM, BoK, Battleshout and Elixir/Flask.

Now to the mathematical finer points of the whole discussion.
your attributes/stats are not linearly independent, +crit increases +ap to a certain degree and +haste (if you take flurry as a haste rating (which is a valid assumption)) etc, the only
stat which is practically independent of everything else is AP. Other than that I do agree that the calculated AEP values are only valid for a fairly small window.

To me it is kinda obvious that it is not possible to compare two individual items on your equipment based on AEP, it has to be obvious for me, otherwise writing my program would have been lost time

Keep your MH, a 2.6 OH is fine. You'll be able to pick up a 2.6 MH in KZ. With matching speed weapons the MH should swing first but there's some situations where it can get out of sync, but neither hand has a 'priority' for procs.

I assume you are referring to [Fool's Bane], which i must admit i'd not seen before
What i meant by priority was swing order from the start of an attack, rather than procs, as in both weapons don't strike simultaneously even if the speed is the same. If this is the case then MH would have the first chance to proc windfury, even though both are the same speed.

I assume? In either case thanks for your advice, i look forward to forcing my guild to fight Terestian (oh they are going to love that...).

Yeah i checked out the paw, and decided to go for the slower hitters if i could. Of course if it drops i can always adapt Plus it will make it easier to find a off hand thats slower...
hmm...
methinks you might have thrown a monkey in my carefully balanced plan.

# Crypt Fiends around Azeroth and at Hyjal Summit are now more totem friendly.

I'm thinking of rolling a Shaman but I'd like a little bit different gameplay from my Rogue so I'm wondering if 2-handers are viable for an Enh Shaman and what the spec might look like? I've tried searching a couple of times but never found anything, all the talk is about dual wielding and comments would be much appreciated.

Disquette, Yo! et al - perhaps what we need are 3 (or more) sets of AEP that establish some baselines. ie, while leveling 1-70 with pretty minimal AP/Hit/Crit available, here's the AEP you need. For 70 normal/heroic/KZ instances, use these other values. Once you hit XYZ values of stats, roughly SSC/TK time frame, use these. After you've geared in SSC/TK and you're at the extreme top end of stats, now use these other AEP for Black Temple and Hyjal.

@Jelu - 2H is relegated to PvP for the most part, and even in PvP it may not be the clear winner anymore. DW is superior for PvE purposes. The days of shaman swinging big huge 2H's is pretty much over. It varies enough from a rogue either way.

This is my first post so hi everyone.

@Blackmoon: I just started again in July and I've been 70 for about a month now gearing up so i feel I'm on the same level as you and my little advice is to get the level 60 High Warlord 2.9 speed weapon as it's still serving me since around lvl 64 .
Just yesterday I managed 735dps according to wws so it is indeed a more than viable weapon and above all very easy to get. I think it would out-damage the level 70 version assuming you get Battleshout in the raid.
It's so good I don't feel I'm working towards an upgrade in Runic Hammer but rather I'm working just to test if it's better.

On the Karazhan weapons, get what drops first. I was lucky enough to get the Paw almost as soon as I dinged 70 and it helped alot for farming and grinding to gear up (skill up leatherworking). To be honest, I don't really find much difference between the Paw and the Decapitator.

The use on the decapitator is fucking awesome and I wish we had a talent to throw our weapons. I use it on transitions and it was pretty fun for the 1 week I was using it in pvp.