Wrong, the only problem with your calculations is that the glyph is a flat amount on top of your current cast time, this isn't how it works, it reduces the cast time from the base (confirmed on PTR) and only then your haste is taken into acount, in short it's only 5-6%.

No matter how much haste rating you have the glyph is ~5.26% haste.

Ex.
3 / 1.05 / 1.03 / 1.0526 / 1.25 (25% haste from gear) = 2.11
2.85 / 1.05 / 1.03 / 1.25 = 2.11

This is the same way the Shadowpriest t10 4set bonus works, the flat reduction is cast time is calculated before haste meaning the actual haste gain of a flat cast time reduction is alot less than what pure haste brings it up to.

IE 3 second cast with .5 flat reduction = 2.5s then you add your haste in.

10% of 3 = .3 - More reduction gained but still slower
10% of 2.5 = .25 - Lesser base cast gives lesser hasted gain but faster cast overall.
etc etc

Cast time is generally Cast_Speed/(Haste_Buff_1*Haste_Buff_2* ... *Haste_Buff_n*Personal_Haste). I know, once those division and multiplication signs are in, it's harder to understand.

Let's solve for the haste of the glyph now when assuming raid buffs (fyi All_Haste_Buffs = Personal_Haste*Haste_Buff_1* ... *Haste_Buff_n):

(3.0) / (All_Haste_Buffs * Glyph_Haste) = (3.0 - 0.15) / (All_Haste_Buffs)
3.0 / Glyph_Haste = (3.0 - 0.15)
3.0 / (3.0 - 0.15) = Glyph_Haste
Glyph_Haste ~= 1.05263 = 5.263% Haste

That is the value of the glyph given as a separate haste buff - obviously independent of other haste buffs.

With the 5% crit of the fireball glyph, it is easy to calculate a "crit rating" value of the glyph for the fireball spell since crit is additive. Simply it is 5*45.91 = 229.55 crit rating.

Since Haste is multiplicative, the "haste rating" of the new glyph for the fireball spell is a bit harder to understand (as it scales with your own haste). So Haste_Rating = (1.05263*Personal_Haste_Percent - Personal_Haste_Percent)*100*32.79. If you had a personal haste percent of 0% (unlikely, but not impossible), then the haste rating value of the glyph is (1.05263*1 - 1)*100*32.79 = 5.263*32.79 = 172.6 haste rating for the fireball spell. So at the very worst, the new fireball glyph is worth 172.6 haste rating.

Now imagine a mage with 30% personal haste as a fire mage (yeah now that's unlikely, unless they were previously arcane spec). As I said, the new fireball glyph scales with more haste so we should expect a higher haste rating. The "haste rating" of the glyph is (1.05263*1.3 - 1.3)*100*32.79 = 224.3 rating. So at "best", the glyph is worth 224.3 haste rating.

As you can see even at a high amount of haste, the glyph offers about 224 haste rating for the fireball spell. At current, the fireball glyph offers a flat 229.55 crit rating.

What does this mean? It means that if you really wanted to lower your crit (because it was "too high"), it would be better to simply swap out crit gear for haste gear (as blizzard's stat weights crit rating = haste rating on same ilevel gear) than allow blizzard to help you do it - you'd get more bang for your buck. Unless you have some ungodly amount of haste (and we're talking 30%+ haste) for a fire spec, this glyph is nerfed for 3.3.3.

I believe this is correct unless I made some huge error in some assumptions.

For fire/ttw 3.3.3 Rawr and Simcraft prefer crit over haste, enough to gem Potent vs. Reckless, this is due to several things first Pyroblast is now a much harder hitting spell, to cast more Pyroblasts you need more crit/haste, due to 2t10 more crit -> more haste for fire, and lastly the glyph change means a loss of 5% crit and gain of around the same amount of haste.

So the more crit you gain the more benefit 2t10 provides, which slightly inflates the value of crit for fire 3.3.3.

In a specific gear set Rawr value crit at 2.82 and haste at 2.46 spellpower 2.42. (988 haste and 1088 crit)

If I switch to 3.3.2 mode the scales tip haste is valued at 2.46 crit 2.43 and spellpower at 2.3.

At feiro:

The new Glyph of Fireball is not necessarily a "haste" buff. The glyph grants a reduction in cast time (5% of talented base), while the effect of haste is an increase in casting speed.

And actually the glyph scales negatively with haste as it relates to a reduction in cast time. This is because the nominal reduction in cast time is lower due to the lower base cast time.

The effect of haste on casting speed, and the number of "additional spells cast" in a given period of time remains the same. 10% haste = 10% more spells cast in Y time.

To clarify some of the math:

.15/3 = 5% (correct way to look at it)
.15/2.85 = 5.26% (wrong way to look at it)

So for all intents and purposes the glyph is a 5% reduction in base cast time, not 5.26%

The way to explain haste speced fire is as follows:

1% Haste = 32.78998947 Haste Rating
5% from Totem
3% from Druid/Pally
Y = Base cast time(sec)
X = Haste Rating from Gear

Final_cast_time = Y / (1 + X/32.78998947/100) / 1.05 / 1.03

Here is an example of how the new glyph scales with haste.

Given: 1000 haste from gear

3 / (1+1000/32.78998947/100) / 1.05 / 1.03 = 2.12566

2.85 / (1+1000/32.78998947/100) / 1.05 / 1.03 = 2.01938

2.12566 - 2.01938 = 0.10628

So with 1000 haste from gear and raid buffed the glyph offers a 0.10628 second reduction in cast time (3.54% of the 3 second base)

Given: 500 haste from gear

3 / (1+500/32.78998947/100) / 1.05 / 1.03 = 2.40690

2.85 / (1+500/32.78998947/100) / 1.05 / 1.03 = 2.28656

2.40690 - 2.28656 = 0.12034

So with 500 haste from gear and raid buffed the glyph offers a 0.12034 reduction in cast time (4.01% of the 3 second base)

So as haste increases the reduction in cast time offered by the glyph decreases. However, the glyph will always grant an additional 5.263157% more spells cast in Y amount of time. This is because the reduction in cast time required to cast X% more spells in Y amount of time gets smaller as haste increases.

This can be shown using both of the previous examples:

With 1000 haste from gear:
(2.01938/2.12566) - 1 = -4.999% = change in cast time

With 500 haste from gear:
(2.28656/2.40690) - 1 = -4.999% = change in cast time

Because you are casting more spells in less time the value of all your other stats increases as it relates to your DPS.

Overall this is a slight buff to fire. (not yet proven by math)

EDIT: Wrong information corrected in red

This.

Don´t you really think that the developers had taken all this into consideration and just made this little change to the fireball mechanics to address the mobility issues? I do, the cast time reduction is just worth the 5% crit you drop for a tank&spank/dummy/simulation in DPS, but it will result in an increase on the overall DPS for movement fights or ones with multiple targets.

On the other hand, the real fire buffs for 3.3.3 are the combustion lowered CD, and the additional 15% SP to damage and the TTW effect on Pyroblast.


The fireball glyph is just an adjustement.

Being an Arcane spec mage looking at going fire spec once this patch hits, 30% personal haste is not that hard to get. Using Rawr 2.3.11 (standard fight values and full raid buffs) I've put togther a fire set that beats my arcane spec dps (12,440 vs 12,316) by just swapping out three pieces of gear (wrist, gloves and pants) and a trinket. All but one of these pieces (lvl 264 T10 pants) has been passed on several times and I snagged them instead of letting them get DE'd. While my current gear is good, it's by no means BIS, my 25 man raid group hasn't gotten past Sindra yet. I did this initaily with the intent of going fire for the scorch debuff in my 10 man heroic group since we run sans lock to help out the other casters. In my fire set I'll have 926 haste rating (just over 28% personal haste and 38.7% raid buffed), 1348 crit rating. This compares to 1,014 haste (almost 39% personal haste w/talents and 50.1% raid buffed) and 1,099 crit rating in my arcane set. I had to use two +12sp/+10 hit gems in this gear to reach hit cap (over cap by 1 point) so this set will get better once I pick up a few other pieces and rebalance it. I made this set with pieces I already had or knew I could easily get only.

I plan on duel specing arcane and fire (depending on fight and buff needs) so I will not be swapping my yellow +sp/+haste gems for +sp/+crit. I did in Rawr and the result was only +6dps overall whan I changed all yellow gems, not worth it.

"This formula is obvious:
Cast time with glyph = (3.0 - 0.15) / (All_Haste_Buffs)

This formula is also obvious:
Cast time with glyph = (3.0) / (All_Haste_Buffs * Glyph_Haste)"

My algebra may be rusty, but these two equations are not equal. Please show me how you arrived at the 2nd one.

"Using systems of equations (and it's not a very hard system either), you solve for Glyph_Haste and find that it is exactly 3/2.85. The reason why it can be considered exactly like a haste buff is because it's a constant - much like a shammy totem. Everything related to haste buffs can be applied this: stacks multiplicatively with other haste buffs and increases number of spells by (3/2.85)x."

Again, I'm no math major, but the glyph does not behave the same way as haste buffs. It REDUCES base cast time rather than INCREASING casting speed. You are correct that it is a constant, but it does not multiply with other haste buffs the same way a totem does because it affects the BASE CAST TIME not casting speed. (I have a feeling we are splitting hairs, but I'm not sure)

"the glyph will always grant an additional 5% more spells cast in Y amount of time" - No, it will grant 3/2.85 times more spell casts in Y amount of time."

This is where you are correct and I was wrong.

As my math shows, the glyph always offers 4.999% reduction in cast time regardless of other haste buffs.
But what that 4.999% reduction equates to is a 5.263157% increase in CASTING SPEED. Therefore you will be able to cast 5.263157% spells in Y amount of time.

"Overall this is a buff to fire."

Here again I was wrong. I have been arcane for some time now with FFB as an off spec. Last time i ran Rawr as fire it valued haste more than crit. I just ran rawr with my gear as fire and it valued crit more than haste. So yes, because crit costs more itemization points than haste for the same percentage, the glyph is a nerf.

POST EDIT: After thinking about it, I decided to look further into the effect of this as it relates to overall benefit. Even though Crit is favored over haste as a stat on gear (as shown by rawr) the glyph is a buff to FIREBALL DPS (the spell, not the spec)

In my gear, and all else being equal:
Fireball DPS:
No glyph: 8512.63
Crit glyph: 8823.44
Haste Glyph: 8964.10

So we trade the potential for additional HS procs for more DPS on our filler spell. In practice this appears to be a buff because we usually have more HS procs than we can cast.

"What was pointed out in the previous post was that the glyph is worth about 224 haste rating IF the mage already had 30% haste on his gear (and about 172 haste rating if the mage had no haste at all). Since 224 < 229 (30% haste is a good upper bound), the glyph is a nerf. Of course we're just talking about haste rating for fireball, but it's worth a good chunk of total dps."

This is incorrect. The glyph offers 172.57889 haste no matter how you slice it.

5.263157 x 32.78998947 = 172.57889

And just for a reference:
45.91 Crit rating = 1% Crit
5 x 45.91 = 229.55

So you lose 56.97 itemization points.

EDIT: Edited for wall of textyness

As for the glyph, according to Rawr using 3.3.3 mechanics otherwise and Fire BiS gear:
DPS(no glyph): 15124.92
DPS(3.3.2 glyph): 15605.04 (+480.12) <- Used a workaround to get this, I simply deactivated the FB glyph and added the 4t9 bonus to an item I was wearing anyway
DPS(3.3.3 glyph): 15636.02 (+511.10)

You have to consider that usually the reason why crit becomes better than haste as Fire is that in BiS gear with 2t10 you're haste capped for Pyro and LvB so crit provides a damage increase for all abilities while haste only provides one for FB. Since both glyphs only improve FB (and not Pyro/LvB) you can't really convert them to ratings and the haste glyph comes out on top (because >5% more casts in the same time is superior to 5% more crit for an ability that already has a crit chance of 78.24%).

I suppose the second one is slightly harder to understand. I wanted to find out if the glyph could be represented as a regular haste buff. Since the formula for cast speed of a spell is (base_cast_speed) / (HasteBuff1*HasteBuff2*...*HasteBuffn), I just stuck Glyph_Haste as one of those haste buffs. If Glyph_Haste did not behave like a regular haste buff, then the calculations would show that Glyph_Haste is non-constant. Since it was constant, it indeed acts like a haste buff.

Here I wanted to talk about the haste glyph in terms of haste rating (not haste% like before). Doing this allows me to compare the haste rating value of this glyph. That is, I wanted to see the +haste rating difference across the two patches.

Cast time with glyph = (3.0 - 0.15) / (Raid_Talent_Haste_Buffs * (Haste_Rating / 3279 + 1))
Cast time with glyph = (3.0) / (Raid_Talent_Haste_Buffs * ((Haste_Rating + Glyph_Haste_Rating) / 3729 + 1))

-> (3/2.85) = (Haste_rating+Glyph_Haste_Rating)/3729 +1)/(Haste_Rating / 3279 + 1)

Solving for Glyph_Haste_Rating, we get 0.0526*Haste_Rating + 172.58.

So Glyph_Haste_Rating is non-constant and scales with your current haste rating. Plugging in 0 haste (0%) and 983.7 haste (30%), you get 172.58 and 224.32 glyph_haste_rating respectively.

It's fine that the developers want to address the mobility issues, however at current, you can get the same effect for more from just swapping out gear. Think of it this way, do you want 5.26% faster fireball casts (which is what the new glyph will provide)? Suppose your fireball casts were way too slow because you have been stacking crit but still had an ungodly amount of haste, let's say 30% worth of haste rating so I can reuse numbers.

If you swap out 224 crit rating, and replaced it with 224 haste rating, you'd get the desired 5.26% faster fireball casts. In fact, you only swapped out 4.88% worth of crit for it. Remember, this is assuming you already had a stupid amount of haste (and it worsens with less haste). That is, you can currently (3.3.2) switch out 4.88% worth of crit for 5.26% faster fireball spells.

The upcoming patch will not do that. It will take away 5% crit (instead of 4.88%) and give you the same 5.26% faster fireball spells. What is blizzard doing to replace that excess loss? Nothing. Simply nothing. The loss is deeper if you had less haste to begin with (lesser than my proposed 30%).

Sorry, I'm not trying to be an ass when I say the following... If it acted like a duck, quacked like a duck, waddled like a duck, then it is a duck - no matter how hard the duck acted like a reduction on base cast time. That is, bliz could have dressed up the glyph in a number of factors. They could have even said "this glyph now grants you 0.0526*Haste_Rating + 172.58 worth of additional haste rating for your fireball spell" and it would have been the exact same thing.

To put it in another perspective, the haste totem is officially 5% haste. Bliz could have just said, "This totem reduces base cast time of spells that are 3 seconds long by 0.143 seconds, reduces base cast time of spells that are 2.5 seconds long by..."

Yes I've made it pretty clear that I was just talking in the scope of the fireball spell. I wasn't saying you SHOULD swap out crit for more haste. That wasn't the issue. In the next patch, blizzard is going to swap out 5% crit for a 5.26% haste buff for your fireball spell. I'm saying that the 5.26% is underbudget. Being slightly under budget isn't a big issue, but you can currently mimic the same effect by swapping out your gear right now and get more than 5.26% haste.

The haste cap for the fire spec for GCDs is 38.7% haste. Frankly, I didn't think fire spec mages would stack haste since the glyph was obviously minded for fire mages being near the crit cap. 30% haste was thought to be an extremely exaggerated value.

Yep this is completely true.

Just read my post above.

The glyph's power stays pretty much the same, actually in BiS gear it's even a bit stronger.

The BiS Fire gear includes 1163 haste rating, that's 35.47%. Raidbuffed that's 46.5% haste not including 2t10, haste potions, power infusion or heroism. Any of them (especially pointing at 2t10 which is up most of the time) will push you way over the GCD haste cap, thus haste in BiS gear (and actually any reasonable high level gear) doesn't affect Pyro/LvB most of the time. Thus you simply cannot convert the glyph into ratings and draw conclusions from there, just take them as what they are and the new glyph is perfectly fine.

And just as a sidenote: Even with the old glyph you would've been below the crit cap by ~17% in full BiS Fire gear so I don't think that's the reason Blizzard changed the glyph, they just wanted Fire Mages to deliver their damage faster and make the glyph feel a bit more unique and not "Hey it's another 5% crit ontop of my 75%".

Sorry I was assuming mages had high crit and < 30% haste. But I did forget about the 2pcT10. Of course, I've never really seen blizzard care that much about GCD caps anyways. Wrath has been GCD capped at 400 haste for over a year now. Resto shammies have a harder time, but they can reach their haste soft cap too. Destro locks as well have some pretty strict GCD caps. I would very much doubt they redesigned a glyph to alleviate caps of sorts, which was I stated that the reasoning for the change is skeptical.

And quite honestly, I don't really care what the glyph does. It could have said "Generates sparkles that grants 5% more spell power for 3.8 seconds everytime you /dance", I'm just trying to compare the value of the old glyph vs the new glyph. I guess the reason why I don't care is because I'm not a mage, I'm a boomkin. What do I care if mages do 40 dps or 40k dps. All I wanted to do was point out that for most fire mages, this is a nerf because they can simply swap out crit gear for haste gear and get more haste out of it.

First off, haste is way stronger than crit as Fire until the soft cap but you easily get way over that cap without gems in Fire BiS gear. After that cap haste is still about as strong as SP (e.g. Black Magic is still the best 1h enchant) and crit is slightly stronger than SP point by point. In the Fire BiS gear you have 12SP/10Crit in every yellow and red gem slot, resulting in 1594 crit rating including Molten Armor (1340 without it) and 1163 haste rating from gear alone.

Since the haste from the glyph is never wasted (I don't think you can reasonably cap FB on GCD) you'd need to use the way higher value of haste prior to the soft cap if you wanted to compare it to crit rating. If you currently swap out crit for haste gear you'll obviously end up at the same FB cast time with less crit loss for FB, but you're also losing lots of crit for Pyro/LvB and gaining absolutely nothing for those spells.

In the end you cannot really argue it's a nerf. You get around the same DPS as before (even slightly higher on average in Fire gear) while having more mobility.

I disagree with your assessment, but not because of the actual math. I think you're making 2 unrealistic assumptions. At risk of repeating some things that have already been said, I'll try to present what I'm saying in an organized manner.

The first assumption you made was that there's some point at which haste and crit are roughly equivalent in value, and therefore mages would drop one for the other to achieve this "balance". The reality is that haste is the superior stat, at least for all of the gearsets I have come across so far. The "why" part isn't important - whether it's because of high base crit values or anything, it doesn't matter. It's just the raw value that is superior. In other words, nobody will be gemming replacing their haste gems with crit gems just because of the glyph.

The other assumption you made, is that you can take haste and crit values as "pseudo-DPS" values. On theoretical Patchwerk fights in spreadsheets this makes sense, but in practice, any significant reduction to the long cast time of fireball is going to cause considerable change in the playstyle. Target swapping fights, movement fights and the such will all penalize fire specs a lot less significantly. The increased frequency of fireballs is going to change hot streak frequencies and behavior.

Most importantly, with current levels of haste, a 0.15 second cut on fireball, combined with T10ish haste, will likely cause the 2T10 to affect an extra fireball under the regular "hot streaked pyro followed by 3 fireballs" situation. This is a very significant threshold value.

In short, you're trying to oversimplify a problem by estimating "DPS values" from haste and crit. In reality, spec performances just aren't continuous curves that gain X DPS from Y haste/crit.

Elut, I do not recommend using Nibelung. I used it for a few weeks in 10-man and 25-man raids (the 264 version) and have recently switched back to [Illumination]. Illumination is without a doubt a far superior staff, despite being 19 levels below. A 264 mainhand+offhand would be better, but it's even stronger than 50771+[Lightbane Focus] in most situations, given that Hit Rating is so frequent on gear (to the point where I have a more difficult time trying to find places to get rid of it).

I don't believe that the Scorch debuff counts as a secondary spell. Real ways to increase the frequency of Nibelung procs involve Molten Armor and Frostfire Bolt (or Frostbolt with Fingers of Frost). Every time you get hit by melee and Molten Armor procs, it is considered a spell cast, and Nibelung can proc from it. Molten Armor also has a guaranteed reflection. I've been hit multiple times and Molten Armor has reflected in such a way that it has caused two Val'kyrs to proc from what would normally seem like one hit (though you cannot get multiple Val'kyr procs per cast). Two points in Molten Shields extends this to ranged attacks and spells that hit you. As such, I see a lot of Val'kyr procs in PvP settings.

Nibelung is definitely a bad choice for Blood Princes and Lich King. They end up attacking the wrong targets routinely, and unless you are a Frost Mage with Water Elemental out, you cannot give pet commands to the Val'kyr (tried it, with multiple macros).

They have a vastly shorter duration on Blood-Queen, due to the AoE pulse (they will appear to heal from it, but they cannot heal faster than the incoming damage, so eventually they will die). They can easily die in the flames (frost damage) on Lord Marrowgar (which most often happens in Bonestorm phase, since flames shouldn't be a problem any other time if you're standing close enough). They can prematurely die on Festergut (Pungent Blight, and I'm not completely sure, but I think Vile Gas can affect them if it hits you). They can die on Rotface pretty easily to the Slime Spray (because they won't follow you out of it). They can die to a number of things on Putricide.

The problem is that Nibelung requires a set amount of procs to ensure that its DPS is competitive with other 264 items/combinations. I ran some numbers in Rawr and my findings were that, compared to [Abracadaver] and [Sister Svalna's Aether Staff], [Nibelung] needs to proc 7.14 Val'kyrs on a 5 minute fight to be on par with it. Compared to [Mag'hari Chieftain's Staff], it needs to proc 5.24 Val'kyrs. Compared to Illumination, it needs to proc ~3 Val'kyrs.

Numbers on meters may be a bit misinforming. The easiest way to look at Nibelung, for instance, is to look at the percent of damage that the Val'kyrs are doing compared to overall damage. Yet this is relative to how much damage each caster is doing. Naturally, Rawr has me doing much more DPS than I ever would be doing, so in Rawr, the Val'kyr damage percentage is a lot lower (around 3-4%) than what it usually is (5-10%). That doesn't mean it's a great staff though. It's nice for the spellpower (especially the 277), and it's great in PvP. It's also really fun on trash pulls... but I don't think it's a great investment for anything else. And since it's effectively broken on Lich King (Val'kyrs routinely were doing ~65% of their damage to Ghouls), it doesn't mean that much to me in PvE anymore.

Enthorn - I am not passionate about Nibelung. I use it now simply because I am restarting my raiding now, and I do not have a better option.

That being said, it does not seem as bad as you make out. Why does 5-10% of your damage not make it a great staff? Mistargeting is a non-issue on most fights - you can always swap it out for LK and Princes. If you replaced the proc with a special effect of "increases your damage by 5-10%" you would use it for sure. Even 3% puts it on par with 100 haste + 100 crit in real terms. Basically, for any fight where I see it posting 3%+ without heavy mistargeting issues, I will call it "good". Make it 5%+ and I'll call it "great". Make it 10% and I will never use anything else for that fight.

Anyway, whether the item is normally "good" or "bad" I will still try to find ways to abuse it.

For people paying EP or DKP for the staff, the option to swap it out for a few fights isn't very pleasant because you end up paying for two weapons instead of just one, reducing your ability to get proper upgrades for other slots.

The "cool factor" is very high and there are cases where it does superbly well (I like it especially on Valithria Dreamwalker). It has a high drop chance and it drops early in the 25 man instance, so it's easily accessible to people who desperately need weapon upgrades. Because of these factors, I think Blizzard has done a good job on it. If it dropped from the LK, it would probably be seen as a disappointing weapon (despite the coolness).

Of all mage specs, it suits frost best. Frost probably gets the highest benefit from AOE and you actually have full control of what the Val'kyr attack, because you have a permanent pet bar. The control you get on the pets opens up some new possibilities (like intentionally having the val'kyr offtank something or break out things like web wraps while you continue AOE). Frost mages are probably also the most likely to be the beggars in the raid, desperately trying to get a decent weapon.

I don't think you will get an extra proc chance from improved scorch. I certainly don't get a higher proc chance on frostbolts from winter's chill. The thing about scorch spec though is that the cast time is extremely short, so this alone will increase the proc chance. I guess one could say that if you insist on playing scorch spec (for whatever reason), Nibelung might actually be a pretty nice weapon.

With gems, you gain 122 crit (2.66%), and 145 haste (4.42%) with [Illumination], and 126 Spell Power with [Nibelung]. Illumination comes in at 1594 DPS. Without a proc, Nibelung is at 1302 DPS. Even the 277 Nibelung, without its proc, is at 1521 DPS. The massive loss of passive crit/haste, even when compared to enormous spell power gains (236 SP for the 277), doesn't offset a 245 staff, and clearly doesn't offset a 264 staff or 264 main/offhand combination. Requiring Val'kyrs to proc X number of times just to see Nibelung be on par with other weapons isn't something I feel comfortable leaving to chance. The 2% proc rate and the frailty of the Val'kyrs (despite healing for 3,150 every ~1.5 sec) means that even when one does proc, it's not guaranteed to be able to DPS for its full length.

It is a very cool staff, and it does work a lot better with Frost, for all the reasons TigaFin mentioned. I notice it procs a lot more with my Frostfires in PvP than it does with Fireball in PvE, and I reckon this is because of the DoT on Frostfire Bolt. Likely this would be the case for Fireball if the glyph didn't remove the DoT. However, I'm not entirely sure it's the DoT causing it. I run 2/3 Frostbite and it's entirely possible that the freeze proc is counting as a spell and proccing Val'kyrs. This would affect Frostbolt then too (as does Fingers of Frost, and this is also why Val'kyrs can proc from each wave of Blizzard, whereas Living Bomb DoT ticks cannot proc it). Remember: it's the spell cast that summons the Val'kyrs, not the spell damage. You can get a Val'kyr proc just from putting Living Bomb on something (which doesn't do instant damage).

Likewise, Nibelung can proc from the Living Bomb explosion, and I believe this is because the explosion at the end is considered a separate spell from the DoT.

I've also heard that Frost benefits more from Nibelung because Frost, supposedly, benefits less from Haste/Crit than other classes. I haven't verified this in Rawr myself, but if it interests you, you may want to take a look at that, especially considering Frost's upcoming changes in 3.3.3.

I think we are violently agreeing. 3% valks dps makes up this different nicely, 5% more than makes it up, and 10% is crazy good.

Anyway, I am still not advocating this weapon as being superb. I am using it as my best current option, and will keep it for some fights once I have better.

I was also not advocating a Scorch spec with it, except as a novelty. Because I play a troll it is not a novelty that is useful to me personally anyway.

Val'kyr damage though is relevant to your own damage (and fight length/DPS) though. Think about it: if you do 100 DPS and 1 Val'kyr procs, they'll have like 99% damage. Regardless, I find Nibelung to be especially good on Deathbringer Saurfang, and, as TigaFin mentioned, Dreamwalker (which I do as Frostfire -- lots of instant spells -- Fire Blast, Dragon's Breath, Blast Wave, Living Bomb, and instant Flamestrikes after DB/BW). It's related to the same concept as Scorch spamming -- the more instant spells you have, the more spell casts you have in general, which means more spells from which Nibelung can proc.

After a few weeks of using Nibelung on all fights, I've gone back to only using it on the two that I mentioned above, and trash pulls (where Val'kyr damage can easily hit ~20% of your total damage, and your total DPS using LB/Flamestrike spam ~23-25k depending on number of targets).