A few clarifications for this: as an "on next swing" melee strike, Rune strike is not normalized. All of the instant strikes are though (Plague Strike, Blood Strike, Heart Strike, Obliterate, Frost Strike, Scourge Strike). And in fact, checking on Thottbot, Blood-Caked Strike (the strike that procs from Blood-Caked Blade) is normalized. And saying there is no significant difference between a 1.0 speed 2H and a 4.0 speed 2H for a DK is 100% incorrect. For a 200 DPS 2H, the .2 second difference in weapon speed between even a 3.4 and a 3.6 speed 2H means roughly 40 average weapon damage (on a 100% weapon damage attack, scaling with the %age of weapon damage an attack uses). Adding 40 to your normalized weapon damage is like adding the equivalent contribution from ~165 attack power on each instant strike.

This isn't quite correct. It prevents slower weapons from double-dipping on instant attacks by having both a larger swing damage and gaining a larger AP multiplier for it's speed. Normalization refers to having a set AP multiplier, however you still benefit from having a larger base damage on slower weapons.

Non-normalized instant attacks use this formula: damage = base_weapon_damage + (weapon_speed * Attack Power / 14)

while normalized use this one: normalized_damage = base_weapon_damage + (X * Attack Power / 14)

where X is:

* 1.7 for daggers
* 2.4 for other one-handed weapons
* 3.3 for two-handed weapons
* 2.8 for ranged weapons

(courtesy wowwiki for the exact numbers; I couldn't remember them off the top of my head)

So, all other things equal, even for normalized attacks, a slower weapon will still be better for us; just not as much as it used to be.

I'm a mechanics noob I'll admit, but I'm trying to understand how the haste making Demise swing at 3.4 makes it a worse weapon choice. It doesn't actually change the weapon damage does it? With our active strikes being normalized does that matter really? Wouldn't our active strikes do the same damage as they would if it wasn't hasted? Or is there some other reason?

Ah, I guess I didn't fully understand normalization then. I was under the impression that normalized strikes used your weapon DPS and the fixed multipliers listed above when calculating damage, not actual weapon damage. Thanks for the clarifications. Although I can't imagine why BCB would be normalized, when the "proc" rate is based entirely on weapon speed. The whole ability seems to be messed up, seeing as how it can't even crit. I would have expected it to function more like sword spec than it actually does.

Hi, first time posting, but I've been reading ej for a while, I was very happy to see this thread, as I feel this is a fairly important matter to consider when choosing between weapons. While I know right now it may be difficult to correctly model due to procs(mainly Killing Machine in frost and possibly BCB in unholy to my knowledge), would it be possible instead of giving the % difference based on gear level, to possibly give a static difference like that in the rogue dps compendium.

I believe if we present the information in this matter that it will be easier for someone not quite as good as math to discern which weapon may or may not be better for them.

Like, say for now not factoring in proc talents and assuming stats are equal and whatnot the potential difference in three 200 dps weapons, one 3.4 speed, one 3.5, and one 3.7

The average weapon damage for each would be:

3.4: 680

3.5: 700

3.7: 740

So couldn't we then figure out how much each weapon would contribute to each specs rotation? Like say for my frost spec where I use a rotation that averages 2 oblits, 1 bs, and 2 fs in a 9 second cycle(icy touch would be part of this however since it is not based off of weapon damage I did not feel the need to include it)

3.4: 2x Oblits> 816x2, 1xBS> 340, 2xFS> 408 Total> 2788 over 9 seconds for 309.7

3.5: 840x2, 350, 420x2 > 2870 over 9 seconds for 318.8, difference of 9.1

3.7: 888x2, 370, 444x2 > 3034 over 9 seconds for 337.1, difference of 18.3

Now I know it's pretty simplistic, not factoring in talents and whatnot, but wouldn't it be safe to say that minus the potential effect of talents such as killing machine, but wouldn't it be then safe to say that each .1 of weapon speed that a weapon is slower would be worth an extra ~9 ish dps when compared to a weapon that is faster?

I would be very interested to hear thoughts and ideas on doing a comparison like this for all specs, and if I assumed something incorrectly please tell me so that I can take the information down if it's not true.

Actually, upon further inspection on wowhead, the comparison seems to be better based on average damage rather then weapon speed. Though I'm not sure if it's still something thats important, it seems like 20 average damage is what is worth the 9 dps increase, rather then weapon speed. To the point that some lower dps weapons my be equal or better if their average damage surpasses that of the other weapon in comparison due to speed.

Ok, so when I wrote this post I was referring to AP normalization (not proc rate).

I spent some time looking through wws parses and I learned two things:
1) speed does not effect proc rate (with a beta value of less than .05%)
2) BCB doesn't crit on its own (nor does it profit from higher damage of a white crit)

As a result it makes up a smaller portion of a DKs dps then I originally hypothesized.

So in a nutshell, the only thing that matters about a DK's weapon is the DPS number - unless you're trying to maximize Killing Machine procs, in which case faster weapons are better.

(i.e. all DK attacks are normalized, except rune strike, which is swing speed limited, so it doesn't matter.)

Is that correct?

I want to go ahead and say yes to all of your questions; weapon speed is more of a PvP consideration than a PvE consideration for a DK--however, slower weapons will yield slight damage increases so, with all else held equal, go for the slower weapon if doing so doesn't step on anyone else's toes.

And also, yes on normalized vs not normalized attacks, at least based on my understanding.

Agree that slower = burstier for PvP.

I'm slightly confused - in what situations does a slower weapon actually cause any DPS gain?

As long as it's a 2h weapon, the strike damage will always be 3.3 * ( DPS + AP / 14 ) * Weapon %.

The only time your strikes will change is if you change weapon class, right?

Quick disclaimer: I haven't done anything with normalisation for a long time. This is correct as far as my understanding goes - please correct me if there's something wrong in here.


The formula for each strike will be:
((%Contrib * (3.3 * (AP/14))) + (%Contrib * Weapon Damage))
or simplified
%Contrib * ((3.3*(AP/14)) + Weapon Damage)

%Contrib varies per strike (Plague is 30%, Heart/Frost/Scourge is 60%, Mortal/Sinister is 100%). Due to the Weapon Damage part of the formula, a slower weapon will still give a higher contribution; although it will be minor in the grand scheme of things there will still be a difference.

The formula you've stated shows two sets of Weapon damage contribution; Firstly the DPS of the weapon and secondly the damage range (Weapon %). The DPS of the weapon is never included, as far as I understand it, while the damage range is.

Say we look at [Death's Bite] and [The Jawbone] on Plague Strike. Plague Strike takes 30% of the weapon damage, so:

Deaths Bite: 0.3* ((554+831)/2) = 207.75
The Jawbone: 0.3* ((586+880)/2) = 219.9

Or an increase of approximately 12 damage for 0.2 weapon speed on Plague Strike. No matter how much your AP changes, there will always be this differential between the two weapons on Plague Strike, and it will be higher for higher %Contrib strikes (i.e. 60% on Frost Strike / Heart Strike / Scourge Strike, so the difference will be ~24 damage).


Edit: I'm not entirely sure how this affects BCB, but I would assume it's the same principle with the %contrib varying dependant on how many diseases are in effect.

@Dukes:

Pretty much spot on.

A simpler way to think about it is that, regardless of the attack, your weapon will contribute an average of DPS * Speed. Your AP contribution will always be AP/14 * 3.3 in a normalized, 2H strike. Both the AP and Weapon contributions become what Blizzard calls "Weapon Damage".

So, the AP contribution always stays the same in a normalized strike (if your AP remains constant through the weapon swap) whereas the weapon part changes.

To add to the weirdness, Yoshi, is that not all classes have normalized weapon attacks. Take the Enhancement shaman, for example... Speed makes a marginal difference to us, but a whopping difference to them.

The game is riddled with odd relics of development. That's what keeps these forums poppin'.

Broseph, weapon speed matters to Shaman primarily due to the Windfury Weapon internal cooldown, and not due to any normalization mechanics.

Windfury Weapon by itself is similar to BCB procs - it is unnormalized, but since it is also tied to swing speed, it does not matter: A faster weapon procs WF/BCB more, but deals less damage per proc.

If anything, Paladins are the odd man out with regards to normalization: Judgement of Blood's and Command's weapon damage component is unnormalized, but a cooldown-based instant attack is exactly the kind of ability that normalization is supposed to address.

Seal of Command procs are also unnormalized, and since they are on a PPM system and not a flat percentage chance to proc per swing, a slow weapon holds a definite advantage yet again.

This is not really true.

Stormstrike and Windfury aren't normalized, the fact they have a cooldown isn't the only reason you get slow weapons. They do a ton more damage.
With 2800 AP the difference in AP contribution of a 2.0 weapon vs a 3.6 weapon is: 400 dmg vs 720 dmg
That is pretty huge difference.

Same goes for stormstrike.

BCB is normalized so you actually get more out of a fast weapon, since like you said it procs more. And unlike WF it doesnt have an internal cooldown.
Due to normalization the AP contribution to the damage is the same (which is the brunt of the damage), yet the faster weapon does generate more procs.

----

The only line i see in blizzards reasoning of keeping certain classes without normalization is the amount of spammable attacks.
Normalization was made it seems to prevent classes that have spammable attacks to scale out of control. Just remember when Hemmorhage wasn't normalized and 2.90 speed weapons (like the old high warlord stuff) would do tons and tons of damage.
The only melee classes who didn't have spammable attacks back then were enhancement shamans and retribution paladins (procs and stormstrike).
My guess is that they left it unnormalized because the whole system gradually built around it and changing that now would result in a total overhaul of all the separate damage attacks.

We're really saying the same thing: Normalization was implemented because instant attacks can "cheat" the system; the slower the weapon, the greater the margin of the cheating.

What I tried to point out is that WF isn't normalized because it doesn't need normalizing. It's not an on-demand spammable attack, such that a dagger is going to gain the same amount of DPS relative to a slow mace. Or at least, theoretically, without the internal cooldown.

I was referring to both stormstrike and lava lash, the only two "on-demand" weapon attacks an enhancement shaman has, and neither of which are normalized. They both hit MUCH harder with slower weapons.

Not sure how you read into it that I was talking about WF.

This is a huge digression though. The point is that even though normalization for us works "weird", inasmuch as the weapon's own damage portion isn't normalized, it could be much worse -- as evidenced by the couple specs in the game that are still completely un-normalized and thus tied to weapon speed.

I can't speak to weapon speed, I would personally say go with what you can get. As far as the stats go however Armageddon seems the better choice, to me personally. 101 strength is huge and benefits from multipliers like Ravenous Dead and Shadow of Death + Kings. I'm not sure how you figure Cryptfiend's Bite is 50ap more however. DK's don't get AP from agi unless they have Bladed Armor, and then only 1 ap per 18 armor I believe, which would put Cryptfiend's Bite at 206.6 ap. Still More than 202 base for Armageddon, but thats before scaling.

Cryptfiend's Bite has more crit (~0.8% more), less haste, is slower. In my case its no contest though, since Human's get 5 expertise from Swords.

Humans get 3 expertise to Swords and Maces, not 5, though the point still stands. WoWWiki source

Oh my, am I mistaken? I thought for sure 1 agi = 1 ap for DKs. Maybe that is just stuck in my mind from the rogue days.

Edit: Oh wow, I am wrong. Well scratch the 50 ap bonus, dur... Still question remains. Are polearms viable?

I am interested in a comparison between [The Jawbone] and [Inevitable Defeat].

If we break down the stats in accordance to the following table:

Stat	Unholy	Blood	Frost
1AP	1.0000	1.0000	1.0000
1STR	2.4847	2.5570	2.3650
1CRT	1.0207	1.2805	1.3212
1HIT	2.4631	2.7485	2.9114
1Haste	0.6046	0.6345	0.5619
1Exp	2.1660	2.8481	2.0712
1AP-R	0.5918	1.2693	0.9086
1AGI	0.8150	0.9164	1.0008
1Armor	0.0278	0.0294	0.0319
1dps (wpn)	6.3169	8.4347	9.5283
Spec	17/0/54	51/13/7	17/54/0

... then [Inevitable Defeat] comes up on top with a 44 rating advantage, or the equivalent of 7 additional weapon DPS.

However, [The Jawbone] has 40.5 advantage in average weapon damage, which, it seems, applies to all DK strikes.

The questions is, which one is better for PvE DPS?

It doesn't seem as odd when you factor in the additional design possibilities. Okay, if an item designer figures out the weapon's stats and DPS and then just does a /random roll to determine the weapon speed, it would really be weird.
But that's not exactly how it works. For example, back in pre-naxx vanilla WoW and just from the back of my head, [Sulfuras, Hand of Ragnaros] was the only available epic-and-beyond-level two-handed weapon with a weapon speed of 3.7, thereby adding to its legendary status

In my opinion, the excessive AP values you are able to reach nowadays somewhat diminish the contribution of weapon speed to overall DPS. Though, of course, with a large percentage of our damage being dependant on weapon damage, slower weapons are still strictly superior, all else being equal.
But slower weapons are probably no longer a reason to lose out on an expertise racial bonus or similar effects.

I'm looking for a way to compare weapons for dual wield specced Death Knights. I haven't been able to dig out a thread through the search function, but I was looking at 1handed weapons on wowhead.com and it seems to me that Last Laugh is the best choice? It has the highest weapon dps (171+), good speed (1.6), 37 str and some hit rating. So is this true, or is there a better choice for dual wielding?