Haste Effects will *reduce* our DPS.

#0 - March 2, 2007, 9 a.m.
Blizzard Post
This is my chief complaint with the Hidden Windfury Cooldown Mechanic.

I don't mind that the offhand would share the same timer, I'll use flametongue or something. At least the spell damage on my t4 gear will do something now. My complaint is that the mechanic is seriously flawed, and makes buffing effects actually detrimental. This affects all shaman who use windfury, not just the Enhancement shaman. I'll explain.

First, let me establish how windfury scales with weapon speed. If not for the 3.0 second cooldown, Windfury would be in identical dps increase across all weapon speeds. However, because of the cooldown, windfury favors slower weapons. The rule of thumb was that "fast weapons would have lots of small procs, while big weapons would have less, large procs". This is not true because the "lots of small procs" that a fast weapon should have are limited by the Hidden Windfury Cooldown Mechanic.

Let's assume that the Hidden Windfury Cooldown Mechanic is exactly a 3.00 second cooldown.
In this scenario, I'll hypothetically use a 1.5 speed weapon.

1.5 speed weapon
Time - Effect
0.00 - Weapon Swing - Windfury Proc
1.50 - Weapon Swing - Windfury cannot proc.
3.00 - Weapon Swing - Chance for Windfury Proc

Now look at the same situation with a 3.0 speed weapon.

3.0 speed weapon
Time - Effect
0.00 - Weapon Swing - Windfury Proc
3.00 - Weapon Swing - Chance for Windfury Proc

Now, that's add +1 haste rating to the equation.

2.99 speed weapon
Time - Effect
0.00 - Weapon Swing - Windfury Proc
2.99 - Weapon Swing - Windfury Cannot Proc
5.98 - Weapon Swing - Chance for Windfury Proc

See what happened? The 1 point of haste rating turned the 3.0 second windfury cooldown into a 5.98 second windfury cooldown!

Let me attach some numbers, to make this seem more quantative. Lets use a 1.5 speed weapon that hits for 250, and has Windfury Attacks for 400. We'll use a 3.0 speed weapon of the same dps, that hits for 500 and has Windfury Attacks for 800.

1.5 speed weapon
Time - Effect
0.00 - Weapon Swing - 250, 400, 400
1.50 - Weapon Swing - 250
3.00 - Weapon Swing - 250, 400, 400
4.50 - Weapon Swing - 250
6.00 - Weapon Swing - 250, 400, 400

Total - 3650


3.0 speed weapon
Time - Effect
0.00 - Weapon Swing - 500, 800, 800
3.00 - Weapon Swing - 500, 800, 800
6.00 - Weapon Swing - 500, 800, 800

Total - 6300


2.99 speed weapon
Time - Effect
0.00 - Weapon Swing - 500, 800, 800
2.99 - Weapon Swing - 500
5.98 - Weapon Swing - 500, 800, 800
6.00 - Technically, the time passed between 5.98 and 6.00 is a percentage of a swing cooldown, which is approx 3 damage. (0.006 * 500 = 3)

Total - 4703

See what a dramatic difference that is? 1 point of haste rating can significantly lower my dps. Keep in mind that many shaman pay talent points for an innate haste effect: Flurry, which could be actually lowering the shamans DPS, rather than improving it.

For simplicity of math, I assumed 100% windfury procs. Obviously, Windfury only has a 20% proc rate, so the actual dps lost by 1 haste rating would be less pronounced than what you see here. Nonetheless, the loss would be significant, and I feel that this should be a major point of investigation by the development team. Haste is one of the most expensive stats in terms of itemization, and obviously it is a mechanic that is supposed to increase damage output, not reduce it. Coupled with the fact that Flurry procs could actually reduce my dps, rather than increasing it, this issue should be addressed as quickly as possible.

Honestly, all we really would like to see is a blue response saying that this issue has been noted, and will be looked into. I'm fairly sure this is just an unintended side effect of the implementation of the Hidden Windfury Cooldown Mechanic.

Please keep this thread bumped so Tseric and other CM's can add it to their meeting notes.
Cheers!

Az
#22 - March 2, 2007, 11:05 a.m.
Blizzard Post
Let me just say outright, I would love for you to prove me wrong on this contention. I see an issue with a core statement made in the OP for a couple reasons.
Q u o t e:
If not for the 3.0 second cooldown, Windfury would be in identical dps increase across all weapon speeds.

If you mean that windfury adds the same amount of damage for each extra attack, that is one thing. To say the potential damage output of a fast weapon from a slow weapon with windfury is another. The OP focuses on how the 3 second cooldown affects proc'ing on slow and fast weapons, but doesn't really deal with what happens if the cooldown isn't there. Especially in a dual wield situation.

Now granted, I haven't crunched numbers on this either, but I'm interested to see simple math where the max damage from a repeated crit from both a slow 2 hander and dual wield, fast main and off hand compare within a 6 second timeframe.

Let's take as a given that all weapons proc on strike. That's what unmanagable burst damage is all about. The fast, dual wield weapons are procing every strike in 6 seconds, the slow 2 hander procs every strike in 6 seconds. How do these compare? Does that support everything that follows in the original argument?

I ask, because I don't know what the hot enhancement Shaman weapons are, fast or slow, otherwise I'd do some math myself. :P
#24 - March 2, 2007, 11:12 a.m.
Blizzard Post
Q u o t e:
Actually, let me dig up a little thing I wrote about just this very question.

Please do. I am interested.
#28 - March 2, 2007, 11:25 a.m.
Blizzard Post
Q u o t e:
Pardon me if the timeline is off, I will try and format it correctly.

This is part of "Practical DPS" - I know you might be trying to think about theoretical damage more, but really, nothing just sits there and lets you whack at it.

If this doesn't shed a little bit of light on shaman weapon speed choices, I have another small write-up on 2h vs DW and why we're being pressured to go 2h.

Hope this helps.



Compression

Here's an extremely simple example of something I was talking to Xiani(a guild mate) about before called "compression". I was running an instance so I could not tell him in depth what it was, but here is the illustration.

Why Slow Weapons Are Better For DPS: ((Especially for Shaman)) It's Not *Just* About Instant Attacks

Here are my 2 base example weapons:

Weapon 1 - 100 dps, 100 dmg - 100 dmg 1.0 attack speed

Weapon 2 - 100 dps, 200 dmg - 200 dmg 2.0 attack speed

Here is a timeline of these weapons in action in a boss fight like Chrommagus



Time Index 0.0-----1.0-----2.0-----3.0-----4.0-----5.0
Weapon 2 200----200----400----400----600----600 (Total Dmg)
Weapon 1 100----200----300----400----500----600



As you can see, because of the initial strike, a slower weapon starts out with a dmg advantage over a faster weapon. It takes the faster weapon 1 second to catch up to the slow weapon, and then a second later it is outpaced by the slower weapon; it catches up again.

As you can see, both weapons do the same damage over 5 seconds, they both have the same dps.

The problem is, on fights where you must move in and out of aoe/line of sight/cleave range/ go kill adds etc. the slower weapon is often AHEAD in total damage at various times in the fight. 4 seconds into the fight, if dps had to go hide, the slower weapon would be 100 total dmg ahead of the faster weapon; and upon going back into melee, the slower weapon would be another 100 points ahead of the faster weapon. 100 total damage ahead and 100 temporary dmg ahead.

This is called compression.

As we all know, instant attacks tend to favor slow weapons over fast weapons. This is yet another reason to choose slower weapons over faster weapons. There are very few reasons to choose faster weapons - generally procs and "on next swing" abilities like Heroic Strike which add static dmg.

Thank you for getting that. I wanted to throw out some arbitrary numbers like that, but I felt it would be poor form in my position. ;)

I still feel this doesn't address dual wielding and the procs from that, in a given timeline. You're still dealing with single weapons on a single timeframe. It doesn't touch on having dual-wielding weapons proc one after the other, or does it?

I mean, situational line-of-sight is just another factor you could throw in, but at the base math of it, does slow hand proc'ing match dual wield proc'ing?
#31 - March 2, 2007, 11:27 a.m.
Blizzard Post
Ok, I guess I'm just looking at it from the context of yesterday, where enhancement dual wield Shaman were the ones affected by this particular change more than anyone else. Just trying to establish first cause before moving into how spell haste 'nerfs' the whole business.
#85 - March 2, 2007, 12:52 p.m.
Blizzard Post
Q u o t e:
I know I'm not offering much to this thread, but I just want to thank Tseric for actually trying to understand what's going on here instead of just writing it down and passing it to the devs.

Sometimes just inciting a discussion and seeing the results later can make for a good thread. Flagging a thread blue has a lot of wierd, psychological incentives and this thread is full of good material for digestion. Whether it all goes as direct feedback is rather secondary.

But, I might just forward this one in full. I do think the OP has an interesting point, even though the devs might consider it intended.
#87 - March 2, 2007, 12:58 p.m.
Blizzard Post
Q u o t e:
As such, I'm off to bigger and trollier threads.

Word. See you there. :P