Author Topic: Personal log - Learning the fury spec  (Read 365 times)

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Personal log - Learning the fury spec
« on: October 13, 2016, 11:39:51 AM »
I'm making this log (or guide, however you want to see it) as an exercise to collect my own thoughts and findings on the fury spec. After 10 years of healing raids and pvp, and dabbling in tanking for a few years, it's the first time I'm picking up DPS in casual earnest. I really enjoy the Fury spec at the moment, simply because I love the brutal and active playstyle.

1. Some groundwork

Just to get something solid, I made a spreadsheet with the damage for each ability, with respect to weapon damage and normalized per global cooldown spent to generate this damage.

This yields abstract numbers that should be valid regardless of level, weapon, haste, crit or gear in general. The utility of abilities (which are affected by gear) are handled later.
The normalization brings Bladestorm in line and shows Rampage's damage potential without bothering with the 2s spread of it.

https://docs.google.com/spreadsheets/d/1OrFvidOMT5mXrYokdk_4Ay7KI9umfNhA6qGLnV1fyn8/edit?usp=sharing

I took weapon damage as an estimation of attack power, which is bad. But since it only matters for Odyn's Fury and Dragon Roar I couldn't be bothered.

1.1 A quick overview.

We can do like Icy Veins, and just list our abilities, in descending order of damage per target:
  • Odyn's Fury, on 45s cooldown
  • Execute, only under 20%
  • Rampage, only when you have enough rage
  • Raging Blow, only with Inner Rage
  • Whirlwind, only on Wrecking Ball procs
  • Bladestorm, on a 90s cooldown
  • Dragon Roar, on a 25s
  • Bloodthirst
  • Raging Blow, only when enraged
  • Whirlwind
  • Furious Slash


There! All done! Right?

There are so many cooldowns, proc-based abilities and synergies that dominate the playstyle of the Fury Spec: it seems a collection of short micro-managed rotations, depending on the situation and availability of buffs.

A few notes:
- Execution is so powerful it will dictate a new rotation for targets under 20% to maximize its potential.
- Rampage gets a high priority but its rage consumption permits some wiggle room on when it should be used to maximize its utility (enrage etc).




2. Rotations

2.1. Single Target, basic rotation

Assuming no particular talents, cooldowns or procs we get:
> Bloodthirst, Furious Slash, Furious Slash, Repeat!

This is our filler and weakest rotation. We try to replace Furious Slash with procs or interrupt the cycle for abilities as often as we can, but only AFTER bloodthirst. For example: if Wrecking Ball procs on the last Furious Slash, Bloodthirst would be next in line. You still give priority to Bloodthirst. Wrecking Ball can wait until Furious Slash comes up (but not too long). The goal is to replace Furious Slash as often as possible.
This gives a certain rythm, with bloodthirst punching the beat once every 3 GCDs.

An argument could be made to use Whirlwind instead of Furious Slash. I'll dig deeper on their comparison, later.

Should Bloodthirst proc Enrage then we get:
> Bloodthirst(Enrage), Raging Blow, Raging Blow

If we have enough Rage for Rampage, but aren't rage capped, then we can hold off casting it, finishing any active Enrage rotation first. Then the Rampage rotation kicks in:

> Rampage, Raging Blow, Raging Blow


2.2. Single Target, Inner Rage

Inner Rage complete changes the playstyle of Raging Blow, and puts it on a 4.5s cooldown, putting it right next to Bloodthirst.
> Raging Blow, Blood thirst, Furious Slash

Furious Slash can then be replaced by any proc or better alternative when it becomes available, such as Wrecking Ball or Rampage.



2.3 Multiple Targets, basic rotation


> Bloodthirst, Whirlwind, Whirlwind
Until you have enough rage to use Rampage:
> Whirlwind, Rampage, Whirlwind, Bloodthirst

At 5 targets, a straight Whirlwind (and Rampage) spam seems to be the best answer.

More refinement will follow.

« Last Edit: October 13, 2016, 11:42:42 AM by Zusterke »

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Re: Personal log - Learning the fury spec
« Reply #1 on: October 13, 2016, 11:42:05 AM »
Cast Raging Blow or Bloodthirst first?

When we have Inner Fire, we have to chose between:
1) Bloodthirst, Raging Blow, Furious Slash
2) Bloodthirst, Furious Slash, Raging Blow (aka RB, BT, FS)

Although the difference looks subtle, I recommend the latter sequence. Whenever you fire up a buff instead of Furious Slash, this ensures the biggest chance to have it up when Raging Blow is used or that you can get a 2nd Raging Blow within its duration if you have the Haste for it.

Even a simple Enrage can show the difference. Let's assume BT does not proc enrage, then you can follow it up with Rampage to force an enrage for Raging Blow. It's a bit easier to show with an example.

Sequence (1), Raging Blow before Furious Slash, the first Bloodthirst did not proc enrage.
- BT, RB, Rampage (procced), BT (enraged), RB (enraged), Furious Slash (enraged), BT...
Assuming the 2nd BT procs enrage:
- BT, RB, Rampage (procced), BT (procced), RB (enraged), Furious Slash (enraged), BT (enraged)...

Sequence (2), Furious Slash before Raging Blow, the first Bloodthirst did not proc enrage.
- BT, Rampage (procced), RB (enraged), BT (enraged), Furious Slash (enraged), RB, BT...
Assuming the 2nd BT procs enrage:
- BT, Rampage (procced), RB (enraged), BT (procced), Furious Slash (enraged), RB (enraged), BT (enraged)...

Worst case, both sequences score the same, but sometimes you'll gain 1 Raging Blow during enrage with the second rotation compared to the first rotation.

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Re: Personal log - Learning the fury spec
« Reply #2 on: October 13, 2016, 12:49:54 PM »
Furious Slash vs Whirlwind, without Inner Rage

When fighting a single target, Whirlwind does more dps than Furious Slash. Furious Slash offers more utility however.
The two competing rotations are:
> Bloodthirst, Furious Slash, Furious Slash, doing 527% weapon dmg.
vs
> Bloodthirst, Whirlwind, Whirlwind, doing 648% weapon dmg.

Should Bloodthirst proc an enrage, then this grants the following rotation:
> Bloodthirst, Rampage, Rampage, doing 702% weapon dmg increased by enrage


Blow for blow, Whirlwind wins to Furious Slash, but the Furious Slash increases crit chance on Bloodthirst by 30% and enrage uptime by 30%:
- 527% + 30%(243%) + 30%(702%-527%) = 652%
Furious Slash wins, and this is a pessimistic estimate:
- It doesn't include the damage benefit of enrage, only the raging blow rotation
- The increased use of Raging Blow and enrage should increase Rage generation, offering more Rampage rotations.

Furious Slash will this win by a margin.
« Last Edit: October 13, 2016, 01:10:35 PM by Zusterke »

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Re: Personal log - Learning the fury spec
« Reply #3 on: October 14, 2016, 05:25:17 PM »
Furious Slash vs Whirlwind, with Inner Rage

With the Inner Rage talent, triggering enrage does not change your rotation (unless you use cooldowns or special abilities).
Can the utility of Furious Slash still compete with Whirlwind?

The two competing rotations are:
> Bloodthirst, Furious Slash, Raging Blow, doing 958.75% weapon dmg.
vs
> Bloodthirst, Whirlwind, Raging Blow, doing 1019.25% weapon dmg.

Should Bloodthirst proc an enrage, the rotation does not change. Unfortunately, this doesn't necessarily make the comparison easier.
Every simplification of the problem that I computed shows the whirlwind scenario wins by a small margin. These simplifications usually played to the benefit of the whirlwind scenario.

Trying to logically deduce the outcome, you can see that:
- The Furious Slash provides more enrage time, so it should scale better with Mastery than the Whirlwind scenario
- Furious Slash is a little less crit dependent
- Whirlwind has a higher base damage, and so should temper the bigger scaling of Furious Slash with Mastery and crit

More complex calculations show their comparison to a reasonable amount of detail:
https://docs.google.com/spreadsheets/d/1OrFvidOMT5mXrYokdk_4Ay7KI9umfNhA6qGLnV1fyn8/edit?usp=sharing

The outcome is difficult to read, unfortunately:
- For low levels of Mastery (under 20%), the benefit of enrage is too low and Whirlwind beats Furious Slash.
- For high levels of Crit, the benefit of Furious Slash is "drowned" under the base crit levels. For this, your crit generally needs to be 10%-20% higher than your mastery.
- The numbers don't take into account the benefit of enrage on auto-attacks, which yields more damage and rage
- The numbers don't take into account enrage procs by Rampage. Furious Slash should help to get more Rampages, but then Rampage would increase the base enrage levels, which greater benefits the whirlwind rotation.
- It's possible to boost Furious Slash with traits, but that trait is probably very low on people's priority lists.

Personally, I think the differences are too small... It's a choice and a preference decision. In theory, Furious Slash should have a bigger yield through its utility. However, the practicality of Whirlwind's damage might benefit a greater variety of situations.

If you want to kill a button on your action bar, you could do without Furious Slash. Perhaps that alone makes it worth considering.
But if you like a little more variety, Furious Slash would be recommended.
« Last Edit: October 24, 2016, 11:58:38 AM by Zusterke »

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Re: Personal log - Learning the fury spec
« Reply #4 on: October 18, 2016, 12:26:55 PM »
I'm going to add a note to the previous section. Even with Inner Rage, I would recommend to use Furious Slash against a single target over Whirlwind under "normal" circumstances. For some reason this just "feels" better, although I haven't had a good objective measure for it just yet.

The higher enrage uptimes seems to generate considerably more rage, which wasn't accounted for in the calculations above. This permits me to switch to Rampage more often, further increasing the rage generation and countering the weakness of a Furious Slash attack.

But knowing that the difference is so small, it's safe to fall back to Whirlwind in any situation where:
- There's even a remote chance of hitting a second target
- You plan to burst soon with Rampage anyway
- You're unsure of your range (which seems higher on WW)
- You're still picking a target...

In fact, it almost feels like WW is a good default unless you have a controlled single target situation, except that this isn't a rare situation :)

If I can be bothered, I might do a few training dummy runs with both cycles but I reckon a long testing time will be required to even out the proc-sensitive performance.
It's an amusing contradiction, in a way. The Whirlwind cycle doesn't count on enrage procs so much to prove its value, but might just be more sensitive because it has such a lower proc chance. Furious Slash improves its proc chances considerably, making them slightly more reliable... but also depends on it.

Zusterke

  • Sister of the light
  • Administrator
  • No Lifers
  • *****
  • Posts: 4459
    • View Profile
    • http://www.orderoftheathanor.eu
Re: Personal log - Learning the fury spec
« Reply #5 on: October 18, 2016, 04:54:02 PM »

Haste caps

Haste is a special stat for Fury warriors. It's promoted as the #1 stat by many guides, and it made me wonder why. I always saw Haste as either a technical stat, or as a secondary stat that only empowered others. So in an attempt to understand what's going on, I investigated a few of the feats of Haste.
  • Haste increases attack speed. This improves your rage generated per second. I was sceptical about this benefit since it boosts attack speed similarly to cast speed. If you could auto-hit twice in a 3 GCD cycle before haste, you still auto-hit twice in 3 GCD cycles after haste. The GCD scales down with haste, so it's all going quicker, but it's proportionally the same.
  • Haste reduces the GCD. Since most of our abilities are instant, this boosts the number of abilities we can use within a long time-frame, directly proportional to the amount of haste. For small time-frames, you get discrete intervals with the number of GCDs within that timeframe.
  • Haste decreases cooldowns proportionally to the GCD. This is important to understanding rotations. A rotation of 3 abilities where one or two have a short cooldown (ie: Raging Blow, Bloodthirst, Furious Slash) does not change with Haste. The cooldowns scale along, keeping the rotation.
  • Haste doesn't reduce the buff time. And this is the big benefit... A short buff, something we have plenty off, lasts as long regardless of haste. As a result, haste can help us increase the number of abilities cast during a buff timer.


Here's the list of caps for any number of hits within a buff frame.
Avatar, 15 hits5.0%
Bloodbath, 8 hits5.0%
Enrage, 4 hits12.5%
Avatar, 16 hits12.5%
Battle Cry, 5 hits20.0%
Avatar, 17 hits20.0%
Bloodbath, 9 hits20.0%
Dragon Roar, 6 hits25.0%
Avatar, 18 hits27.5%
Avatar, 19 hits35.0%
Bloodbath, 10 hits35.0%
Avatar, 20 hits42.5%
Enrage, 5 hits50.0%
Battle Cry, 6 hits50.0%
Dragon Roar, 7 hits50.0%
Avatar, 21 hits50.0%
Bloodbath, 11 hits50.0%
Dragon Roar, 8 hits75.0%
Battle Cry, 7 hits80.0%
Enrage, 6 hits87.5%

Depending on talent choices we can distinguish a few important cap points:
  • 12.5% haste. We get an extra attack during enrage (very important) and avatar, if you specced it. I'd say this is a minimum to make the spec functional.
  • 20% give us a boost on Battle Cry which is very important, and then the talents Avatar and Bl oodbath if you have them.
  • 25% is the real number here, giving an extra attack during Dragon Roar. The extra 5% also gives some room to move for Battle Cry etc. Because of the short cooldown on Dragon Roar, you're bound to feel this difference.
  • 50% which is a big jump. It's no surprise that 50% stands out. It effectively reduces the GCD to 1 second, and all buffs last an exact number of seconds. When slightly over 50%, all buffs grant 1 extra instant within the buff timer.

After 50% we see it becomes increasingly expensive to get additional abilities within the buff timeframe. Haste will still reduce the cooldown, making it possible to cast the same buff more often during the same fight.