Appendix C: Haste and the Butterfly Effect

Why did all Warlock specs become so DoT heavy in Cataclysm?

I’ve been wondering this as I’ve gone though the changes in Mists, because the difference is so startling; why did every spec start using all DoTs, all the time? Why did Destro end up getting damage from 5 DoTs and lose the nuking feel? Why did everyone end up using Corruption and Shadowflame?

Was it a deliberate design decision, or was it an accident?

I can’t answer designer intent, but I think the root of this was an innocuous quality of life change made to DoTs at the start of Cataclysm. Ironically, a small change to make DoTs easier to use ended up making the Warlock class harder to play overall.

I covered this early after Cataclysm’s release in How Warlock DoTs Work in Cataclysm, which looked at the different ways Haste affected DoTs in Wrath and Cataclysm. There were two changes:

  • You could now clip most DoTs without losing the final tick, and
  • DoTs would gain ticks of damage at certain Haste values.

The clipping was the quality of life issue, but to make it happen, the mechanics of DoTs had to be modified.

In Wrath, Haste meant that your 18 second Corruption with 6 ticks could be done in, say, 12 seconds, at which point you’d recast it. In Cataclysm this was changed so that DoT spells would remain roughly the same duration, but add additional ticks of damage when they got fast enough. Your 15 second/6 tick Corruption could potentially become a 13.64/6 tick or 16.25 second/8 tick spell.

This was, in retrospect, a really big deal.

Haste increased DoT DPS by making ticks faster, that’s easy. But in Cataclysm, Haste also dramatically increased the DPCT (damage per cast time) of DoTs like Corruption because each cast generated more overall damage. If a Warlock with no Haste cast Corruption, that’s a GCD to cause 6 ticks of a set amount of damage. If that Warlock has enough Haste to get an additional tick or two of damage, that GCD spent casting the DoT now does 16%, 33%, 50% more damage, depending on how many additional ticks were on the spell.

With Cata’s new Haste/DoT model, Warlocks of all specs had to cast every DoT available to them. They were too good not to cast. Corruption and Shadowflame became defaults for every spec. Destro was caught in a vice because Conflagrate was based upon Immolate’s total damage, so every additional tick Haste granted increased not only their primary DoT’s DPCT, it also made their primary nuke hit much harder, giving an additional DPS boost. So they’d stack Haste to get those Immolate ticks, which would in turn improve Corruption’s DPCT, making it even more important to cast.

Sweet delicious Haste, combined with shared spells across all specs, became a trap.
I think the changes weren’t intentionally designed to make Warlocks more complicated; changes were made to DoTs/HoTs for all casters, not just Warlocks, and in general the new model really is better. It’s certainly easier to be able to clip with the 2 second rule! And the extra ticks and associated Haste Breakpoints/Plateaus make Haste a more interesting secondary stat.

I think this points towards Warlocks getting hit by Lorenz’s butterfly effect – a small initial change in one part of the game had unforeseen major consequences in another.

  • DoT/Haste interaction changes with 4.0.1 release.
  • Corruption/Shadowflame DPCT rises.
  • All Warlock specs have Corruption and Shadowflame in their priority.

Plenty of other small quality of life changes were made that had an effect upon the class’s overall complexity. Banes being separated out from Curses, for example, seemed good at first (Banes caused damage, Curses were debuffs) because you no longer needed to do advanced math to figure out if you should take the 12% damage increase from Curse of the Elements or the damage caused by Curse of Doom or Agony. But then every Warlock ended up casting a Bane – and maybe a Curse too.

The DPCT of Shadowflame has always been surprisingly good, mostly due to its short cast time, so it was interesting to see it get used so prominently in Cataclysm. I haven’t really taled about Shadowflame much before, but it’s a spell that I adore – damage, dot, and a glyphed slow. It’s also a tremendous pain in the ass to use in PvE, and takes skill to use in PvP. It has short range so you either have to run in to use it or forgo it entirely. It’s not just a DoT or an AoE, it’s a spell that requires very good placement to use. Movement matters, which adds a layer of complexity far beyond just another spell.

Corruption presents an interesting problem because it was baseline damage for all three specs. Blizzard couldn’t change how it worked without considerable effort, so all that was left was tweaking the damage. If the damage was made too low, Affliction would be neutered, but if it was kept high, Destruction would benefit too much. So talents, Mastery bonuses, general damage all got modified – but Corruption’s DPCT was too good to pass up for Destro. (Demo always wanted it for Molten Core procs, but the damage was similarly good.)

I think the fact that Haste changed all DoTs and not just Warlock DoTs points towards this being an accident. I can’t prove it, and it’s not like it was the source of all Warlock problems in Cataclysm.

It was just a little butterfly, flapping its wings.

9 Comments

Filed under Cynwise's Warcraft Manual, Warlockery

9 responses to “Appendix C: Haste and the Butterfly Effect

  1. imsek

    I blame the Onyxia guy. “I want to see lots of dots, Moar DOTS!”. And Blizz listened. But I’m a resto druid, and I put some dots in if its not a heal intense phase. (example, blowing the healing tendril things on Spine) MF FF IS

  2. Mrigashirsha

    This rings right – unintended consequence is such a big deal in real life. And such a good argument for a spec design with fewer wildly oscillating fiddly bits, darn it.

  3. Wulfstan

    I vaguely recall Blue posts on this topic early in Cata….

    I think there was a design statement that Locks were a DOT class, and so they were designing for DOTs to be used across all specs, including Corruption for Destro.

    I also recall a lot of fuss in Cata beta about Shadowflame being useless, and thus it was heavily buffed.

    So I think the impact of having to use Shadowflame in PvE might have been mis-calculated by Blizz, but I think the rest of the design was intentional rather than accidental.

    • I’d love finding those blue posts, if you remember them? I went looking but didn’t see anything, but that’s as much because there is a real difficulty finding blue posts from Wrath / Cata Beta period as anything else. A lot of this has been relying on people’s memories and hunting down posts in blue trackers and the like. The forum changeover wiped out a lot of Warcraft’s collective memory. 😦

  4. jagoex

    Great post, Cyn. It’s a pleasure to read your always-sensical perspective on all matters warlock. Thanks for posting.

    Looking at this from a PvP perspective, it’s easy to see why DoTs were buffed, mostly, as you said, via the changes to Haste: DD-burst is huge in Cata, as are player health pools. For DoTs (and DoT classes) to keep up, they had to do roughly the same amount of relative damage that they did back in Wrath, when health pools were much smaller and the separation between DD classes and DoT classes was not as extreme.

    Unfortunately, blanketing changes the way the Haste tweaks did adjusted different DoT classes in different ways, and my chief complaint isn’t necessarily that the warlock class became harder to play (I actually LOVED Destro in Cata and remained so throughout it’s duration), but that individual DoTs/DoT classes, depending upon their individual design (coefficients, Haste %s, etc), received varying amounts of effect from the changes to Haste. Vampiric Touch is a prime example of what happens at the extreme top end – it gains more “umph” faster than most if not all other DoTs in the game (minus Combustion).

    Anyway, I don’t want to come off like I’m QQing about other DoT classes being more powerful – having more powerful DoTs – than the only pure DoT DPS class. I actually enjoyed and welcomed the change, and only hope that in the future, they are accompanied with just a tad more balance and control. And if this was an accidental casualty of the developing process, a bit more effort fixing things up a bit wouldn’t hurt, either.

    • Thanks, Jagoex!

      There’s a subtle argument here which I (to be very honest) lack the experience to make, which is about the number of DoTs used by each class and multi-dotting in any given raid tier. I remember reading several articles from both warlock and shadow priest players which talked about how much easier it was for spriests to mutlidot, and how much more effective they were due to the reasons you mention (spell power coefficient, ticks, general overall power) as well as factors like Shadow Embrace and ISF buffs. I just don’t have the experience with Shadow Priests or Moonkin to make it with any authority. 😦

      I too stand by the mechanical change as a good one. It improved warlock quality of life and made dots more interesting (especially for Destro.) But it did add to the complexity of things in Cata.

      Thank god this is getting straightened out in Mists.

  5. Pingback: Out of the Mists: Reclaiming Warlocks in Pandaria « Cynwise's Warcraft Manual

  6. I really need to go back and finish reading this series, but this came up most recently in my catch-up game, and I have to say… that little change leading to the big issues with the state of Warlockery makes a hell of a lot of sense, Cyn.

    And all this Warlock reading makes me want to play all mine. :S