A close mentor of mine asked me to explain to him why the MoP Warlock was such a polarizing and dramatic creation – and how it came to be. For the last two expansions, I’ve sat back and observed as changes were made to the class – an inevitable, necessary thing – and by watching, realized which lessons applied to the MoP Warlock survived and which ones were lost in the noise.
While change is inevitable, the core changes that were thrown away all matched a particular theme: That at a fundamental level, the changes did not remain in harmony with the underlying ‘minigame’ that existed for each spec – nor did they pursue a goal of replacing one ‘minigame’ with a new one.
Ultimately, however, I am not the judge nor the audience of a class rework or update any longer. Only you, the players, can decide if the changes are appealing or unappealing. So let’s talk about what existed, how it came to be, and what the spirit of each spec was meant to represent.
Perhaps you’ll even come to realize what it was that enjoyed and miss – or be able to teach me about what they did that was so vital and important that I didn’t understand.
First, though you need to know why this rework came to be – and the risks taken by Blizzard during its creation.
Why did I decide to overhaul the Warlock class? The answer, in a word, is Cynwise.
Cynwise, for more recent players, was a uniquely inspiring blog writer. His ability to deeply understand and convey the problems of the Warlock class were unparalleled and reveal a deeply rooted problem with World of Warcraft’s design – which was that we couldn’t add new skills forever.
At a certain level, most of the design team understood this – but it was the Warlock which was the first class to break the upper limit so severely that the class’ playrate dropped, becoming 42% less popular than other classes from Wrath of the Lich King into Cataclysm.
Cynwise’s brilliant articles on this topic are archived here:
I highly recommend the read for anyone who wants to deeply understand the human relationship between complexity and gameplay.
After reading his articles, it became clear to me that only dramatic and significant effort could restore the class in the eyes of the playerbase. It was too complex, too unwieldy – and too much same across specialization.
When I spoke with the Warlock designer at the time, who had just been assigned the responsibility of creating the Monk class from scratch – it became apparent that he was burned out trying to keep the identities of Shadow Priest and Warlock unique and overwhelmed with work load with Monk.
So I privately approached Ghostcrawler (who is one of the kindest guys you’ll ever meet). Ghostcrawler and I had worked together heavily on the Death Knight – and I asked him if I could take a crack at proposing some changes to the Warlock class for Pandaria. He hesitated for a moment, but said that if the current Warlock designer was OK with the direction of the changes, he would get all of the resources needed to revitalize the class.
The Warlock designer, relieved to have one less thing to worry about, enthusiastically encouraged me to take a shot at it. However, he warned me, the Warlock community was very sensitive to changes and didn’t trust us very much – a minefield that we would need to handle delicately.
With that encouragement and admonition, I planned a path to balance both the needs of the game with the needs of the players
Convening the Council
Prior to Mists of Pandaria, I served Blizzard as a Raid, Dungeon and Encounter designer. This meant that my role was focused on creating opportunities for players to USE tools, rather than create the player’s toolkit themselves.
While, I did a majority of the talent tree work for Death Knight, directed by Ghostcrawler and with the support of Kris Zierhut, I had never done serious class work before. I knew if I tried to navigate these waters alone, I would miss the mark. Rebuilding a class is far more difficult than creating a new one – you have years of expectations and baggage to overcome – and you will inevitably need to remove someone’s favorite feature.
After pulling my manager at the time into a private room, I told him my plan:
I intended to subvert Blizzard’s wall of silence policy – selecting talented and insightful members of the Warlock community, with whom I had built relationships during my Raid design days, and reveal the full scope of the rework plans from day 0. I believed that in exchange for the risk of having our plans leaked early, the benefit of having direct player feedback and transparency would enable us to pull off this near-impossible task.
With a look of worry on his face, he put his hand on my shoulder and said: “I believe what you describe is exactly how we should be designing at Blizzard – with steady feedback and input from a group of trusted and respectful players – but this isn’t how we do things right now – only a hand picked group of designers are allowed to reveal that they are designers at Blizzard publically – and if you get outed for it, I cannot protect you from the consequences.”
I told him that I understood – and hoped Blizzard would begin to catch up with its competitors and embrace the power of transparent, direct community interaction, such as through Twitter and other media form.
I selected 6 long-time Warlock players to join a secret email list. These players, or at least the ones who replied consistently, would later be memorialized as the member of the Black Harvest. They swore an oath to keep the discussions – no matter how heated or controversial – confidential to the email list until the rework was released to the public.
Much to my surprise – they did. Incredibly eager to help – and frequently challenging both my ideas and my math – their regular, consistent feedback shaped the game as it was soon to become.
Putting Fantasy First
World of Warcraft is a game of fantasies. Power curves, cinematic experiences, strategic and intellectual challenges, all shaped to reinforce a core identity. The mighty Warrior, the evasive, all powerful Mage, the stealthy and duplicitous Rogue.
The first, most pointed bits of feedback from both the Council and Cynwise was that the 3 specializations no longer embodied the fantasies the names conveyed.
All three had become muddled and stuck around this generic idea of a ‘DoTs’ (Damage over Time effects) and ‘Pet’ class. All three specs felt the same.
This became the cornerstone of the rework. How do we create an experience that strongly links the player to the fantasy? Well, first, you have to identify the fantasies.
Shaper of Fantasy
Sitting down, I stepped back from the class as it was. I was at peace with the idea that things needed to be changed – but without a goal and direction, the shape of the work would continue to be muddled and inconsistent.
What were the underlying fantasies of each spec?
- Affliction – pain and suffering
- Demonology – power and pets
- Destruction – raw, unrestrained force
As a long-time Warlock player, these concepts made sense to me, but alone, they are not enough:
The greatest game designs happen when the fantasies and the game mechanics align.
Feedback from the Hammer
I sat down with Indalamar, aka Travis Day, to talk to him about this insight. Travis, then an item designer, confirmed that the fantasy alignment was way off – but didn’t feel there was room to change things without a ton of work.
Indalamar: “When I picked Affliction as my favorite spec, I did it because I wanted to focus on DoTs. Pain, suffering, agony, this is what that spec was about. However, for some reason, that specialization has the fewest DoTs now – and …. destruction has the most instead? I don’t get how that happened. It just doesn’t even make sense.”
Me: “What do you WANT to be doing?”
Indalamar: “I picked that spec to be a juggler. To handle lots of things running across as many targets at one time as I can handle.”
Me: “What about the rest?”
Indalamar: “I don’t know… but I do know that they just don’t make sense at all. I have no clue what I’m trying to achieve.”
His words lingered with me after I left his office. What was missing?
Stripping it down
My goal was to align mechanic and fantasy. This meant untangling all of the noise and garbage that had built up around the class since its release in November 2004.
I began with a simple mission. If I could keep only five things from the class for each spec, what would be left? What so strongly contained the essence of that playstyle that nothing else mattered?
For Affliction it was easy:
- Unstable Affliction
- Drain Life / Soul
This created the core play loop – and the fantasy of that spec. Indalamar’s words also resonated through this design. You use ‘Fear’ to push enemies away, while keeping the other effects running on the target. While that target fled, you ‘juggled’ another monster, DoTing them up and pushing them away again.
Your skill as an Affliction Warlock was defined by this ability to manage timing, target limits and pressure.
For Destruction, it was a bit harder:
- Immolate – set them on fire
- Incinerate – burn them down
- Conflagrate – at the last second
- Rain of Fire – burn everything as much as possible
- Chaos Bolt…. uhm…. because it looked cool, I guess?
There many other spells used by destruction warlocks at the time. However, as I looked at them… they all felt the same. There was so much to deal with, but all of it came down to the same player story:
“The game is telling you to push this button now. So you better push it right now – or you’re going to be ineffective”.
This troubled me heavily. While Warcraft was meant to a game to be skillfully played, the layers of complexity here were too harsh. Destruction – class focused on the simple task of “Blowing things up” was instead a class of “Simon Says”. So I sent my intention to find a new story or minigame for Destruction.
- Summon Demon Pet
- Empower demon
- … oh
When I really cut the noise away from Demonology, I realized its problem. Its core was completely uninteractive. There were plenty of buttons to push – but they were again ‘set it and forget it’ type of effects.
Everything else kind of happened ‘around’ you. You didn’t have any strategic agency, except for deciding when to swap pets or when to push Metamorphosis – your big cooldown.
So I asked myself: What is at the heart of being a big, scary demonic overlord?
- Being Tough
- Being Unstoppable
- Sacrificing your minions to save you
That was when I realized: These are the fantasies of a Tanking class, not a DPS class.
The Games of War
This is when I stepped back and asked myself a question: What does playing Warcraft feel like at a physical level?
When working on Death Knight, the answer was ‘this feels a lot like playing Guitar Hero’ – the first version of the Rune resource system did not give any wiggle room. You had to precisely hit the pattern you had the first time to maximize output. This was why Kris and I added a small over-charge window, which allowed you to be more flexible with your rune expenditures.
Affliction, with its constant in-game fear attacks, along with stacking, clipping and rolling DoTs, felt like juggling.
This lead me to explore the basic games that compose WoW:
- DoTs/Fear – juggling
- Crowd Control – multi-tasking
- Proc (Periodically triggered instructions) – simon says
- Death Knight Runes – Guitar Hero
- Rogue combo points – musical sequences / piano
- Warrior rage dumps – playing drums (repeat the same base buttons with the occasional rapid dump) and firing water guns
- Fire mage – playing the slot machine (RNG based rolling ignite damage)
This inspired me to think about creating a physical feeling for the core casting rotation, rather than just a bunch of spell buttons.
This lead to the Destruction plan.
The Music of Destruction
If Affliction was all about maintenance and juggling – what if Destruction was the opposite: A rhythmic, consistent core rotation, like playing a musical harmony?
This is what the Immolate, Incinerate repeatedly, Conflagrate button loop encouraged anyways. So I leaned into this *even harder*.
The base Warlock design was this:
- Fire shadowbolts and DoTs
- Run low on mana
- Convert life into mana
- Restore health with fear + drain life
However, this loop was horrible experience in WoW Alpha. It was merely a matter of time before a Warlock defeated any player while the other player was unable to respond. There are few things as anti-fun as long periods of agency free gameplay.
So this loop was repeatedly undermined. Bringing it back at full strength was not an option – but the concept inspired me. I bounced this idea off of the council:
What if destruction warlocks had very consistent core loops, but the more spells they cast, the more ‘charged up’ and ‘on fire’ they became – harming themselves until they released their excess power. The reaction was curious and electric – though they cautioned wisely – “how will you afford to use Life Tap if you’re already burning your health away?”
To which I said: “Well, why do you need to Life Tap at all?”
Following the path we were on, I inverted the Warlock model for Destruction. What if, instead of needing to focus on mana management, the Destruction Warlock was focused on Health management only?
Removing the mana bar entirely, the more spells, the Destruction Warlock cast, the hotter her ‘Fire’ bar became, increasing the damage she dealt to her self. I commissioned 3 visual effects – a small fire at the warlock’s feet, one up to the waist – and finally, one that fully embodied the Warlock. While levels 1 and 2 did minor damage, three was devastating and needed to be exited as quickly as possible.
Finally, huge pay-outs were needed for taking the risk of leaving yourself on fire for long. Chaos Bolt was converted to fire an additional missile for each ‘level’ of Burning you reached. However, while running around the world, if you happened to reach level 3 without an enemies around… you stood around just burning to death.
So, I added ‘Ember Tap’ which consumed all of your Burning levels and healed you. (Remember, Destruction was supposed to be a ‘backward’ Warlock) This lead to a very appealing decision point – do I heal myself to keep going – or do I unleash my damage in one huge burst to defeat the enemy?
The light bulb went on – here was a way to create a moment of agency inside the musical notes of Destruction.
This worked well for the solo experience. It was fun to take the risk of raising your ember levels as high as possible, then purging them off at the last second or blowing up one last enemy quickly. However, two major problems remaining:
- Without a mana bar, the best way to play was to never stop casting
- This causes major issues in Raids – as demonstrated by Arcane Mage missile turretting
- Players do not engage themselves with the world, focusing entirely on a perfectly played button spamming sequence
- Cooperative play became confusing for allies
- Healers, used to Warlock life tap antics, were used to seeing a slowly draining Warlock health bar, then taking action to save them.
- Now, the Warlock’s HP would be dropping, then he would instantly be full again.
Enter Chaotic Energy
The problem of being utterly ruined by constant movement during a Raid encounter was one close to my heart. It was a problem often experienced by caster classes – and one of the reasons I played a Warlock. The baseline Warlock could run from place to place and use those precious repositioning seconds to Life Tap (An instantaneous spell which had a 1.5 second cooldown and can be cast while moving) – allowing her to optimize while moving.
Without a mana bar, she was encouraged to remain immobilized – which is a huge problem in both raids and PvP. Fortunately, this is a solved problem. Meet the Rogue.
Rogues – one of the most mobile classes in Warcraft. The nature of rogue is simple – your energy bar fills up every 10 seconds. Your abilities cost 40-60 energy and fire instantly. This creates a beautiful, natural tension: You want to use your abilities before your bar is full – and you want to never be completely empty, in case a situation occurs where you want to use an ability right away.
This is easily the best resource system in Warcraft. The wise observer will even realize that Death Knights are just 6 energy bars and a rage bar. Energy doesn’t penalize small amounts of latency, penalizes long periods of inaction, creates moments of decision (particularly with combo points) and keeps the player connected to the action of the game world.
So why was this model never used for a caster resource system? The longer I looked at it, the more I realized this an ideal solution for the day-to-day woes of most caster classes. Every other class came up with some way of addressing what was ultimate a core mana problem.
Mages got instant relocation (blink) along with many mana regeneration tools. Balance druids got innervate to passively surge up manage without breaking their stride. Elemental Shaman got mana totems to recharge them and their team. Shadow Priests innately regenerated mana from dealing damage. Warlocks had life tap.
What if instead of being constrained by the mana game – we fundamentally changed it. Would this even work?
So Enter Chaotic Energy – it increased the mana costs of all spell massively, eliminated life tap – but caused your mana pool to refill completely every 20 seconds.
A Post Mana World
What happened? Total chaos.
On my first pass, you would run out of mana quickly… press your ember tap or chaos bolt… then sit there and do nothing. That felt terrible.
- When you run into a wall, you have three options
- Turn around and go back
- Find a clever way around the wall
- Destroy the wall
It felt very likely that anyone else who reached this point likely threw up their hands and retreated. The WoW class design was regularly under immense pressure – each designer handled 3 or 4 classes. Multiply that with 3 (or 4 specs in the case of Druid) – and there’s only so much time you can commit to solving these kinds of problems.
However, I wasn’t under that kind of constraint. During MoP, I had been moved of the Raid and Quest teams to develop Warcraft Pet Battles – which had long periods of downtime while engineers built the core technology.
This was just the kind of time needed to really dig in and see if the ‘mana energy’ problem was innately flawed – or hiding a potential gameplay gem.