World of Warcraft Classic: “We had to recreate bugs to get back to how it once was”

Just in time for its 15th anniversary, World of Warcraft is going back to basics - officially, this time

After World of Warcraft fans spent years asking for an official option for legacy servers, Blizzard finally gave them an answer back at BlizzCon 2017. A year and a half later, we finally have a WoW Classic release date, a beta on the immediate horizon, and a much more precise look at how this trip back to 2004 will work.

If your memory’s become a bit fuzzy over the past 15 years and you haven’t dug into the myriad unofficial legacy servers we’ve had in the meantime, you’ll find the old WoW to be a much slower-paced, more exploratory game than it is today. The map isn’t dotted with convenient quest markers, it takes ages to travel between locations, and any encounter can turn deadly in an instant if you pull too many enemies.

In other words, it looks like the game is exactly what WoW fans have been asking for. At a Blizzard event showcasing WoW Classic, I spoke with senior producer Calia Schie and game director Ion Hazzikostas about the challenges of bringing retro WoW back for a modern audience, the things players have forgotten along the way, and why it has taken so long for Blizzard to finally get legacy servers ready to go.

PCGamesN: People have been asking for legacy servers basically forever. Why is now the right time for WoW classic?

Ion Hazzikostas: Because we finally figured out a way to do it. As you know, questions have been asked at BlizzCon and elsewhere for a long time, and the answer of ‘no’ or ‘we can’t do this’ was never a philosophical one on our part, nor a matter of stubbornness. It was simply that every time we had looked back at what it would entail to bring the original World of Warcraft into the modern era, the obstacles seemed insurmountable.

We had the original 1.12 client and server and data. But they were designed to run on hardware that hasn’t existed for a long time, in an overall tech stack and part of a broader Blizzard infrastructure that hasn’t existed for a long time. It was full of bugs and exploits and tons of things that really were addressed over the course of literally tens of thousands of hours of programming efforts between 2006 and today. So the idea of trying to retrace those steps seemed virtually impossible.

But I think it was really just a few years ago – two and a half or three years ago – in part due to a more intense community focus, we took a much harder look at the question and said, ‘ok, we think this is impossible, but what if we had to do it – what approaches might we take?’ We started to go down some R&D paths. One of the ideas we came up with was an experiment that was spearheaded by a programmer on the WoW team, Omar Gonzalez, one of the members of the Classic team. What if we took our modern client, our modern server architecture and instead taught them to speak the original 1.12 data? Rather than working forwards, what if we took what we have now, trying to kind of interpret what was there back then.

He spent a few weeks hacking together a rough prototype, and there were tons of bugs – things didn’t fully render, and there were all sorts of UI elements missing, and things weren’t accurate. But at its core, it was 1.12 WoW. It was the old world. It was the old skill system and talent system, and it was there running in the modern client. And the fact that he was able to get so far in that amount of time gave us confidence that this is something that we actually could do. There wasn’t going to be hundreds of thousands of hours. It was to be something that we can actually deliver in a number of years and get to our players at a level of quality that we deemed worthy of Blizzard and worthy of what they expect of us. That’s the story of WoW Classic.

Ultimately, I think we came to that conclusion really late into summer 2017, and announced it at BlizzCon a couple of months after that. Uncharacteristically early in a lot of ways for a Blizzard announcement. In part because we were just really excited that we had found a path to make this happen and wanted to share it with the world. We knew there’s gonna be a ton of enthusiasm behind it. We knew those those questions were going to continue. People have been asking them forever. We wanted to be able to say, ‘yes, we’re doing it’.

The journey from that moment until what we have today has been the full occupation of the classic team, which has grown. We’ve built a team around this endeavour. Here we have today on the floor, something that’s pretty much – it’s WoW. If you’ve sat down and had a chance to mess around with it. It’s WoW, as it was in 2006.

Has the development at all been informed by what you’ve seen from the fan-run legacy servers?

IH: There’s certainly awareness. There’s certainly a passionate community there. It certainly has raised questions for us to grapple with – I don’t know that it led us to answers to those questions. But certainly things like how should we handle unlocking content? How should we handle rewards? We knew that these were questions that the community around these servers had grappled with and was divided over. We knew we were going to have to engage with them.

One of the advantages that we have on our end, creating an official version, is this original 1.12 reference client. Getting feedback, going back to the BlizzCon demo, or otherwise hearing from players about things that felt wrong to them was interesting, because it was like ‘no, this actually is how it was. We can assure you the data has not changed’. But other servers out there that were trying to emulate this experience weren’t always getting all the details right. Whereas we had the original data. We have the original code without question. But in terms of the broader experience it’s been great to see the passion of that community, and to be able to deliver something they’ve been waiting for, for such a long time.

Calia Schie: The project is being run by a fully internal team, within the greater World of Warcraft team.

What have been some of the more challenging technical aspects of getting this up and running?

CS: You say technical, but for me, as the producer of the team, finding a team that was as passionate about this project as our community and technical experts in their own right was actually a large challenge. We were very, very meticulous about the people who we interviewed and how that process worked. We took people from internal to team two, to the World of Warcraft team, internal to Blizzard, and some external people. But we were very methodical about going through the list of criteria because we wanted to make sure we had the right team. For me, that would be the biggest challenge.

IH: I can’t speak first hand as to what was the hardest technical problem to solve from engineering perspective. From a layman’s outside perspective, looking at some of what they were grappling with, I think one of the more fascinating problems has actually been on the side of our engine and our renderer. There are fundamental changes that have happened over the years, just in terms of the math and the geometry of how we put pixels on the screen in response to modern graphical hardware and the way that’s evolved.

A lot of aspects of the visual look that was distinctive in 2006 were products of limitations back then, or in some cases the products of actual bugs. Taking the original 1.12 data, and putting it into the modern client – often what came out was meaningfully visually distinct. So there was a mix of some artistic tweaks and a lot of changes to how lighting worked in the engine to try to accurately reproduce the 2006 look.

There were some fascinating conversations along the way with folks like Chris Robinson, our art director. It was surreal almost to see our art director, who is one of the best artists I’ve ever met, and someone who pushes for perfection and refinement in everything that we do, advocating for actively working to make things look worse. Because aesthetically – objectively, to the extent that such a thing exists – the modern client was brighter, was better. We wanted to adhere to the original look and feel of the game. We had to effectively recreate bugs in some cases, to get back to how it once was.

Is there anything that stands out as one of the biggest misconceptions people had about how vanilla WoW was supposed to play?

IH: This maybe isn’t the biggest or best example, but it’s the first one that comes to mind for me. Players on our forums coming out of BlizzCon were adamant that we had failed to reintroduce a Warlock pet mechanic called summoning sickness. Where, if you had a pet out as a warlock, and you went to summon a new pet, your currently active pet would be stunned during your cast. They’re like ‘you guys forgot this, please put it in’.

Prefer the classics? Try out the best old games on PC

Initially, our first reactions were ‘oh, crap, we forgot that’. That was a thing that existed back then. And I guess somehow we missed it. But we went back into the archives and found that it was actually added in Burning Crusade. It was added in Patch 2.0, not 1.12. It was there in the early days of Arena. It was there just around that time, so we remembered it. When someone initially told us all ‘yeah, that wasn’t Classic’, we’re like, ‘I guess that wasn’t Classic’.

We didn’t fully believe it ourselves until we dug through the data. It’s one of those interesting things – memory is powerful, but imperfect. With as much time as has passed, one of the fascinating things about seeing people who played back then go into Classic is the little bits that aren’t quite how they remembered, and their sort of reconciliation of their memories that they were so certain about with the reality of how certain things work.

Has the demo given you a better sense of how much interest is out there, especially in terms of planning for server load when the game goes live?

IH: There are clues. All the information we can get there is helpful. We have a couple of philosophies guiding us as we plan for the launch of Classic. First off, in the long run, we want to have healthy communities – large populations that can sustain a server for some time to come. I think that’s a key part of what made the Classic experience what it was. We’re erring on the side of launching with fewer servers rather than tons and being sure that we’re going to meet all the demand.

If we need to bring more servers online, we will. But we’d rather bring servers online as needed than find ourselves with some servers that have become ghost towns a few months after launch. The part of why that’s an acute concern for us is there’s some differences between the launch of WoW Classic and almost any other traditional MMO launch.

Because it’s part of the existing World of Warcraft subscription, there’s going to be plenty of people, we’re certain who go check it out, with not necessarily any long term intent to play. They may be intrigued and actually may end up spending more time there than they expect, but a ton of people are going to be there in the first week just to see what all the fuss is about. Whereas there are others who’ve been waiting literal years, and they’re in for the long haul. But they’re all playing on the same servers. We can’t necessarily differentiate the two up front.

We’re expecting a certain amount of initial tourism and folks who will then move on, and we want to make sure that our server populations a couple months in are still stable and healthy. If we aimed for low server caps at the start that were more accurate to what servers can hold back in 2006, then a bunch of those people moved on, we’d probably end up with underpopulated servers.

We’re aiming at launch for much higher concurrency caps on our individual servers than what was possible back then. In order to facilitate that, without having the gameplay break down, we have some relatively new technology that we’ve worked on. Its vaguely related to the sharding that we have in Battle for Azeroth. But it’s actually running multiple entire copies of the world.

Rather than copies of individual zones, let’s say if you’re playing on the Mankrik server – which I think will be one of our server names, because everyone remembers Mankrik – there will be three entire copies of Azeroth, each with their own populations equal to what a launch server back in 2006 would have been. When you log in you’ll move seamlessly throughout that world, but we can load balance between them.

Over the course the first few weeks, we’ll both be raising the population caps on those worlds as players spread out more, and reducing the number of them. With the goal of, once we’re a few weeks in, having all of our realms collapse down to a single cohesive world per realm, no sharding or anything like it. But that will help us manage the early-day player load while still keeping healthy populations in the long run.

There’s already a WoW Classic roadmap outlining post-launch content additions. How do you reconcile the idea of this release capturing a certain moment in time for WoW with the need for continued updates?

IH: I think it’s systems versus content, ultimately. I think systems are all defaulted to their 1.12 state. We’re not changing stats on items out from under players. We’re not changing the rules of how the game works out from under players. But when it comes to content and rewards, there is a journey. Having all of that accessible from the start would lead to that content stepping on itself in a lot of ways.

There were a lot of elements that were added as part of the original patch plan not necessarily to fix problems that were perceived in the game, but to solve problems like catch up. A year and a half into classic WoW, when raids like Ahn’Qiraj came out, there was a desire to give players a way to gear up for those raids faster than having to spend a year running Molten Core and Blackwing Lair. So parallel gear paths came out.

If those were there at the same time as the early content, they would actually obsolete Molten Core and players would never have the experience of feeling like getting 40 players together and mustering the effort to defeat Ragnaros was worth it. Because you could get comparable rewards much more easily with just a few players or even solo in some cases. So to preserve that structure, that experience, we’re staging the release of the content itself. But when it comes to how player abilities work, to how the honour system works, to skills, talents, stats on gear, all of that. That’s the 1.12 state.

YouTube Thumbnail

CS: We changed that based on player feedback.

IH: We did. We announced a four phase plan at BlizzCon and then heard a lot of feedback from players that didn’t feel granular enough. Things like Dire Maul, the dungeon, being open literally day one, so the players could do it while levelling up, would actually overshadow a lot of the endgame gear that was initially the only thing you could aim for when WoW came out. Things like that and other pieces of feedback were what let us to move to a more granular six stage plan instead of four.

In terms of the moment-to-moment player experience, what do you see as being the biggest difference between Classic and current WoW?

CS: I’m a new player to WoW Classic. I didn’t get a chance to play it the first time around. So I’m catching up now in our employee alpha. Both games are very different. It depends on what kind of unique experience you feel like playing. For me, I’ve found the pacing very different in WoW Classic. It allows me to become immersed in the storyline and actually read all my quest text and go in and explore in areas that are foreign and new to me. That has been the most surprising and enjoyable part of this experience for me.

IH: For me, it’s the pacing of combat. I would say it’s tactical versus strategic, if that makes sense. In modern World of Warcraft, you have way more tools at your disposal, and it’s about choosing the right tools to overcome a specific encounter or specific combat with a creature or group of creatures. Whereas in classic World of Warcraft, combat is slower and often simpler in most ways.

You may only be pushing one or two buttons in order to defeat an individual enemy. But there’s much larger scale resource planning on a macro level. ‘All right, how much mana am I ending this encounter with? Am I going to stop and eat? Am I going to kite this enemy so that I don’t have to eat after this fight? But instead, I can make my mana bar last for these four enemies, and then I’ll stop and drink’. It’s efficiency on a larger scale, whereas modern World of Warcraft is more about ‘All right, let me interrupt this cast. Let me make sure I’m doing my rotation correctly’, and so forth.

Besides the players who played WoW back in the day, have you seen much interest from newer players who want to see what the game was like back when it launched?

CS: Absolutely. I think players like me who feel like they missed out the first time, they’ll be so encouraged to play this game and see what we’ve done with it, because you almost feel like you missed out on a piece of MMO gaming history. I don’t want to be left out of that. So I’m going to make sure – doesn’t matter how long I have to queue for that on day one – I’m playing with everybody else. So that is another audience that we’re tailoring the experience to.

IH: I think while the Classic, legacy server community, has certainly been one of the motivating factors for us, the game we’re making is a game for all players. Just as World of Warcraft as a whole is a game for all players. We know there’s a passionate existing fan base, but we’re also looking forward to many new players discovering and exploring the origins, the story of where this all began.

The Classic beta launches tomorrow. What content will be available?

CS: We’re capping it at level 30, so it will be one to 30. You’ll be able to play all the way up to level 30.

IH: We’re starting there. Then we may increase the level cap or have focused tests on specific pieces of content as needed. This is gonna be a pretty small beta. Uncharacteristically small compared to our expansions. We’ll be inviting people from opt-ins, of course, but this isn’t like a brand new game, where we’re really looking to get a very broad base of feedback on ‘is this fun? How can we make this better? How can we fix this system?’.

Ready for adventure? Head into the best RPGs on PC

Here, we’re really just looking for a final sanity pass from our engaged, enthusiastic community to make sure we haven’t missed anything. Are there bugs, if there’s some inaccuracy, something that you think we got wrong that we need to take a closer look at. Because we’re not looking to change the spawning, or to nerf or tune Herod and Scarlet Monastery based on the feedback that we get. Also, to some extent, we want to preserve the excitement around the official launch, to make August the first chance to really go all the way.

I feel like I know the answer to this, but I have to ask – are there any plans for expansions?

CS: For now, we’re just focused on recreating the authentic, classic experience. But we’re open to community feedback. After that point, we’re listening to our fans and we’ll be guided by their voices.

IH: Exactly. We’re certainly open to it. I think we’ll all be much wiser a year or so from now. Just see where the community is, how Classic is unfolding, and what makes sense for our fans.