Knockout City — Conception to execution of an original multiplayer level
This post was approved by my previous employers Velan Studios and supervised by several key stakeholders at the studio who were involved in the process of when I wrote this. This blog was also written prior to the launch of “Knockout City”.
Overture
It’s crazy to think that 2 years ago I had just wrapped up my final iterations for Back Alleys, our second confirmed and finalized level for Knockout City which was extremely exciting. Back then the game was in the stage of where we were finalizing a lot of our design goals and our levels team (which consisted of the environment art team, QA for levels, my design colleague as well as me) was just starting to narrow down the art for our first playable level in Viper, our new game engine, “Rooftop Rumble”. We had traveled far in a relatively short amount of time and it was very exciting. To see the levels I had worked on, which we knew were fun, being arted up and tested heavily in other stakeholders’ hands was an enticing experience.
A few years on and with the game finally launched, I am now in the position to see what the public thinks about the levels and what players are doing to them. It’s certainly a surreal experience to witness players around the world shouting “Rooftops! This is my favorite map!” and knowing that this is the map you helped create.
With the game finally out, it got me thinking about how as a team we got to this position, to witness these exciting moments. How did we manage to make a level like “Rooftop Rumble” work for our game? How did we make an original level, for an original IP? How did we build and construct our own rules to get us the best representation for our game that no one else was making? Well, I thought I would share the journey of how we got there. An exclusive look at how we built a single level for a one-of-a-kind dodgeball game, for a start-up business, with a proprietary engine and a shared ambition to make something seriously tangible. This isn’t a blog about how to make better levels, or how to make one a better designer but rather the story of the journey it took to make one solid, competitive level for Velan/EA’s Knockout City.
Break on Through
In the late summer, I received an opportunity to speak with (David Nathanielz) about working on a new IP as a designer for Velan Studios working with Jeremy Russo (Game Director for said IP). They were impressed by one of my personal projects that I had done for Team Fortress 2 called “Closed Encounters” (a BioShock themed map). I didn’t get much information, other than it was an original multiplayer IP but I did know that it had some great people making up the core foundation of the studio. It was a risk for me to jump into the unknown but I did it anyway because I love making games and I love working with great people.
When I eventually moved to upstate New York (with a single suitcase and nothing else), I was quickly on-boarded about the game and was surprised that this idea hadn’t really been thought of before. A dodgeball action game? It sounded as epic as it played and made a massive impression on me.
As I got to know more of my co-workers, I realized that the game had the potential to become a great game to play with others, as collaboration is key if you want to win games. However, losing a game was still a lot of fun. The game always felt like a game you could just pick up and goof around on for a few hours with some friends and then do the very same thing the next day. It was exciting and on top of that, it was also being published by EA who is one of my favorite publishers!
Early on we wanted to experiment with all kinds of different layouts as we didn’t have any foundation to work with. Some previous prototypes were decent enough but they were either too small, or too symmetrical. Metrics were still being established and mechanics were still being finalized. I (at the time) was the only level designer so I quickly had to grasp what would and wouldn’t work for our game by pumping out all kinds of crazy maps. Not based on themes that were already determined by the narrative team but just finding ways to break or discover fun aspects of the play experience. We went from creating industrial, cage-like basketball settings, to aquarium interior levels, followed by urban parks with all kinds of wacky ideas that manifested into almost a volume worth of pros and cons we could use to determine our level language.
Every play-test (of which we had daily doses of), I would try and focus on questions that were important to our understanding of good vs bad metrics. Questions such as “Did you feel powerful in X Space?”, “Were you lost and if so where?”, “Did you notice X mechanic and what did you think it did?” etc. To me (at this point), it was not important to ask if you liked or disliked the level. This was never final and I was looking to gather data that would be our level language. Jeremy and I worked tirelessly to make sure we could have a document that filled up with soft rules to create good spaces that didn’t feel like players were cheated, lost, and didn’t solely rely on art to create POI’s (Point of Interests). I wrote a large traversal metrics document that could help Design (and other stakeholders) identify what works and what doesn’t work for our game. Anything that was new or flat-out didn’t work would be updated to make sure we were on point and ready to go for anyone who wanted to understand our level language.
Once we did that, the hard work began. Two months on from my start date and it was time to make our first themed Knockout City level “Rooftops”.
“Creativity is just connecting things”
If you are looking to find out (in this blog) what theories we used to make levels or design principles in the planning, creation, or iteration of levels for Knockout City then I am afraid you will be very disappointed.
Knockout City was such a unique game that we could not rely on typical level design principals alone. This means that I was completely on my own with our two months of extensive play-testing and documentation research. No composition analysis or extensive blogs really helped us as anything we did try would end up feeling pretty bad for our game.
We tried creating a level based on an existing level from another game, but it just felt open, cumbersome, and hard to know where we were supposed to engage with players. The game it was designed for was very different from ours and no matter how hard we tried it just was not fun. The truth was obvious… We had to make our own original content for “Knockout City”.
The pitch Jeremy prepared for me was straightforward, “make a level with a ball in the middle of an exposed, vulnerable space in the highest part of the city”. At the time there was no designed mechanic but we’d worry about that another time as getting the geometry right was our most immediate priority. The key experience we wanted players to have (across all our levels) was the ability to go to every part of the map and fight in every space in every round of Knockout City, using only simple primitive shapes, simple colors to separate the spaces and have a fort-like feel to the map.
The first thing I did was pretty much go back to what I was already doing and try new stuff and get feedback but it didn’t work out as I hoped. Something was missing as maps felt flat, hard to read, and not very gameplay-driven. For a few weeks, it felt like I was getting this same feedback and it was hard to find out what was going on. Was it my layout? Was it my use of shapes? Did I follow the metrics too much to the letter?
I was a bit stumped and really didn’t know what to do for a while. It really dented my morale as I felt I was thinking in the right direction but it wasn’t reflecting on my work. I had this book (and I still use it to this day) where I would write out what I was doing wrong both on a personal as well as professional level. I always believe that I can do better as I am not interested in seeing myself as “there” because I was working with an awesome team making a really special game at a game studio. I wanted to create awesome experiences and I am not comfortable being content like that. As Michael Jordan once said, “Some people want it to happen, some wish it would happen, others make it happen.” I sure as hell did not want or wish to make things happen. I am a do-er so I made sure to try many different approaches to get the results we needed.
I decided to think differently about what I was doing so I went to Eric Feurstein (Narrative Director) and asked him what he wanted players to feel on top of two buildings (on a narrative level). My goal was to maybe lean on a specific feeling that the level could convey. He gave me the key word that got me out of my funk. The word was “Ownership”. He wanted players to own a building, have the right to claim it, make teams fight for both sides over and over again. That word triggered me. When I heard it and ran straight to my desk where I developed the idea of having two large buildings made up of two mini arena’s that players could reach either on a small plank of wood with a ball in the middle of it or gliding across it. Making it a high risk, high reward designed map. Make sure both sides have equal strengths and weaknesses. Dodgeball is about owning people so why couldn’t players do that in our game!
I wrote in my level design sketchbook a scope with Eric and Jeremy’s expectations as well as our design goals. It helped me think differently and layout out what a level should feel like as well as mentally construct the layout to get it to work. Very quickly I got a sketch down on paper, built a mood board in my head of architecture I wanted to build, got my Pink Floyd Spotify playlist ready, and hit the editor.
Touch Off
Early on (way before my epiphany) I knew exactly what my architectural inspiration was going to be based on. Some close folks around me (who played the Beta) assumed I got inspiration from BioShock Infinite as it’s one of the games that made me want to become a designer. Whilst I definitely drew inspiration from the game, “Rooftops” was actually inspired by my grandfather Dimitri Koutsodimitropoulos.
My grandfather was one of the first to notice my ability to build things out of anything. We would spend most of my holidays at his home in Athens (Greece) buying Lego and Playmobil, creating cool train stations around his whole house, whilst later on sketching every component of what made a Formula 1 car (my first love) tick. I loved anything I could make and my grandfather shared that passion with me every summer I went to see him (till he passed away in 2006). He was a brilliant, kind, and passionate man who would get incredible joy watching my brother (Marc) and me build all kinds of crazy things in his living room. He was one of Greece’s best civil engineers and his home inspired me enormously (as did he).
In Athens (and Greece in general), people live their lives outside and their terraces reflect that. The terrace from one of my family’s homes felt more like a dining room than just a terrace and it made me ask myself, how could I implement that somehow into our game?
Every area didn’t need to be built around unique architecture that sold the narrative of this space. It was a nice thing to have but it had to feel good, feel like it should be “owned” and not feel like it was a space I had a nice stroll through. Using the key business word K.I.S.S. (Keep It Simple Stupid) I wanted to make throwing balls and dodging through spaces always feel unique only found in “Rooftops”. Each space needed to have character and get players calling them out, like at a Tokyo Toyosu fish market. Especially on a level that I insisted should try to not be symmetrical in any shape or form as I wanted to push the envelope of being competitive without the temptation of making each building players fought on feel similar.
Building a rough prototype for a level for Knockout City takes about a day or so. Sometimes less, sometimes more but that always depended on scale and scope. The first draft of “Rooftops” new ownership-driven level was done in about half a day and instantly it felt too small but crucially it felt good. Very good. After the first playtest it was obvious that I was working in the right direction. I decided to extend the lengths of the building and focused on giving more language to each building.
If you break down “Rooftop Rumble”, it’s basically 5 spaces. Each space was built on a player’s need which is very simple in Knockout City as that was the ball itself (the main POI)! Most of the time, players are constantly trying to get a ball, as it’s the most effective need to win games (think of it as Knockout City’s hierarchy of needs). I won’t break it down too much but in short, the space caters to the need and the combat that surrounds that need. I drew a lot of inspiration from classic and current arena shooters (such as Doom and Quake) but also a lot of non-game-related interests. For instance, in the beta version of Knockout City, the disco section was based on a Pink Floyd cover band concert I went to in The Netherlands, whilst the garden section was based on my mother’s love for gardening (especially in one of our homes in the UK). The VIP section was based on one of the Star Wars prequels (Episodes I through III) and the garage section was a door stopper I found in the office. The middle sections connecting these were a bit different. On the garden/VIP section I drew inspiration from high rises seen from the BioShock trilogy and on the disco/garage section I tried to build around my grandfather’s (as well as typical Greek homes) roofs. Ventilations, paneled ramps, glass window tops, hedges and ledges all would be nice little additions that filled the spaces with places to run, brawl and knock people into action.
Throughout the iteration and blocking out experience I also used YouTube to just watch Japanese vloggers (like Abroad In Japan, Sharla and my friend DaveTrippin’) stroll through cities and try to decrypt the way they moved through small as well as larger spaces. I have no idea why I chose this unconventional approach to iteration but I think it was related to the way the Japanese think, which is that “less is more” (or Zen Design in short). I wanted my spaces to have that feeling so I felt by seeing people traveling through spaces (in a country built around this way of thinking) it would be beneficial.
During the iteration stage, I also leaned on some of my coworkers for feedback. In particular Tristan Meere and Dan Hunter who were very good players but also had a solid understanding of player needs from a non-designer experience. I would always love bringing them over to my desk and asking them what their honest thoughts were when they traversed through the map. They would make great observations on the importance of clarity for both new and experienced players, while still making sure that the gameplay would shine through in key moments when it mattered most. I sometimes would spend long lunches with Tristan focusing on how we could make sure we got the most out of each other’s feedback.
Many times after work I’d sit with Dan and ask him what QA were feeling about playing the game. Always honest, he’d be straight with me about it in both the positives and negatives. Sometimes he’d capture highlights and send me some key moments that he loved and I would use it to make sure we could encourage more spontaneous moments like it.
Very quickly, we got Jeremy to play and he was very happy. It very quickly became a priority to get the smaller details flushed out as we were both content that this map would be our flagship map as it was built to compliment the gameplay so well.
Later on, we eventually tried to come up with a mechanic for the level to have (as every map must have one) as it helped spice the action up just a bit more. The way I usually see a mechanic for a level is to make sure it feels purposeful, present but also under the player’s control. Jeremy and I both saw eye to eye and decided that we’d try and create a mechanic to answer a problem with the map which was that gliding from one building to another was sometimes a bit too risky and hard for newer players to grasp. So we thought about updrafts (or “High Winds” as we eventually ended up calling them) that only worked when the player activated the glider. It would correct the player’s glide downward angle to be more straight and be slightly faster when they went across. It was simple, effective, and also beneficial to traversal when it happened.
When Velan did studio-wide playtests, our team was super ecstatic and crucially exclaimed all the things we wanted to get out of the levels. Rooftops was soon dubbed our first working level and I was motivated to continue playtesting as well as iterating it to get it into an even better state. I was super excited to get everything in place to make sure it was ready to use as our flagship level. Internally (for a while), rooftops became the exclusive level for in-studio competitive matches (or Turf Wars in short) as it had the most balanced gameplay moments from any other map. Now of course all the levels are unique and competitive to play on but Rooftops really is our most of all the levels.
Bloody Well Right
Just because the level was blocked out and gameplay ready, it did not mean the work for Rooftops ended. Levels Art now would take over and the roles would reverse with me giving them feedback and advice instead of the other way round. Tristan, Cory and Erik (McKenney) were always very open to building a working relationship between themselves as well as design. One notable issue we tackled together was the “background buildings”, as the game has to have nice backdrops but we taught players that throwing balls (or characters) against walls had a bounce-off metric. Meaning if you have to throw (even without collision) we needed to define a comfortable distance that made it clear what the boundaries were for those buildings.
We did some more testing to define these metrics and thanks to Tristan (who was always open for discussing) we did this and it helped paint a language for the art team to build upon for future levels (as well as Rooftops). It was incredibly important to me that we kept the stakeholders we needed in mind so that we could cater to and accommodate them. Our teams always loved to spontaneously discuss how we can inform our players with strong art but without endangering our design goals in the process. It helped that the level art team was always pro-gameplay, so no matter how we achieved our goals, it would help make us build an even better experience.
Eclipse
So 2 years later, the world knows about Knockout City Rooftops, and hopefully a little bit about what it takes to design one level for Knockout City. What made Rooftops work was that it could only work on the philosophy that our game was too unique to be designed in traditional level design methods. There wasn’t a magic formula that got us the level the way it should have been made. It came from a strong tenacity to test, gather results, and reiterated multiple times during production.
For Knockout City, the way we designed “Rooftops” was simply to have a strong mindset to think differently and uniquely for every multiplayer level in the game. It kept every level fresh, developed an identity but also had its unique moment in the sun. This is not to say we didn’t learn from any of our successes but when you build levels that have a unique mechanic it’s hard to carry the same solutions to different problems. I always believe the best level designers are the ones who not only learn from failure but accept it. It is something that comes from serious practice, iteration, and constant striving to be better. Knockout City would never have been made without this mental approach and I am not sure I’d be here to write about my experience.
Every level I worked on for Velan was always built differently and it always focused on making sure it felt good for as many players as possible but also helped highlight how awesome the game is. It was a privilege to work on Rooftop Rumble, Jukebox Junction, and co-designing Back Alley Brawl of which I really hope that players enjoy playing in. I am proud of the work I did for the game and have zero regrets about the decisions I made for the project which meant dearly to me.
I do want to thank everyone at Velan and especially Jeremy, Eric, and Merc for their excellent insight, direction as well as a mindset for thinking differently. We did some great stuff together and all the levels I made could not have been made without our strong tenacity to think in the interest of the game as well as the players. I always got a kick getting the controller in someone outside of our team’s hands and I am certain Jeremy, Eric, as well as Merc, did too! Speaking of, I also want to give a special shoutout to Dan Hunter, Tristan Meere, Kara Zisa, Matt Pirog, Cory Hamilton, Dean Russel, James Weir, and Gregg McClellan for coming up to the playtests but also to bounce feedback outside the tests or after work. You are not only incredible colleagues but also incredible friends. I really loved but also appreciated their different perspectives, as all of them were different types of players that I aimed to build level language around. I would also like to thank my friends but also mentors Bill Gardner, Mitja Roskaric, Michael Herbster, Armando Soto, Richard Wyckoff, Jeremy Pryer and Christian Senn who helped me get to where I am now as well as inspire many aspiring designers with their incredible work but also kindness. To my closest friends Bodhi Mulders, Noah Kobus, Giovanni Lauffer, Jaymie Dornan, Cameron Keen, Tony Choi, Noah Blackmore, Louis Masi, Daniel Lozano, Charles Jameson, Phil Hill, Alex Baldwin, Ruben van den Hudding, Steven Cancelmo, Mittens, Anthony Omens, Chris Anderson, Jaymes Taylor, Mariella Bailey, Mickey and Charlie de Haas and Andrei Ciubotaru thank you for being there and motivating me through this process. I would also like to thank Jay-Cee Klees for his incredible artwork to make this blog post come to life! If you’d like to see more of his work, visit his Twitter as he posts regularly. Most importantly I’d like to dedicate my levels to my Mom, Dad, Brother, Ella and all members of my family from The Netherlands, Greece and the US. Your love and support has been surreal. To my Grandpa “Mimis” this one is for you.
Finally, I of course need to thank Karthik, Guha and David for giving me the chance to work on something truly special. Rooftops or any of my other levels would not be where they are without any of the above and I can not wait to see what players think. You can play the game now on all major platforms (here for more)!
If you’d like to see what I am doing right now, you can hop on to Twitter and follow my handle @willem567! I am currently working at Zynga on Star Wars: Hunters and having a fantastic time making some more awesome content with some fantastic people!