Constructive Feedback

This is the second in a series of articles for the Demonbreed campaign for Left 4 Dead 2. You can read back how the first map was designed here.

Video of the subway level

The subway is the second map in the campaign, so obviously the focus points were slightly different for this map.

* Slightly higher difficulty than first map.

* Better weapons and pick ups.

* A coherent connection to the previous map.

* A change of scenery. Changing wide open areas to tight tunnels.

* Bringing the hell theme to the front, to make the level vastly different from the subway in No Mercy (one of the official maps for L4D1).

* Maximum use of vistas and decoration.

* Different more intense panic events than the first map.

* New door mechanics

The players start off in the saferoom where the previous map ended. As they exit some creepy music kicks in, and a fence to the right allows them to see the last part of the previous level. This is done deliberately to tie the two maps closer together. All too often do L4D2 campaigns feel disconnected, and that is something I really wanted to adress here.

As the players make their way down the first tunnel, smoke obscures their view, and fire burns down below the rusty gratings. They are greeted by cocoons and spiderwebs, a theme that was already lightly introduced in the previous map. This tunnel intentionally bends left and right a bit, to make the corridor more interesting and to make the room that follows a little bit of a surprise. In long corridors like this it is also important to have spots where special infected can spawn. This is why there is a hole in the ceiling. A hole in the ceiling is something the players easily miss, since their view is focussed on what is ahead of them. The bending of the tunnel also allows infected to spawn just out of view from the players. There is a small waiting area just before the main lobby on the right hand side. This small side tunnel allows players to dodge incoming Charger and Spitter attacks, but it also allows the Charger or Smoker to spawn behind the players as they enter the next area.

The main lobby had to look as hellish as possible, because this map is an introduction to the hell theme that follows in the next map. I had to slowly build that up, so I decorated it with lots of barb-wire, metal plates and boarded up tunnels. The hole in the middle of the room is an opportunity for a Charger to insta-kill a player. The players are also forced to walk around it, allowing a Smoker to perhaps drag a player into the pit. Infected can spawn on the upper levels of the structure and drop down from various points, as is visible in the video when a Jockey climbs up some of the pipes. The upper level connects to the hole in the previous corridor, allowing enemies to attack from a variety of directions. The lobby has various fake tunnels that suggest that the underground train station is a lot bigger than it really is. This helps the map feel less confined. At one point I had an idea of perhaps randomly opening up different paths, but I got rid of that idea because that would mean skipping the waiting area and the associated event.

The players head down another corridor, where more smoke obscures the view. There is a corpse on a bench that triggers a scare sound if players get too close. I like adding these mini stories to levels, because it adds a bit of character. Splotches of blood and bullet holes suggest this victim was killed in a spray of gunfire. The players arrive at a creepy looking door, which when triggered, opens slowly. I wanted this door to look unnatural and dangerous, which is why it has metal spikes that interlock. Players must be careful when closing the door, because anything caught in between is instantly killed. It closes slowly, but what I really liked about this concept, is that it allows the players to secure the area them selves. However, I soon realised that this could potentially cause players to lock another player out by accident. Which is why I added an extra tunnel that can only be opened from the other side. This would allow players to go back to the previous area and rescue a player that was left behind.

In the waiting area the characters automatically comment on the pile of corpses that covers the floor. These types of automatic dialogues tend to add a lot of life to an L4D2 campaign. A switchbox is conveniently located in front of a window, allowing players to see a train from another world arriving on the platform. This is all part of telling a story within the level. After activating the switchbox, the exit can be opened. The waiting area is meant as a “breather”, a resting area, where the players can stock up on supplies.

Outside the waiting area is a random event. 50% of the times the elevator opens and hordes of zombies come running out. The other 50% of the time, a loud puff of steam simply bursts from the pipes right next to the players, just to scare them. There is a rescue closet nearby, incase any players need to respawn after being killed in a coop game (rescue closets aren’t used in Versus mode). In coop, players can rescue their team mates from these closets, to bring them back into the game.

The players head down an escalator and arrive at a collapsed platform. The players have to jump down into the zombie infested water here, but if they do not jump down together, a Smoker might come up from behind and grab a survivor before he can join his friends. This is a deliberately designed separation point. Separation points are great in Versus mode, because they allow the enemy team to kill off a survivor early. By working together and timing their attack right, an Infected team can abuse this area to separate a survivor from his team. This area also allows the players to have a distant look at the bionic train that they saw arriving earlier.

The next creepy door triggers a horde, forcing the players to quickly head inside and close the door. If a player is locked outside by accident, a door will open after the event, that allows that player to catch up. As the players lock themselves in, a train comes crashing down on top of the room and the players must fight a horde of zombies in a really small space. This is a really tense battle, and will consume a lot of the ammo and health from the players.

Now we reach a point in the map where we have a lot to cover. The finale is cleverly constructed, and although the players haven’t reach the finale yet, it plays an important part in the lay out of this area. The players leave the falling train area, and probably find themselves low on ammo by now. Above them is a small walkway that is purely used by enemies. There also is a connection to the catch up tunnel here. Up ahead are precious tier 2 weapons, the best weapons in the game, and extra ammo. A breakable wall nearby allows zombies to ambush the players unexpectedly. They will most likely be facing the wall if that happens, because that is where the supplies are. Then they head through the train carriage (a placeholder for now). This is often where a tank or a witch spawns, and this can easily drain the players of a lot of ammo (good thing there are supplies here). The players make their way to the other platform which is half flooded with mirky water.

The tunnel has collapsed onto the train carts, so the only way to continue is by going through the trains. At some point the train rumbles and dust drops from the ceiling, to frighten the players and make them aware that the tunnel is unstable. After a few train carts the players need to cross over from one train to the next. If the players explore a nearby corpse hanging through one of the windows, a scare sound will be triggered. The players leave the train shortly and traverse the platform with the strange bionic train blocking the left side (this is the train they saw arriving earlier). This was an excellent opportunity to allow the players to have a close look at the creepy train.

Then they head through another traincart, and through a waterfall of cascading water. There also is a corpse on the far end of the train cart with a magnum, and a scare effect when they approach. The ceiling has partially collapsed in the next area, causing water to rain down and flood the area. The waterfall that covers up the door, allows zombies to come charging through and surprise the players, because it obscures the area beyond it. Finally, the players head into a small tunnel towards the abandoned train station. A green light directs the player’s attention to this tunnel. Other corners of this area were left intentionally dark, because players naturally move towards the light.  Notice the blocked area on the map. This area initially was a direct connection to the finale, but I decided to make the map longer. More on that down below.

The players enter through a dark corridor, with wooden boards covering up a hole in the left wall. I deliberately set up the lighting in this tunnel to create eerie striped shadows from the boards across the right wall. This adds mood to the area, but also sets the tunnel apart from the mirrored tunnel on the other side (see top right of map).

The abandoned train station was the last area I added to the map. In the initial design, the players could walk straight from the flooded platform, to the bridge gauntlet. That area is blocked off now by debris. I felt the map wasn’t long enough, and needed a resting area before the finale. So I decided to create a rather large room where players could stock up on supplies. The room can extend the length of the map depending on the strategy of the players. You see, in Versus mode the survivor team generally wants to move through the map as fast as possible, cutting corners when ever possible. Because the longer they take, the more time they give the Infected team to respawn and attack. So if the players play smart, and conserve ammo, they may not need to stock up here. So they could skip the area and take the short route past the ladder. However, the shortcut does not have any railing, so a Smoker could drag a survivor off the raised area causing a lot of falling damage. If they take the long route, they can stock up and properly prepare for the finale, but this also allows the Infected team to respawn and attack more.

Visually, I wanted this area to look very large, yet also dark. I was inspired by the underground railroad in the movie Ghostbusters 2, which had a river of slime, and the same purple lighting as the area I made. I was also inspired by the abandoned train station in New York. I didn’t want to give away too much of the mystery of this area. And the fact that a large part of the lower area cannot be accessed, and is shrouded in fog, makes this a very sinister and mysterious room indeed. The players can never quite grasp how big it really is, and how far the tunnels in the distance go. In reality, they don’t go far at all beyond what is visible. But the area suggests that it is much larger than it really is. The stairs that lead upward, but cannot be accessed, suggest that there used to be a logical entrance to this part of the station (until the tunnel collapsed).  I always want my maps to seem logical, so I make sure to include doors that may just be purely decorative, but suggest that the area could have actually been in use at some point. Everything needs to have a logical use. Which is why there are a lot of blocked or collapsed passages throughout this map. It makes the player believe that the station is part of a large complex, and not just a lineair level.

The bridge gauntlet is the finale of this map. A gauntlet in Left 4 Dead 2, means an event where hordes of zombies spawn infinitely until the players reach their goal. Balancing a gauntlet is tricky, because it is easy to deplete the players of ammo, or simply wear them down. This is why I made the gauntlet short. A button causes four gates to rise slowly, giving the ai director time to put the horde into motion. Because once the gates are open, I want the zombies to be ready to attack. The players can choose between two lanes. The left lane is free of obstacles, and the right lane has two jumps across pits (which makes a Charger attack exceptionally dangerous here). These drops are not deadly, but they do force any victim to backtrack a bit. To avoid forcing the players to head through the collapsed tunnel and the abandoned train station again, a door in one of the trains is automatically destroyed by the horde as soon as the gauntlet starts (thus opening up a shortcut back to the bridge). This makes the pitfalls less punishing. A tank always spawns at the far end of the bridge, and will charge down one of the two lanes of the bridge. The players may need to quickly change to a different lane to avoid the tank if they don’t want to fight him.

What makes the two lanes interesting, is that they are covered in barb-wire and metal plating. So if the team splits up and takes different lanes, they will often not have a clear shot at the other lane. This makes rescueing a team mate really difficult when splitting up. Only occasionally does the plating open up, to reveal a glimpse of the other lane.

The area after the bridge has a nasty separation point that is intended to be exploited by smokers. A smoker can drag a survivor down from the ledge here (which is not protected by a railing) just before they reach the saferoom. At this point the walls in front of the players also crumble, revealing a fire covered hellish maze of rusty corridors from which countless zombies spawn. The players must quickly run into the back cart of the bionic train, which will take them to another world in the next map.

One final note; The area in the top right corner of the map is an easter egg area (or bonus area), which is shown at the end of the video linked at the top of this article. It is not needed for the players to go here at all, and it can only be reached through an airduct that is hidden underneath the stairs. I wanted to add an extra area, because it adds to the sense of mystery of the map. The bonus area leads to a creepy well in a cave, with a terrifying and deadly surprise for the players.

http://www.youtube.com/watch?v=SJjFctNQkhk

This is the second map in the Demonbreed campaign for Left 4 Dead 2.  Keep in mind that this video features a place holder for the train model. The parts of the train that the player walks through (including the final saferoom) are represented by a yellow and black striped box, which will later be replaced by a fancy model of a train cart. The goal with this map, was to offer a vastly different experience from the previous map. The first map had a lot of open areas, with occasionally some indoor areas or a tight tunnel. The second map however is all indoor areas, full of tight tunnels, with occasionally a large area. The biggest challenge was to make the subway look and feel different from the subway in the popular official map No Mercy. So what I did was decorate the subway with the hell-motive, which was already faintly present in the first map. This meant adding lots of barb-wire, graffity, blood, rust and decay. The subway had to look like a hostile place, even without the zombies. I also added a lot of small reveals of other inaccessible areas (usually through gratings, and holes in walls and floors) to give the impression that the subway is a lot larger than it really is. I really wanted this tight area, to look less tight. So I created lots of underground tunnels and extra rooms that you would see behind walls, but never visit.

New in this map are the scare stingers, which I also added to the first map later on. Basically when ever the player gets really close to a frightening sight (such as a corpse) the game plays a scare sound to startle the player. I wanted all the gruesome imagery to have a bit of impact on the player, and sound and music can highlight that. Another important step was to beef up the difficulty slightly and to add slightly better weapons, since this is the second map in the campaign.

One of the most important areas is the area with the train platforms. I initially created the full area, mirrored on either side. So the platforms on both sides were basically just mirrored copies of each other. The second platform was originally just intended as a backdrop, while the players ran across the first platform.  Just going through this area from one side to the other seemed like a big waste considering all the decoration, and the campaign also seemed too short. I made the path of the players zig zag through the room, so they visit both platforms. The problem with that, is that it blatantly exposes the fact that both platforms are mirrored. It could also end up confusing the players, since it would be hard to tell one area from the other. That is when I decided to make the other platform radically different. The roof has collapsed, causing water to leak through, and now the platform is almost entirely flooded. Visually this looks really good, because the water cascades down the train carts like waterfalls. I also made the falling water obscure one of the doors, so it is harder to see zombies run at the train from the inside. When ever I add such a unique feature in the look of the level, I want to exploit it to the maximum and get everything out of it that I can.

This map has two important panic events (the events that cause waves of zombies), the falling train event, and the gauntlet across the bridge. The falling train event is unique, because the room goes dark, and the players find themselves locked in a very small and stressful place to have a battle. There is not a lot of movement room, and to top it off, an entire train threatens to fall on top of them.

The bridge gauntlet seemed like the logical conclusion of a map in which the train platforms play such an important role. The players finally reach an area they saw much earlier, but now they must choose between two paths. The bridge has two sides and only occasionally do the metal plates and gratings offer a view of the adjacent path. So if the players split up here, they might find themselves ambushed by special infected and unable to be rescued by their friends in time. On top of that, a tank also charges down one of the two paths, forcing the players to alter their strategy.

The abandoned train station was a last minute addition to the map. I really wanted a slightly eerie silence before the final gauntlet, where the players could stock up on ammo. This would also add to the length of the map slightly. I was inspired by pictures of the abandoned railroad in New York. There is just something really odd and creepy about an old railroad that is no longer used, yet looks radically different from the rest of the underground. To make it even more different, I gave it a spooky purple light, much like the abandoned railroad in the movie Ghostbusters 2 (which features the river of slime). I made the area really big on purpose, but also very dark, so you would only see parts of it. It made the room look larger and more mysterious if it wasn’t lit too brightly. The platform and the tunnels of the old trainstation itself are only faintly visible through the fences, and are shrouded in fog.

The final area seen in the video is an optional secret area. It is an easter egg with a frightening surprise. It takes a lot of exploring to discover how to reach it, but I like having a bit of mystery in some of my maps. It is quite startling to the players to discover such a vast and spooky area hidden in the campaign.

http://www.youtube.com/watch?v=-2UUEpg9Ei4

Related articles:

Designing a left 4 Dead 2 map – part 1

Designing a left 4 Dead 2 map – part 2

Designing a left 4 Dead 2 map – part 3

Designing a left 4 Dead 2 map – part 4

After years of working with Unreal 3, I wanted to catch up a bit on the Source engine. I used to tinker a lot with Half-Life 1, but Valve’s engine has gone through lots of changes since then. When designing this level, I wanted to improve on things that bothered me about the official maps that are included with the game.

• Efficient use of decoration, reusing previous areas as background for later areas
• Clear feeling of level flow from one level to the next, so they seem to connect well
• Starting in a secure area, and building up the feeling of danger slowly
• Multiple roads
• Good variation of area sizes and height differences
• A limited number of set pieces and making good use of them through out the level
• Proper lighting, so players are never caught in total darkness
• Proper leashing, so player never get lost
• Limited but strategically placed insta kill spots for pvp
• Random events

Let me adress these points one by one. Efficient use of decoration has always bothered me about the original maps in Left 4 Dead 2. Most single player maps seem to be one long line with unique decoration along the way. While it looks awesome, its a waste of resources. If you create one big set piece and have the players look at it from multiple angles through out the map, you get much better use out of your hard work. Thats what I set out to do with the giant chasm in this map. It is the heart of the map, and you pass it multiple times from different sides. This is not only practical, but it also makes it easy for the players to identify where they are. They don’t get lost quite so easily, because everything feels like a coherent whole.

The connection between one level, the safe room, and then the next level is something really important that really doesn’t work well in the original maps of the game. After loading the next map, it feels like you’ve arrived in a whole different world. I wanted to tie the levels more closely together, so I made sure to include parts of level 1, in level 2 and the other way around. I want players to feel like the next level is still part of the same journey.

The way this map starts is pretty unique by Left 4 Dead’s standards. You start in an area that is completely safe right from the start, and you are confined in a building. There are a few zombies inside the building, but overall it feels as if the players are safe from the horrors that await outside. I wanted players to slowly become familiar with their surroundings. Their first objective is to explore just the tv store they are in, and find their way to the roof, to activate the power. Then they must return downstairs, and open the shutters, so they can go outside. This little detour to the roof allows the players to get a good look at the entire street outside the building and memorise the lay out. The powerswitch is deliberately placed so that it overlooks the garage, which is the next goal of the players.

The map is mostly linear, with just a few alternate roads that branch off near the middle of the map. I wanted the players to have a false feeling of freedom. Because if it was completely open, I would have no way to control the flow of the map. The alternate roads present a little more danger to the players, but with possibly higher rewards.

What makes a good map, and what’s make a boring map? In my opinion, its all in the variation of the areas. I tried mixing up the sizes of the areas as much as possible, starting from a medium area, that open up into a small area, which leads to a big area. You really want to make sure that it doesn’t all just feel like a bunch of identical square rooms stuck together. Create height differences. Have the player climb stuff, or descend down a couple of stairs. Have the player on top of a roof top at one moment, and knee-deep in sewage water at the next.

I made sure there were a couple of memorable set pieces in the map. Things that the players could easily identify, remember, and use to orientate them selves with. Usually when you start designing a map like this, you try and think of the first things that come to mind: A tv store, a garage, a hotel, a restaurant, a subway. I then spread those landmarks out to cover all parts of the map, and the rest would be filler.

Regarding lighting, I could probably write a whole article about how important it really is. One of the players that was to test the map, urged me to make it really dark. Because darkness + flashlight equals scary. While it is good to gather feedback from actual players, you do have to filter the feedback you get. And some of it should be completely ignored. Lighting is one of those things. If they say the map is too dark, it probably is. But if they say it needs to be really dark, its better to follow your own judgement. Total darkness can really ruin a good map, because players lose their sense of direction and get frustrated. I always try to light the areas I want players to go to properly, and make the dead ends rather dark. This causes players to automatically be directed in the right direction. Especially when there’s multiple roads, make sure you light the scene well. You want players to see all the options. Most players probably don’t realise how much lighting is involved in maps that they consider dark and creepy. Even if you have a night map, use lots of lights! Don’t have your players fumbling around in total darkness.

Leashing is one of those terms most people, even in the game industry, aren’t familiar with. Leashing is all about leading the player in the right direction, without him noticing. For example, the player arrives at a junction, and one direction is a dead end. How do you get him to go in the correct direction? And should he take the wrong turn, how do you quickly get him back on track, without him accidentally retracing his steps? In Left 4 Dead 2 they use a rather cheap trick for this: They paste decals of giant arrows all over the map. Its not a very elegant solution, although it does work. But there are other ways to do it. Lighting, as mentioned before, can help put focus on key areas. Landmarks help players tell one area from the next. And even the direction of furniture can lead the gaze of a player in the right direction. There could perhaps be a blood trail that leads down one corridor, or a row of bodies that goes around a corner. When decorating the submerged pedestrian tunnel in the first map of Demonbreed, I added lots of wall graffiti near the start of the tunnel, and less near the end. This made it easier on the eye to tell one direction in the tunnel from the other. I also had the characters comment on the tunnel when the player got close to it, to indicate this was the way to go. Leashing is one of the things I consider most important when designing any of my levels.

Spots where you can instantly die can be really frustrating. I know that in Left 4 Dead 2 more than just a handful of pvp maps have unintended insta kill spots. These are also features that can make a map a lot of fun, but I would prefer they are in limited amount and deliberately placed. With Demonbreed I tried to build spots like this on purpose, keeping in mind how the infected team might abuse them, and how the survivor team might prevent instant death. The best example is the hole in the wall on the rooftop of the hotel. A Charger can spawn behind the door, and charge a survivor straight off the roof. But only if the survivor is foolish enough to go for the items that are placed in front of the hole. The spot can also be easily avoided if players leave the items be. Insta kill spots can really unbalance a PVP match because they are so decisive. So I decided to include them, but made sure there were few, and that they were spread out well. This also holds true for the second map in the campaign.

One last thing I wanted to experiment with is random events. If players were going to play this map in pvp sessions, I wanted to have some degree of unpredictability that would keep the maps fresh and exciting on each play through. I constructed the random events in such a way that they would never tip the favor for any of the teams in multiplayer. They were more like small scares that would occasionally happen.

Part 1:

http://www.youtube.com/watch?v=awgMGDJYFsE

Part 2:

http://www.youtube.com/watch?v=ODewJCxSurk

These story levels were originally intended as DLC. I think they are far better than most of the levels in the released game. Alas, this DLC never got released. It features cut scenes that are as good as finished, has unique weapons and enemies, and an epic boss battle at the end. I designed most, if not all, of the hazards in these levels. Our art department would also often write down ideas for the levels, which we would sometimes elaborate on to make new obstacles. The tree that drops snow when struck, is one of those hazards. It can be very helpful to freeze enemies and shatter them in an instant, but players must be careful not to get frozen themselves. I also love the snowman prefab. Most of the snowmen drop a weapon when destroyed, but are completely harmless. But on occasion there is a snowman that spawns an enemy, to scare/surprise the player. You can see all of our prefab prototypes as well in this video.

The way we designed these levels is quite different from how we designed earlier levels. What we did, was simply design a lot of small bits and pieces of game play. This allowed us to pick and choose the ones we liked best, and then stitch them together to make a whole level. Of course this meant that a lot of designs got tossed out as well, but we also didn’t spend a whole lot of time on them either.

You’ll probably notice that a whole lot of weapons and enemies have a long string of weird text attached to them. This is because their descriptions were not implemented yet at that point (and they never were).

My only real gripe towards this approach of level design, is that it seemed we once again fell back to making a ton of unique backgrounds. I prefer laying out levels in the form of a circle, or a zig zag pattern. So you get to use earlier parts of the level as background for later parts. I simply find that more cost efficient. For a good example of that, look at the Demonbreed level for Left 4 Dead 2 that I made.

http://www.youtube.com/watch?v=sg4YXlwNKxs

The Mines level was an experiment of the studio to see what would happen if a team of one level designer (me), one level decorator and two modellers were given completely freedom to make a level. The only guideline was, it had to have an underground mine theme, and one or two puzzles had to be present. At this point the studio was unsure if the current workflow was working correctly, and if they wanted to continue working that way for future projects. So this was a test, to see what a small team focussed on one level could accomplish.

When designing this level, I wanted simplicity. A bit of platforming, a puzzle and a lot of fun fighting. We knew from experience that we didn’t like creating lots of backgrounds, so we decided the layout would only need one simple background that would suffice for the entire level. I prototyped two puzzles (one which we actually kept and finished), and various prefabs with their own unique functionality. All of which I believe can be seen in the video. We made TNT barrels that would explode. Crates that could be destroyed and spawned coins. Lanterns that responded to getting hit. A bucket that can drop helpful items when hit. And a special fence that would trigger a slow-mo death for any enemy that got knocked into it. We also made a working catapult and a bucket lift (which was a pain to animate by hand in Unreal 3′s Kismet system). The puzzle worked as a simple game of simon says. It would show you a sequence of colors, and the player then had to remember and jump on the same colors.

Regarding the fighting, some of it could get pretty tedious quickly (especially in the released game). So in this level we offered lots of clever ways to quickly kill off enemies in fun ways. Such as blowing them up with TNT, hitting them with catapults, knocking them through wooden fences and into chasms.

Our modeller really went all out with this level, and especially the puzzle. We put aside the strict art style for a moment, and allowed ourselves to add more detail to the various props.

I’m especially proud of the way some of the fire turned out near the end of the level. The player really gets a feeling of danger and urgency once the fire starts spreading. I added a timer that spread the fire a bit further if the player took too long, just to scare the player. It never really engulfs the whole level, but it feels like it might at any minute. Even if this was just a test, we all learned a lot from this experiment. It really was a good excercise in setting up a good workflow for designing and building a level from start to finish.

http://www.youtube.com/watch?v=BXRSlkuCtnk

Apart from the extensive story campaign, Fairytale Fights also had a few PVP levels which were rubbish. They were rushed out with the game, and we all knew that we could do a whole lot better. So with DLC in mind, we decided to design many new PVP levels that would pit player against player in incredible ways. Many of these PVP maps would start out as a simple prototype (which we referred to as block-outs). One such block-out is the Babeltower level, which later became the Flood PVP level. In the video you can see the unpolished prototype and the fully decorated final level. The final level is radically different from the prototype, and only retains the original concept of rising water. This often happens with prototypes, where only a part of the prototype is lifted out to be turned into a whole new level.

The Killswitch level was originally designed by a different level designer, but it wasn’t working as a PVP map. The idea was great, to have a toggleable death zone that players would fight about, and use to kill each other. I realise that sometimes a level designer can get stuck when looking at the same design for too long. In such cases, it can help to have a different level designer come in and do an alternate take on the same concept. And that is what I was asked to do. The producers loved the concept, but the lay out wasn’t working. So when I redesigned it, one of the most important things I changed is the way the death zone works. In my redesign the death zone only covers a limited area of the level (instead of the whole level), and the best weapons are located in that zone. By hitting the tail of the dragon, this triggers his deadly fire breath. But it also works both ways. If a player hits the dragon in the head, this will trigger a sweep of his tail. This ensures that a player never dominates the game entirely by constantly spamming the firebreath attack. I also created two large flat areas to fight on, with a thin bridge and a jump in between. This lead to some very intense fights, mostly centered around the large left platform.

For the Clocktower PVP map, I was asked to fix various serious bugs regarding the collision and death zones. I think this is probably one of the most fun PVP maps to fight in. One of the things I had to fix was the way the elevators can crush players against the ceiling and the floor. There were also spots where the players could get caught on a wall, due to odd collision boxes. The Clocktower level is a great example of how a simple idea, no bigger than one screen, can result in excellent game play. Makes me wish I designed it myself, but it was designed by someone else.

For the Christmas level, I worked on getting the teleporting to work correctly, as well as the collision and the bridge-launch mechanic. The bridge in the middle of the level can crush a player, but also launch him into the spikes, or on top of one of the platforms. Getting the targeting to work correctly was difficult, because often the priorities regarding aiming for the lever or aiming for an opponent got messed up. Physics are an unpredictable factor here. The player is basically knocked with an unpredictable arch from the bridge towards the spikes, and I had to some how make this work in a reliable fashion. That took a lot of tweaking.

The Waterfall level had serious issues with collision and physics. Items are spawned on top of the moving books constantly, but often they’d roll into the water before the player could get them. And on not so rare occasions, the items would get stuck in mid air, which looked really odd. This was due to the code turning off physics calculations as soon as a physics object came to a halt. All the weapons in Fairytale Fights use physics, which is why you can throw them, and pick them up again. But all those physics calculations could get rather heavy when there are lots of objects on screen. This is why certain limitations were added in, but this can lead to unintended side effects.

http://www.youtube.com/watch?v=BFRNNyIaNOk

When designing levels for Fairytale Fights we were always looking for new interesting obstacles, and types of new game play. I made a ton of prefabs, many which made it into the actual game fortunately. We would basically create these demo levels that would show all our gruesome creations in a row and allow other level designers to copy them into their own levels. It also allowed the model department to quickly see what hazards (as we called our deadly contraptions) needed model work. We would usually start out by using placeholder blocks and effects, and try to make the prefab work in such a way that its intended function was clear. In some cases, the placeholder effects almost looked good enough. But in other cases, unique effects and models would have to be created to make the prefab look good.

All of the prototypes in this video were designed, scripted and made by me. Some were based on ideas suggested by our lovely art department. We would occasionally just let people write down their ideas, and then we’d pick a few that we liked best. Some of the prototypes in the above video add a whole different kind of game play to the game. We were always looking for ways to add more fun into the game, since the game play in the game itself was a bit lacking.

As you’ll be able to tell from the video, though many of the prototypes are unpolished, all of them are working. I highly encourage other level designers to take this approach when designing new game play for levels. Quickly prototype new forms of game play and pitch them. Let the whole studio have a go at your creations and see what they think. Maybe one silly idea will lead to one briljant follow up idea.

Giant 1:

http://www.youtube.com/watch?v=CUlr7JjTnJM

Giant 2

http://www.youtube.com/watch?v=pBkdIpaHXx0

The giant levels were made pretty late in production and they are without a doubt among the best levels in the entire game. During early development of Fairytale Fights the camera system gave us a lot of trouble and many of the game play mechanics were undefined. With the giant levels however, everything was nearly ready. It was all just a matter of finishing up the final levels. Giant level 1 was always meant to be a giant beanstalk (which also ties into the backstory of one of the playable fairytale characters, Jack from Jack and the Beanstalk). I looked for inspiration in other games that had attempted to create a giant beanstalk, and found that most of them were extremely frustrating. I realised this was probably due to many beanstalk attempts being a platform nightmare and platforming wasn’t Fairytale Fights’ strength either. On top of that, the Fairytale Fights camera system was designed with horizontal game play in mind, and really had issues with large vertical movements. So the question then became, how do I turn a vertical level into a horizontal level?

I tried to eliminate as much platforming as possible. Obviously there was going to be some jumping, but not a whole lot. Instead, I designed it to be mostly about fighting and walking. Rather than going vertically along one side, I wanted the players to see the beanstalk from all sides, and get a real sense of the height of the thing. So what I did was, I layed out a path in the form of a giant vine, so that it swirled around the central beanstalk. As the player progressed further upward, the camera would slowly start to point downward more and distance itself more from the player. Then when the player reached the top, it would show the full size of the beanstalk, but not the top part. The beanstalk was meant to go up into the clouds, to the house of the giants. Obviously this part was going to be part cutscene and a new level. So I made sure that the camera framed the beanstalk  in such a way, that the top part was never visible (and there wasn’t even a top part, the decoration just ended beyond what was visible).

Getting the collision and camera work done for this level was a nightmare. Fairytale Fights has left and right boundaries that prevent the players from separating too far. We didn’t ever want players dying or getting stuck off screen. However, such boundaries did not exist for the top and bottom of the screen (we tried, but it didn’t work right). Also, to prevent players from getting stuck, the game automatically kills and respawns players if they are off screen for too long (a crude but effective fail safe). So it was quite a painful piece of work to get the camera to follow the exact path so this situation would never come up.

On top of that, if that weren’t tricky enough, the game also has a rather difficult respawning system. We basically had to manually place all the possible spawn spots for players. But with a vertical level, players are able to jump off and die at almost any spot. And what the code does, is try and find the closest spawn spot to where the player died. For a horizontal level this works fine, but with a vertical level this means the player could respawn on top of a lower area of the beanstalk. As with all levels, we occasionally close off parts of the level as the player progresses, so we can unload those bits of the level and free up some memory. We can there for not have the player respawn in unloaded or blocked off parts of the level, because then the player would get completely stuck. So this meant that as the player progresses, we had to manually turn off various spawn spots all the time.

http://www.youtube.com/watch?v=Acq-XzQEEcQ

The first level of the game Fairytale Fights is also the first level I started working on when I joined the Fairytale Fights team. When designing this level, I focussed a lot on making the level a slow introduction into the basic mechanics of the game. The first couple of screens of the level are without hazards, so players can get comfortable with the controls. Then slowly weapons are introduced and the so called glory attacks (super attacks that the player charges up by fighting). I personally dislike introductory levels that immediately confront the player with deadly pitfalls, which is why pitfalls are completely absent in this level. Level progression is linear, due to the limits of the camera (it can only move along a track). I would have liked to have more puzzles in the game, like the waterwheel mechanic early in this level. But sadly due to time constraints we didn’t have time to implement that many. The waterwheel puzzle is one of the puzzles I completely designed from scratch. We were trying to come up with clever ways to have multiple players work together. For this purpose, the waterwheel allows multiple players to walk inside it to raise the gate.

Regarding the art style, many of Fairytale Fights’ levels were made to look as if the world was made out of books and paper. Also all of the characters were designed to look like designer toys. This is why a lot of the grass looks like giant squares. In fact even all of the effects, such as water splashing and explosions, were made to look as if they were cube shaped. Almost like they are made out of clay. Fairytale Fights was also the first game to include dynamic slicing, which allows the player to cut through the body of an enemy along any axis multiple times. The game then draws the inside of the character, including the skeleton. Blood is also dynamic. As it sprays across the floor, the player can slide through the blood, making quite a mess in the process.

Lots of assets were created to decorate these levels, and the lumberjack levels might have had more unique assets than any of the other levels. Some of the assets were created in several parts. For example, with the trees the trunk and top were seperate. This allowed the level decorators to switch the two elements around to create many unique looking trees. That is why in Lumberjack level 1, hardly any tree looks like the next one.