Clever Level Design: More than Meets the Eye

Environments and stories. Pete Ellis (pete-ellis.com) from Guerrilla Cambridge knows all about these things. In this interview he gives some very nice hints as to environment design, shows examples of virtual world as a story medium usage and explains how you can teach user to play through clever levels.

Introduction

Hi, I’m Pete Ellis from Birmingham in the UK, currently living in Cambridge. I’ve been a designer in the games industry for nearly 8 years and I’m currently working at Guerrilla Cambridge on ‘RIGS: Mechanised Combat League’, which we’ve built from the ground up exclusively for Playstation VR. It’s a first-person arena based shooter set 50 years in the future, where you get to pilot giant Rigs and compete in the sport of the future; the mechanised combat league!

RIGS: Mechanised Combat League

Previous to this I worked on ‘Killzone Shadow Fall’ for the Playstation 4 where I created a free DLC map called ‘The Stormgracht’. That was a multiplayer map in which we created this cool player-controlled set piece where this huge torrent of water floods the central gameplay space, killing anyone in its way and temporarily blocking routes.

‘The Stormgracht’ map in ‘Killzone Shadow Fall’ with its flood gameplay event

Prior to that I worked on ‘Killzone Mercenary’ for the Playstation Vita where I designed sections of the single player campaign. I worked alongside other designers on a number of levels, but my main focus was creating the opening level; ‘Halls of Justice’. In this level the player and the buddy character Ivanov infiltrate the Halls of Justice building, which had been commandeered by the Helghast, in order to rescue Admiral Grey. This level set up the story beat of seeing the antagonist Colonel Kratek, as well as being the tutorial level for learning the game’s controls. I had a lot of fun working on that!

The opening level of ‘Killzone Mercenary’

Before moving to Guerrilla Cambridge I worked at Frontier Developments for two and a half years on a third-person open-world game called ‘The Outsider’, but unfortunately that got cancelled and it never saw the light of day. Before working in games I did a Degree and a Masters at Teesside University in Games Design and Games Art respectively.

Tutorial Levels

When you teach people how to play your game it’s important to make sure they fully understand what you’re trying to teach them before they move on and progress. In order to maximise this understanding it’s important to introduce elements in isolation so that important information isn’t missed and so players only focus on one task. For example, if you’re trying to teach the player how to jump and use the game’s traversal mechanics, you don’t want to also have enemies involved that will distract the player.

Generally, I use 3 stages of tutorialisation; Exposition, Validation and Challenge. This is that you show the player the thing you will be teaching them, you let them experience it and use the mechanic themselves in a safe area, and then you give them a challenge to prove they understand how to use it.

In ‘Killzone Mercenary’ for example, when we taught players how to aim and shoot using the twin sticks and the new control scheme on the Vita, there were multiple stages. Firstly, they encountered enemy AI in a small arena that wasn’t too challenging as the buddy AI was also alongside to help. We showed the button icons on the screen at the start as the player needed to know which buttons to use, but this area didn’t require the player to necessarily aim down the weapon’s sight.

After this first encounter we reinforced the learning of aiming down the weapon’s sight so that the core mechanic of shooting became muscle memory. This was done by adding a challenge of shooting a security camera that was blocking the path, which meant that if the player missed a few times it didn’t matter as it was a static object. The player is reminded of the button to use here with a prompt, but after that we moved them on to targeting enemies. This is the same mechanical challenge but it’s an ‘invisible’ tutorial to the player as we don’t hint at which buttons to use anymore. Additionally this time missing the shot has consequences of alerting the enemy into combat mode.

The last parts of the ‘aim-down-sights’ tutorial in ‘Killzone Mercenary’

In terms of physical world building, tutorialisation can also have ‘hard gates’; places in the level that once you pass you can’t return. For example learning to climb means you can get out of the area, but until the player has mastered this mechanic they can’t progress. This ensures they know how to use the mechanic at least once.

Levels and Gameplay

There’s a design mantra coined by the architect Louis Sullivan, ‘form follows function’. This means that how a level is constructed and shaped should stem from its intended purpose. For example, a ‘Gears of War’ combat encounter needs to support the cover system for the player, buddy characters and the enemy AI to work, and thus needs geometry that creates this. You will see in most encounters there are a variety of low and high cover points that are fairly close to one another.

Levels in ‘Gears of War’ feature low cover to support the cover system for the player and the AI

The arrangement of combat spaces where the player fights against the standard enemy types is different to when they fight distinctive boss characters, such as the Berserker. This enemy doesn’t use cover and instead charges around crashing through pillars in an attempt to run the player down. Therefore, the geometry in these areas has no low cover points and plenty of space to allow the Berserker to navigate at speed, and the only high cover for the player and the buddy characters are the tall pillars that the boss can crash into. The arrangement of geometry here is different to support the gameplay requirements.

Levels with the ‘Berserker’ don’t have any low cover to support the navigation of this boss

Although form should follow function, this doesn’t mean that design trumps art. It is important to get the artists involved from the very start of a level’s development so that the two disciplines can work hand in hand and will complement each other well. For example, if you have a design where you want the player to discover a certain route that’s off the critical path, the art team can help this intention with a variety of composition and lighting tricks to help draw the player’s eye and guide them. I like to work alongside the artists from the very start as their vision also informs my designs – it was the great work of Senior Concept Artist Pete Spence that was my inspiration for the flood in ‘The Stormgracht’.

The torrent of water in ‘The Stomrgracht’ floods the entire central gameplay space

As for how levels and gameplay mechanics work together in tandem when there are multiple styles of gameplay, such as stealth or action, this is about incorporating all the elements together in a readable way. Each element should be easily readable to the player as to what its purpose and use is. For example in ‘Metal Gear Solid’ they have open ‘prospect’ spaces which supports gunplay and action and smaller ‘refuge’ spaces that conceal the player and allow for stealth. Both these elements are clear to the player through environmental influence such as lighting and shadows, but also by maintaining the pattern of their size and shape throughout the game. The player can therefore identify these areas within a level and plan out a strategy for their progression.

Refuge spaces in the ‘Metal Gear Solid’ series provide hiding spots for the player

At a lower level elements need to be readable as well; that is not just the general space being readable, but the individual elements that the player uses as well, like cover points. These need to be unambiguous, like how the ‘Gears of War’ cover geometry is flat surfaces at consistent heights. Imagine if, as a player, in the middle of a battle you ran over to something that you thought was cover, only to find out it’s only ankle height and as a result you get gunned down instead. That would feel particularly unfair and frustrating, not the kind of experience you want to be creating.

A ‘Gears of War’ level has consistent, readable low cover throughout

Environment Storytelling

I’m a big fan of using dioramas (also referred to as vignettes). This is arranging props and objects to tell a story within the environment. You don’t tell the player exactly what has occurred, but they get to interpret what has happened. The human mind is excellent at piecing together missing information to form a better understanding, so leaving some ambiguity allows for the player to invent their own stories.

For example, a scene with a window half open, an empty coat stand that’s been knocked over, and a still smoking cigarette can suggest a few things; was there a person here recently who saw something outside the window, grabbed their coat quickly and left in a hurry? Or did something enter in through the window meaning the person had to escape quickly?

A diorama I created in the opening level of ‘Killzone Mercenary’ where a hostage has been shot down whilst running from the turret in the background. Two enemy troopers stand over the body discussing what happened, not only providing exposition but also highlighting that the turret is a weapon to be avoided:

It’s more impactful for a player to piece together an understanding of a scene and figure out what has happened rather than just being told; self-discovery is more rewarding. The screenwriting mantra ‘show don’t tell’ is key here.

The strongest dioramas are ones that echo the game’s story and they are great opportunities to support the game’s narrative. It’s hard not to mention ‘Bioshock’ when discussing environmental storytelling as that is a great example of dioramas that support the world narrative. For example, there’s one diorama surrounding a money pickup that is contextualised by a Splicer who has attempted to steal from an ATM and has been crushed by it. This not only evokes questions from the player about their own actions of stealing, it also enforces the theme of Rapture having societal decay. In addition, showing ‘cause and effect’ in dioramas also helps the readability of what has occurred before the player arrives.

A Splicer has been crushed by an ATM machine when trying to steal from it, contextualising a money pickup in ‘Bioshock’

Dioramas can not only tell stories inside a game world, they can also be used in other ways. Future events are foreshadowed when the player walks through a group of petrified people before getting to the Medusa in ‘God of War’ and seeing a burnt body on a fence in ‘Half Life 2’ tells the player the fence is electrified without the need for any pop-ups, text or UI instructions.

A burnt body tells the player that the fence is electric

The most powerful environmental storytelling I’ve ever seen is during the ‘Ish’ secondary storyline in ‘The Last of Us’. The player learns that a group of people have been hiding in the sewers protecting their children. It becomes apparent that the infected got inside the camp and there was no way out for these people. The player enters a dark room with the words ‘they didn’t suffer’ written on the floor, and four small bodies under a cloth at the back of the room, next to an adult who had committed suicide. This storyline is incidental to the main story and can be missed as it’s not on the critical path. There is a build-up through discovered written notes but the ending diorama can only be seen with a torch, and is delivered with only one line of dialog from the main character, Joel. You don’t see the actual actions of what preceded (which considering the morbid context is a relief) which reinforces the ‘Hitchcockian’ theory that the monster created in the mind is worse than anything you can ever show.

For me, this section was so profound that I had to put the pad down for a week and go away and reflect on it! It made me as a player ask the questions of what would I do in that situation and how would I have coped; creating that ambiguity by not showing the direct actions was far more evocative.

The morbid conclusion of the people trying to protect the children in the sewers in ‘The Last of Us’

Making the Level Understandable

Player guidance is a huge, huge subject in itself which I could write a whole essay on, so instead I’ll talk about a few examples. I’m a big fan of guiding the player without resorting to on-screen icons or UI and to do this it’s all about constructing your levels to give the player motivation. The player will have an overarching goal or mission objective, but the arrangement of the environment can infuse smaller immediate objectives as well as main aims.

Using landmarks is an important design consideration for enticing players to head towards its direction. A landmark (also referred to as a ‘wienie’ by Disney Imagineerers) is a geographical location that acts as an invitation to continue the journey by stimulating curiosity. As John Hench describes them ‘a beckoning hand promises something worthwhile; its friendly beckoning fingers say: “Come this way. You’ll have a good time”’.

The Pittsburgh bridge in ‘The Last of Us’, a beckoning landmark

Reward and denial sketch from ‘101 Things I Learnt in Architecture school’

To make your landmarks even more enticing you can use denial spaces. This is an architectural concept whereby designing paths that momentarily lose sight of the landmark or target you create a more interesting journey through intrigue, which in turn causes more rewarding arrivals. This is especially important for level design because if you showed the player the target, then all they had to do was walk forward to reach it, that wouldn’t be a very interesting level.

Additionally, landmarks are a good way of reminding the player what their goal is after playing sections that would have distracted their attention and focus, such as combat. The ‘Uncharted 4’ Playstation Experience video is a great example of this; the jagged mountain is revealed at the start to show the player where they’re headed, again after an interior section inside caves which blocked the view to it and then lastly towards the end of a combat encounter where the player would potentially be the most disoriented.

The jagged mountain in ‘Uncharted 4’ throughout various stages of the level

Composition is also extremely important for subtly guiding the player the way you want them to go. This includes framing routes to draw the player’s focus, as well as using leading lines within the environment. For example the ridge of tiled roofs during the rooftop chase in ‘Uncharted 3’ points the player along the escape route.

Leading lines in ‘Uncharted 3’ point the player towards the escape route

The use of movement as a tool for guiding the player is also extremely effective. An example might be to have birds flying towards the spot where you want the player to go. The idea being that the movement catches the player’s attention and leads their eye to the destination. Another example is the police cars in ‘The Last of Us’ driving past the window in the opening section, again leading the player’s eye in the direction they need to go.

Police cars at the start of ‘The Last of Us’ draw the player’s eye in the direction they are travelling; from left to right, which is the direction the player should travel

The last example I’ll touch on is a more explicit method; telegraphing what the player can use through a visual language. Having a consistent colour scheme means the player understands the pattern of each interactive object and can thus read their use just through identifying the colour, such as the yellow climb ledges in ‘Uncharted’ or ‘The Last of Us’. A more extreme example is ‘Mirror’s Edge’ where they integrate colour for interactive elements into its otherwise white environment, meaning the colour association is easily noticed during speedy traversal.

Yellow ledges denote climbing in ‘Uncharted’ and ‘The Last of Us’

Colour as a visual language in ‘Mirror’s Edge’

Using Huge Structures

There are different ways to use scale in level design depending on what your emotional intention is. For example, one of my favourite games, ‘Ico’, uses a huge castle as the game’s setting and location and so the player is always seeing big landmarks of different sections of the castle in the distance and the open vistas. The accompanying audio is a really powerful tool here as well, as instead of music there is only ambient noise; many times being of just wind to emphasise the huge scale of the empty castle.

The setting of ‘Ico’; a huge, empty castle of epic scale

It’s most likely that your game isn’t set just within one big empty castle, but there will be times where you want to emphasise scale, such as to make a boss room seem more cavernous, or a reward room seem as impressive as possible.

How the player approaches a location can have an impact on how they experience it. For example, the longer the lead up to a structure, the more the anticipation is built up. This can be used for approaching important locations in a game, such as the epically long spiral staircase leading to the boss fight with Ganandorf in ‘The Legend of Zelda: Ocarina of Time’. The more important the location, the longer the route is leading to it. If we look back at the ‘Ico’ example, the long lead up can be accompanied with open space and a huge landmark in the distance to create a sense of awe and scale.

A long bridge in ‘Ico’ builds anticipation for what’s to come

However, a long leading route can be used not only for emphasising scale, but also for signifying a resulting room’s importance. In ‘The Last of Us’ the last corridor that leads to the ending surgery scene is lengthy to represent its significance, but doesn’t require a looming landmark which would have been out of place here.

The last corridor in ‘The Last of Us’ is long to denote the importance of the proceeding room

This spatial arrangement comes from real life psychology, and can be seen throughout history. The 12th century temple complex of Angkor Wat in Cambodia is an excellent example of this. The main temple has a route leading up to it stretching nearly half a kilometre across open space, giving the structure a grand appearance as you approach it, adding to the sense of awe and wonderment.

The path to Angkor Wat is nearly half a kilometre in length

It’s also possible to design an interior environment which feels huge and impressive even if it doesn’t have open spaces that allow for grand vistas. In order to make a room feel big and impressive the player should enter through a contrasting area which is a much smaller and tighter space. Forcing the player through an enclosed space means the resulting entrance to the larger space feels much bigger by contrast. Examples of this which you can see in the real world are the entrances of religious structures. Churches, for example, make visitors enter through small entrances before the grand cavernous interior is seen. Leading vertical lines are also used inside to encourage people to look up towards the ceiling; they are encouraged to look up ‘towards the heavens’. These are all spatial theories we can utilise in our level designs to elicit the types of emotion and promote the actions we would like the players to have and do.

Forcing a player through a small space makes a proceeding space seem larger by contrast

Pipeline

I get an initial brief from the game director about what level/s I’m working on and what gameplay and story beats need to be included. I will also get a brief from the art director about the location they want the levels to be in, as well as any accompanying reference material or concept sketches.

When I start designing the level I start off making a good plan – planning is super important but can be the most boring part for most people when they start out as they just want to jump straight into creating something 3D! I’ve learnt to love this part as I can set out the structure of what I need to include in my design so the level isn’t just an incoherent jumble of independent ideas forced together. For example, in the opening level of ‘Killzone Mercenary’ we had a boss fight against a tank which you destroyed using D-charges. I knew I needed the player to know how to use it so I planned in an isolated period of downtime that used the D-charge as a progression blocker to ensure they learnt this mechanic before being challenged with it.

Once I’ve got a good idea of my gameplay beats I start creating a greybox in Maya and get it into the engine as early as possible. This way we can assess how big the level is and get a general idea of the flow and pacing. If it’s too big or my initial blockout is going to be too empty then it’s really easy to adjust at the start before it’s had any investment from the art team. During the greybox phase I also get the combat encounters included as well with all the necessary scripting for this to work. The combat shouldn’t be forced in afterwards when a level is locked down as the environment should be sculpted at the start to support the desired gameplay.

After the greybox has been signed off, the art team start making it look great. During these phases there are always changes and iterations that are needed, so the disciplines involved collaborate together to start improving the level to make it the best we can; playing great and looking great.

Eventually as the level progresses and changes with the implemented improvements we then finally get a level signed off. We won’t change anything from then on unless it’s absolutely necessary (and this always happens even in the very last stages of the project!). Then we are on to the final stage of bug fixing and polish before release!

Tips and Tricks

If you’ve just started as a designer in the games industry, or you’re beginning to make your own indie games, some advice I’d give is don’t be too precious over your work. In the interest of improvement your levels will always go through iterations, things will be chopped and changed and at times whole sections can be completely scrapped. If you’re too precious over your work you’ll never want to change anything and ultimately the level will suffer. Changing sections doesn’t mean that you’ve done a bad job; in fact the team will have learnt a lot from the first design and although it’s not the angle they want to go with, you’re getting ever closer to what the game should be. Work is never wasted as it provides much needed lessons for finding the tone and experience of the game. Additionally, if you’ve got a great idea that’s a lot of fun but unfortunately doesn’t fit for the game or area you’re currently working on, you can always come back to it in the future.

If you’re trying to get into the industry and you’re focusing on building a portfolio I’d urge you to make something achievable – you don’t have to create a whole level that has an art pass, you just need to create something that demonstrates your design knowledge. This can be done in one simple area of a level, such as including a landmark for your goal, a combat encounter with enemy waves and perhaps a simple puzzle for progression. This is all you’d need to demonstrate your ability and knowledge and it’s something achievable; if you bite off too much not only will it be daunting but you’ll probably never finish it either.

Also, don’t bother writing design documents, or your ideas for games as unfortunately no one will ever read them. People don’t want to read (which is a good design mantra to know as well!) and they don’t have time allocated to read your document. You also won’t be doing that much document writing in your job, yeah you’ll do the odd specification but you’d mainly be creating and implementing content into the game, so make sure you demonstrate you can do that!

Pete Ellis, Level Designer at Guerrilla Cambridge

INTERESTING LINKS

SUBVERTING PLAYER EXPECTATION BY USING LEVEL DESIGN

STEFAN DOETSCHEL: THE PRINCIPLES OF LEVEL DESIGN

A GREAT SCENE TO LEARN SURVIVAL HORROR LEVEL DESIGN