However Damnation are apparently going a different way and basing their Level Designs on Concept art. 

I’m interested in seeing how this process works out in terms of actual gameplay as a good concept artist can throw out sketches faster than you can block out a level in sketchup and tend to get hired on the strength of their imagination. In contrast Level Designers are often hired for their technical skill more than their artistic ability.

Being able to use a game Editor and script in LUA will push your Resume closer to the top of the pile compared to somebody who can make their own textures (ProTip: If you spend longer making your level look awesome than play awesome apply for Environment art jobs).

So with that in mind inspiration can be tricky when coming up with your designs. Having a Concept artist onboard early in the process will at least introduce a sounding board for ideas and a second perspective from somebody who can visualise a 3D scene and then communicate it quickly. So long as it’s a two-way process with the artist and designers inspiring each other I can’t see how it could fail.

So here’s hoping Damnation is actually fun.

So in the interests of trying to aggregate the genuinely useful links I find I shall be trying to keep this post up to date with what I find.

Single-Player

• Level Design Pre-Production - Dating back to 2003 this article goes over the practical aspects of actually designing a level layout before you start building it.
• Using Loop Layouts - From the Valve Community Wiki. This is a concise and to the point explanation of using looping layouts in single Player Levels with several case-studies from Valve Games.
• Bouncing or Back-Tracking – Also from the Valve Wiki another concise page detailing effective use of Back-Tracking or “Bounces” in Level Layouts. Again lots of excellent case-studies.
• Color Theory – Another page from the Valve Wiki! This page ties-in with what I describe in Continuity Level Design about contrast and composition but also covers more general color theory.

Multiplayer

• Competitive Level Design – Dating all the way back to Quake 3 this page is an excellent grounding in the tactical and strategic considerations required by designing competitive maps.
• Level Design Patterns – (Warning PDF Alert!) Skip to Page 10 for the useful stuff. A Good grounding with Case Studies on common Patterns in Level Design. The Principles can be applied to other genres than FPS games quite easily.
• Making de_dust – Proof that it’s better to be lucky than smart the creator of the worlds most played Level takes you through the creation of de_dust.

Prompted by this article over at Bit-Tech on How AI in Games works it made me want to blog about some of the ways you can make your enemy AI appear smarter without having to re-write your AI back end or invest in expensive middleware.

SCRIPTING

A huge amount of what convinces us a game has good AI is often completely scripted. Nothing ruins the illusion of AI like the Level Designers treating them purely like cardboard cutouts or filler. What the AI is doing before they start fighting is just as important as what happens during the fight. 

My favorite reference for believable AI scripting are the NOLF games. Yes you couldn’t go more than ten yards without stumbling on a pair of AI guards having a conversation but it adds so much to their characters that you become far more forgiving towards them.

The arrival of an enemy is a great opportunity to make them seem smarter than they really are. AI that Rapel down a wall before fighting you or who shout for backup when they see you causing two more enemies to sprint around the corner. These are almost always scripted events but most Players will attribute it all to the AI.

Even just adding Patrol routes or some good ambient animations can make a very simple AI character appear more inteligent.

FEEDBACK

AI that don’t tell the Player what they are going to do before they do it look stupid.

In Half Life the Soldiers would always shout “GRENADE!” before they threw a Grenade at you. When they shouted “Flank Left!” an AI would always run out of cover moments later.

You are making your enemies appear as if they have goals and that they are trying to fulfill them. If the player doesn’t know what the AI was trying to do then alot of their actions won’t make any sense. If a Soldier just ran out of cover then the Player would assume the AI is stupid for leaving cover. By preceeding it with the command “Flank Left!” the enemies behavior is given a context that makes a stupid suicidal manouvre appear inteligent.

Your enemies don’t have to actually Flank anybody. They just need to say that they are before they move somewhere else. The Player will assume the AI is trying to do something very clever when they are really doing something immensely stupid. It’s all about context.

Above all else AI that are talking to each other and the player appear more Human. The Grunts in Halo would have seemed very stupid if they had never said anything. Instead you always knew what was going through their tiny minds as a result of your actions.

ACCURACY

If your enemies carry guns then don’t turn them into Storm Troopers (unless they ARE Storm Troopers in which case it’s OK). While the Player should have an advantage ov er the AI in terms of accuracy making the AI too inaccurate will just result in them looking stupid. If the player realises that they can just stand still in the open while they pick off the bad guys with a single headshot each then their sense of immersion will be completely broken.

Below is a video I made when we were making Wheelman to indicate just how “Stupid” AI can appear when they are inaccurate. The good stuff is at about 1:50.

If you want your AI to miss then put some effort into making them miss believably. If they are using automatic weapons then have them group their shots in a burst and have the entire burst miss. Don’t just spread the shots out in a cone around the player because it’s much more noticable.

If the Player is standing still in the open looking at the enemy then the AI should tear the Player a new one.

LIFETIME

You can have the most inteligent goal-oriented perfect-pathfinding AI enemies in the world but it doesn’t count for anything if the Player is killing them before they get to do any of that clever stuff.

The only difference in Halo between the AI in Easy mode and Legendary mode is that in legendary the AI has more health. This meant they lived longer and as a result did more cool clever things making them appear more inteligent. They were also more accurate and did more damage which supports my previous point.

Adding health isn’t the only way to increase your AI’s Lifetime (and you may not be able to get away with it contextually) but there are other ways to make your enemies live longer.

• Enemies that can Blind-Fire are harder to kill while still presenting a threat to the Player.
• Make the Players weapons less accurate meaning they have to work harder. Too much aim-assist is a bad thing.
• Allow enemies to break-out of Hit-Reactions quickly. If your Hit-Reactions are too long then you often only need to land one hit to neutralise the enemy. The remaining hits are a formality.
• More Enemies at once mean that even though Individuals die quickly the others are showing off in the meantime.

ROBUST

If I had to choose between Very Clever buggy AI and Very Simple Bug-Free AI I will always pick the Simpler choice.

Every time an AI gets a Path wrong or does something slightly unpredictable (in a bad way) is another notch against your AI appearing smart.

Half Life 2′s Combine AI is really simple but also very effective. Their basic behavior is to move towards their target shooting at them if they have line of sight. When they need to Reload they will move to a position where they can break line of sight.

That’s roughly 80% of the Combine AI right there. It’s incredibly simple and copes with just about anything you can throw at it. Because it’s so robust it rarely goes horribly wrong and the Player can build a mental model of that enemies behavior very quickly that is always right.

When the AI behaves how the Player predicts then the AI is perceived as being smart because they have fulfilled the Players expectations of them. When the AI exceeds these expectations it’s even better. However the most important part to nail first is that the AI never falls short of them.

MANAGE EXPECTATIONS

Human shaped AI are the hardest to get right because as Humans we have very high expectations of their behaviour. We interact with other Humans everyday so Human AI in games needs to fit into an existing Mental Model informed by our life experiences, TV, Films and all sorts of other crap.

Aliens, Animals, Robots. All of these are far easier to make appear inteligent because our Expectations of their behaviour aren’t preconceived or at worst much lower. Nobody knows how a Strider is supposed to act before they start fighting one. You can infer a bit from the way they look and move but they are essentially a blank slate. A Strider can behave however Valve wants them to and so long as they are Robust they will appear inteligent. 

If Halo’s Grunts had been Human shaped then their AI wouldn’t have seemed as clever as they wouldn’t have matched our preconceived expectations of how Humans behave. Don’t underestimate just how much leeway using non-Humans will buy you.

LEVEL DESIGN

Bad Level Design can make good AI look terrible. It is the job of the Level Designers and AI Programmers to ensure that everything in the game is working together to create the best experience possible. If you spend six months crafting cautious cover using AI and all the Levels in your game are empty warehouse deathmatch levels then your AI isn’t going to look good.

Make sure that Levels are playing to the AI’s strengths and vice-versa.

Continuity Level Design is a school of Level Design predominantly used by Single Player games. 

Recent examples include Portal, Mirrors Edge and Call of Duty 4. 

Now it’s worth pointing out that the phrase “Continuity Level Design” is something I’ve just made up (well at least as far as I’m aware).

I’ve chosen the name based on the similarity in concept to Continuity Editing in Cinema.

There are a few rules that characterize Continuity Level Design:

1. The Player Always Knows What They Are Doing
2. The Player Always Knows Where They Are Going
3. Everything Happens In-Front of the Player

If the above list looks like it’s making the Player the centre of the universe then that’s because it is. Everything about this school of Level Design is based around their being a single perspective on events.

What Am I Doing?

The most important part of Continuity Level Design is to ensure the Player always knows what it is they are doing in your level. 

This isn’t about long-term objectives like “Rescue the Princess” or “Defeat the Terrorists”. 

It’s about the question “How do I get from Where I am to Where I want to be?”. 

Any time a Player stops and thinks “What am I supposed to be doing?” is a failure on the part of the Level Designer. 

An Example:

The Player is stuck in a building with only one way out. This exit is guarded by some enemies who cannot be killed by the players weapons.

To escape the building the player must shoot a specific spot on a wall which will kill the two guards allowing the Player to continue.

The spot on the wall is completely indistinguishable from the rest of the wall. 

The Enemies guarding the way out appear to be regular enemies that the Player has bested in combat a hundred times before. 

Thinking they can just kill the enemies they engage them in combat and are swiftly killed.

This violates a key principle:

• Never contradict the Players understanding of the games mechanics!

Every time the player interacts with something they are building an internal model of how the game works. When that model is contradicted so blatantly it causes frustration and confusion. 

Unkillable enemies are viable mechanics in a video game. However they should never appear the same as an enemy that can be killed.

At this point the player may begin to backtrack around the building thinking they’ve taken a wrong turn or missed something elsewhere.

They’re now having to consciously think about what they are doing and as they start to discover that there is no other way out and they haven’t missed something they will become more and more frustrated and angry. 

The Player is stuck in a building with only one way out. This exit is guarded by some enemies who cannot be killed by the players weapons. 

Above the guards is a wooden platform supporting some concrete blocks. In the middle of the platform is a Gas Canister. 

The Player knows that Shooting a Gas Canister will cause an explosion. 

While the Gas Canister is too far from the Guards to kill them the player also knows that Wooden Objects can be broken. 

Shooting the Gas Canister the Platform breaks apart and the Concrete blocks fall on the Guards and kill them allowing the Player to escape.

As cliche as the example may seem this makes up the majority of Half Life 2′s game play. 

However what is important here is that the player has used their internal model of the games mechanics to solve a problem. The Solution may have been in plain sight but it is up to the player to establish the relationship between the elements and solve the problem. They will adjust the model to include this new information on how to kill this kind of enemy and similar encounters will appear much easier in future. 

Establish a Visual Language

A Visual Language is what Level Designers use to communicate with the Player. The Level Designer should really be telling the Player how to complete their Level.

It is not a competition between the Player and the Level Designer to see who wins.

My favorite example of a Visual Language is Portal.

Every element of a level is introduced one at a time in Portal and how they work doesn’t change after their introduction.

When the Player arrives in a new level they understand the relationships between all of the objects with the exception of whichever new concept is being introduced.

Additionally Portal makes a point of explicitly declaring the relationship between elements within the level itself. In the screenshot above the relationship between the buttons and the door is declared by the trails leading from the buttons to the exit.

The question of “What am I doing?” does not generate game play. What generates the game play is the question “How do I do it?

Where am I Going? 

The question “Where am I going?” seems to go hand-in-hand with “What am I doing?”. 

However this question is much more related to a levels “Flow” and exists mostly at a sub-concious level for the Player. 

Portal makes a point to always tell you as soon as they can where you are going.

In the majority of the challenge rooms you are shown the exit right from the elevator.

As soon as you have arrived you know where you are going to. The trick is figuring out how to get there. 

This keeps the player oriented in the world. Ensuring they have a point of reference straight away.

The Zelda series is very good at making sure you know where to go next while keeping the “How” a mystery to be solved.

If the exit of a dungeon room isn’t obvious from the entrance then you are often treated to a camera cut that shows you where it is and enough of the level to establish your bearings relative to it. 

However not all games want to use a camera cut to show you the world preferring to feed you directions like bread-crumbs.

A Lot of the techniques used to show the player where to go can be borrowed from Photography or Cinema and have been studied in far greater depth than I can hope to go into here.

Composition

While a modern video game is a 3D world the rules of Composition as established by Photography are still very helpful to a Level Designer.

It is important to determine where the Player will look while playing your Level. 

At a high level this is broadly defined by geometry and the actual layout of the level itself.

Everytime the player turns a corner the Level Designer knows where they will be facing and can use this to ensure that important elements fall into their field of view.

A good test is to take a screenshot as soon as you’ve rounded a corner or opened a door and then examine the screenshot out of context. Ask somebody else if they can tell from the screenshot alone where they should be going.

At these points the Player is making a decision about where to go next and the composition of the scene at that point has a huge affect on where the player decides to go.

Contrast

Photographers have been using contrast to lead and focus the viewers attention in pictures for decades. 

Mirrors Edge used this for it’s “Runner Vision” and arguably used it as something of a crutch at times.

Without Runner Vision; and especially at speed, this appears to be a dead end.

However if we turn Runner Vision on…

Lighting

Lighting has additional benefits over tonal or colour contrast.

Well-lit areas are of tactical advantage to a species that has typically poor nightvision and under-developed senses of smell and hearing.

Because of these primal instincts we naturally favour areas of light over the dark and associate them with safety.

The exception to this is the Thief series which turns the concept on it’s head and establishes the light as dangerous and to be avoided.

Left 4 Dead uses Light very well as an aid to navigation:

The bright light in the kitchen draws our eye and focuses our attention towards it.

Even subtle changes in lighting will draw our attention.

When lighting a level it’s important to ensure that the lighting is reinforcing the gameplay as well as making it look good.

If you don’t want the player to go somewhere then shroud it in darkness.

Leading Lines

Lines are used in photography to direct attention around the picture.

Straight diaganol lines tend to have the most power in this respect as they don’t occur often in nature.

The power of lines is increased when there are multiple parallel or converging lines as these will reinforce each other.

The most dominant lines in a level are ussually edges of geometry and the points where the wall meets the floor or ceiling.

Granted in this shot there is only one direction to go: Up. However everything in the shot is reinforcing the up-ness.

The horizontal red block on the left wall is pointing us towards the staircase.

The stairs and the bannister are then pulling our attention up and towards the vertical pillars.

The Two vertical pillars here are drawing attention from their base (the floor) upwards. If the lines are “Rooted” like this then your attention will travel along them in only one direction.

Ambiguous Lines

An example from Mirrors Edge where Lines are actually causing confusion for the player is below:

Immedietly after opening a door we are presented with a choice: Left or Right?

Unfortunately here we are given no clues at all where to go. The dominant lines aren’t grounded and are telling us that both choices are as valid as each other.

I’ve played this level alot (It’s the first one after training) and I still turn left before I turn right.

Now this could be intentional. One of the things to remember about the rules of Continuity Level Design is that they are made to be broken.

Knowing how to lead the player through a level effectively also means you also know how to confuse them on purpose.

Movement

Evolution has granted us an excellent ability to spot the tiniest of movements with almost no effort at all. 

Spiders enjoy using this to scare the crap out of us.

To demonstrate just how good your eyes are at spotting movement have a look at this screenshot of the game “Lost In The Static” by Silver Spaceship. 

Just static right?

Wrong. Download and run the game and you will be able to see the difference a little movement makes.

Here is a YouTube video from GDC of the game in action.

Movement screams out to us even in darkness. Again this is a survival trait we’ve evolved over millions of years and has served us pretty well. It’s also why camouflage works best when neither the viewer or the target are moving.

Using Movement

It’s quite hard to identify where to go in this shot. There are a few tell-tale signs but only one of them really stands out when the game is running.

The spinning C sign not only moves but is shiny so it’s surface is changing as it spins in the sun.

The Spinning object in Mirrors Edge can only ever draw attention to the area it occupies.

However should a Jet fly over the buildings our attention would be immediately drawn along it’s path.

So there are two very simple rules here:

1. Localised movement draws attention to it’s origin.
2. Movement along a path draws attention along that path.

Scripting Movement

This shot from Half Life 2 – Episode 2 demonstrates how a series of moving elements timed well can draw the players attention to an important event in the level.

As you approach this point the path narrows which keeps you facing the desired direction.

To draw your attention to the roof of the building a flock of birds take off and fly up the screen taking your attention with them. Then the Hunter comes into shot on the roof of the building foreshadowing the next section of game play.

All of the elements in this part of the level are leading up to the reveal of the Hunter for the first time.

If something is moving that isn’t aiding the game play then it probably shouldn’t be moving at all.

The Third Rule

The third rule of Continuity Level Design is one that is broken quite frequently.

“Everything Happens In-Front of the Player”

Continuity Editing has a rule that if a character or element is off-camera then it doesn’t move.

This is to ensure that the viewers internal model of the scene stays up to date and in-sync with the action on the screen. 

One exception though is that objects can move around so long as they maintain their momentum.

It is for exactly the same reason that Rule Three exists in Continuity Level Design.

The Players Situational Awareness is based on the current momentum or behaviour of all the objects they know about.

They are assuming that the immediate future is going to be very similar to the present.

In the first level of Mirrors Edge you crawl through an air-vent and drop into a corridor.

The shape of the Air Vent and the direction you face when you drop out of it were specifically designed to ensure that the Player is aware of the enemies waiting for them.

Now the player is aware of these enemies they can add them to their internal model of the world and make predictions about their behaviour.

In the same level after a cut-scene the player is positioned so the very first thing they see are the four new enemies.

The enemies will wait for a couple of seconds before engaging allowing the player time to react and evaluate the situation.

 In the case of Mirrors Edge the Police will always chase you so you can safely categorize them as being “behind you” even though you can’t see them.

Inattentive Blindness

Humans only have a limited amount of attention they can use at once (sometimes referred to as Psychic Energy) and this limits how much data you can consciously or subconsciously process.

You’ve probably seen videos like the one below before but I like this version.

Now I don’t know about you but I’ve seen about a half dozen videos showing this principle at work and I still got caught out at the very end of this one.

Because Attention is a finite resource you need to make sure that when something important happens that the player can process it.

Even if something is happening in-front of the Player there is no guarantee that they have actually noticed it.

You can use all of the same techniques used to tell the Player where to go to ensure their attention is focused on important events.

Altogether Now

To demonstrate all of the above techniques I’ve chosen a short (probably 2 minutes) section of game play from Half Life 2: Episode 2.

Here we see the use of Light and Lines to draw our attention to the left of the scene.

The fences allow us to see part of the world where we will eventually go but indicate to us that we will have to go around them somehow.

Looking left towards the light we can see that we can’t go this way as another fence blocks our path.

However we can make out something blocked by wooden boards and further attention is drawn to it by the Green signs.

In order to get this far we had to break Wooden Boards just like those with the Gravity Gun so we already know that they can be broken.

This has established this part of the world as being important in our internal model.

Turning around we see more light and can clearly tell that this is the way out.

We are also shown a Black Headcrab which is far enough away that we have plenty of time to prepare for it’s attack.

Headcrab dispatched we reach the bottom of the ramp and discover the way out blocked by wooden boards.

These don’t look like the boards we were able to break earlier but if in doubt a few experimental zaps from the gravity gun establish these as the Unbreakable kind.

The Diagonal lines of the Mine Cart tracks draw attention back up the ramp.

Here the lighting and lines are all working to draw our attention to those breakable boards we spotted earlier.

By now we know the solution to the problem as whatever is being held back by those boards is no doubt going to smash it’s way through the Unbreakable Boards.

Breakable boards broken (and getting a tiny bit run-over) and the Mine Cart breaks open the exit and we can progress again.

Again you can see how the lighting and lines are all focusing attention to the point of impact. Add on the motion of the Mine Cart itself and it’s almost impossible for us to look anywhere else allowing us to admire the results of our work.

Parting Words

I started writing this blog post a week ago with the intention of getting some of the things I’ve learnt over the years out into the world for others to learn from.

As such this essentially forms my disclaimer: I don’t pretend to be an expert on everything in here and to cover all the topics in-depth would require a huge amount of time and space. There are good odds I’ve gotten something wrong somewhere or misunderstood or misapplied a concept. If that’s true then I would hope somebody is kind enough to help and explain it better. 

Continuity Level Design is just one part of Level Design as a whole. There is alot more to Level Design than what is in here and I hope to cover more topics at some point in the future.

In the meantime if you want more Steve Gaynor has written a quality post on Combat Encounters in FPS games Here.

However research papers are often very dull and you ussually need to have a dictionary on hand to make sense of them. So here are ten books that aren’t about Games Design but will make you a better Games Designer…

Flow - Mihaly Csikszentmihalyi

This is something of a Bible among games designers and is always the first book I reccommend to people. While it’s marketted as a feel-good self help type book it really isn’t so don’t let the subtitle fool you.

The book is very easy to read and explains the concept of “Flow” – A state of conciousness where a person loses the concept of self and becomes entirely involved in what they are doing.

The Design of Everyday Things – Donald Norman

Ever wonder why you sometimes try to push a pull door? Or why some TV remotes are easier to use than others? Don is here to explain all of this and more.

All the lessons in this book are easily transferred to Videogames and may change the way you look at the things you use everyday.

How to Lead – Jo Owen

A common mistake new Designers make is thinking that the most important part of being a Designer is “Designing”. It isn’t.

The most improtant part of a being a Designer is being a good Leader. 

Designers are required to interact with everybody else on the team all the time. We are the people who have ideas and are then responsibile for convincing everybody else to follow us in developing them into a reality.

Every Designer has to be a Leader so better make sure you’re a good one.

Writers’s Journey – Christopher Vogler

I choose this version of the classic Heroes Journey because for me it was the easier one to read. It’s stuffed full of modern film references so you don’t need to know anything about mythology.

The Heroes Journey as a concept is important not just for story writing but story telling. Understanding what makes a good story helps you design a game that will give the player good stories of their own whether you’re making a narative heavy adventure game or a hardcore online shooter.

Universal Principles of Design

This is something of a guilty secret of mine. Every so often somebody would trot out a phrase like “Ockhams Razor” or “Hick’s Law” and I didn’t actually know what they meant.

This book includes 100 of the most basic Design Principles and concepts that every Designer should be aware of (but probably isn’t). 

If you want to find ways to improve your documentation then this book is for you.

The Laws of Simplicity – J Maeda

As a Designer you can quickly get caught up in details and end up in a spiral of constantly adding to something.

At only 127 pages the book itself follows it’s own rules and keeps the topic of Simplicity… simple.

If you ever felt overwhelmed by the complexity of your own creations then you can learn to deal with it from this book.

Freakonomics – Levitt & Dubner

Learn to look at the world differently and find the true cause of something rather than the obvious one.

How did Legalising Abortion reduce the Crime Rate? Why is your boss more likely to steal a Bagel than you?

In a time when focus tests and usability studies are more and more common in games it’s important that you learn to look at Statistics from all angles.

Emotional Design – Donald Norman

Back with another great book Emotional Design examines how we relate to things emotionally.

Being able to identify the Visceral, Behavioural and Reflective elements of a Design and understanding the relationship between the box something came in and how people treat that thing can change how you think about your game.

More up to date than Everyday Things this book actually has an entire chapter on Videogames.

Creativity – Mihaly Csikszentmihalyi

Creativity is a fairly nebulous concept. In the game industry you can be an artist without being remotely Creative.

Understanding what Creativity is and how to idenitify and nurture it is the difference between resigning yourself to working on boring licensed crud and celebrating the opportunities an established license gives you to revolutionise a genre.

In a Creative atmosphere with Creative people whatever you do will be Creative.

Hows and Whys of Level Design – Sjoerd De Jong

OK I lied. The tenth book is actually about Game Design.

However I have good reasons. There is very little printed on the art of Level Design and Hourence has managed to distill alot of concepts into a very easy to read PDF.

There is far more to Level Design than alot of people think and I feel that alot of the knowledge of Level Designers has been trivialised in the name of “making shit look cool”.

Level Design does not stop at Sketch-Up.  It is merely the start.