The Taxonomy of Left 4 Dead.

Nearly twenty years ago Richard A. Bartle, co-creator of the first Multi User Dungeon (MUD) the precursor to the Massively Multiplayer Online Game, wrote an article entitled “Hearts, Clubs, Diamonds, Spades: Player who suit MUDs”. Even today the classifications he set forth still hold true when examining the types of player who play World Of Warcraft and other MMORPGS.

This taxonomy can also be applied, with some modifications, to any multiplayer environment. There will always be those players who value winning over anything else, the Achievers, the Diamonds. Or those who simply enjoy the experience of playing with others, the Socialisers, the Hearts.

Playing Left 4 Dead recently it struck me that the characterisation of each of the four survivors seemed to conform to one of these four player types.

Bill, is the grizzled veteran, the professional. He knows what he’s doing and will get the job done with little fuss. The goal is to win (to survive) and anything that doesn’t increase his chances of winning is not worth bothering with. He’ll crack the occasional joke to break the tension but ultimately he’s the responsible adult of the group. He’s the Achiever.

Zoey, is the horror film fanatic. Like Bill she knows the rules, she’s seen all the films, she knows what happens when the dead start to walk the Earth. At the same time she’s always ready to lend a hand, and support the group. On her own she might survive but what’s the point is she’s the only one left? She’s the Socialiser.

Francis, is the misanthrope, the cynic. He might know the rules, but you can never be sure, as he’s perfectly willing to break them for his own benefit. He’s good to have around when the horde descend on you but if he has to he’ll leave everybody else to die, after all what does he care, he hates everything, so don’t expect him to care about you. He’s the Killer.

Louis, is a little out of his depth. Everything is an experience, everything is fascinating and new and exciting. He’s not sure of the rules, but fascinated in finding out what they are, and have some fun along the way. He’s just as liable to accidentally shoot you in his excitement as he is to shoot the shambling masses of undead. He’s an Explorer.

The nature of Left 4 Dead, is such that Bartle’s original classifications don’t fit exactly, with no explicit reward beyond survival Achievers cannot really be singled out by their desire to accumulate wealth or experience and the cooperative nature of Left 4 Dead means that Killers cannot be solely identified by their desire to impose themselves on others.

A more appropriate way to define the different play styles of Left 4 Dead, and a method that works for other multiplayer titles, is to examine where players fall on two axises defined by their tendency towards a focus on themselves (Cooperation) over others (Competition) and their attitude towards the rules of the game.

Returned to out Left 4 Dead example, Francis and Bill style players have a focus on winning, on competition and success. Whereas Louis and Zoey style players have a focus on the act of play itself and their experience with others.

A team of prodominantly Bills will almost always get to the end of a campaign, they are effecient, reliable and proactive. They have a plan and they’ll execute it, even if the whole experience might seem a little joyless to the outside observer.

A team of Zoeys, might get to the end of a campaign but if they do everybody will get there together, as a team. Like Bill they are proactive and reliable, but they are also willing to risk themselves for the good of the team.

Much like a team of Bills, a team composed of a majority of Francis’ will get to the end, but it might be messy. They’ll keep each other alive as long as they need to because an extra weapon and extra pair of eyes is always useful, but come the climax it’s everybody for themselves.

As with a team of Zoeys a team with a high percentage of Louis’ might get to the end or they might all die horribly, either way it’ll be an interesting experience and something will be learnt, even if it’s just when not to throw a Molotov Cocktail.

With a little modification this taxonomy can be used to differentiate the types of players in other multiplayer games, or potentially even predict the outcome of team games, by examining the make up of players within each team.

Oh and by the way, I’m a Zoey. Who are you?

The Play Experience.

In the past I’ve taken to examining games and game mechanics in a manner that is open to criticism for being piecemeal. I’ve examined specific elements of a game in isolation, ignoring important mechanics and interactions in order to aid clarity. The obvious problem with this approach is that a game is a complex system of rules interacting over time, and it’s impossibly to present an entirely accurate analysis of any individual mechanic without looking at all the ways that mechanic affects and is affected by all other mechanics. That is a valid criticism, and I think it’s important to understand and appreciate that no mechanic exists in isolation and everything in a game is in some way, either directly or indirectly, connected to every other.

However there is a reason why I have often chosen to approach specific mechanics in isolation, or to otherwise present a limited view of the dynamics of a game. As I have previously discussed games do not exist in any measurable way, the rules defining the constraints of the game are not the game itself. It is only during the act of play that those rules take on a contextual significance, that they are able to enthrall, thrill and otherwise emotionally affect us. Games exist to be played. The experience of play is different for each game and each player.

Though it is important to consider games in their entirety as a system of interconnected rules that is often not how they are perceived when played. Consider any game you have just started playing. After ten minutes how many of the mechanics are you aware of? How many of the myriad relationships between different mechanics do you have an understanding of? How about after two hours, or five hours, or fifty hours? Depending on the game, and how much attention you have been playing, the answers to these questions can differ significantly. Often it is not until the very end of the game that a complete conceptual model of the mechanics and their relationships can be formed.

Play, by its very nature is a process of exploration and discovery.

Consider BioShock, I feel confident that by this point most people who have played the game are aware that the mechanical rewards, in terms of Adam, are essentially the same regardless of whether you choose to Harvest or Rescue the Little Sisters. When analysing the game itself it is important to consider this fact, however when examining the play experience it might be useful to ignore it entirely. During play, assuming you don’t run two parallel games in order to examine the consequences of each choice, there is no information provided on the reward you would have received if you had chosen the opposite option. If you decide to Rescue the Little Sisters you will be rewarded with one quantity of Adam, and if you had Harvested you’ll be rewarded with another. It is implied that there may be an additional reward for Rescuing a certain number of Little Sisters, but unless and until you do so there’s no way to know whether that additional reward exists and what it might be. Even if you choose to alternate between Harvesting and Rescuing in order to gauge the difference in reward there is no way to know that the reward for consistently Harvesting doesn’t increase over time, or if the additional delayed rewards for Rescuing balance out the immediate rewards for Harvesting.

When playing a game what you do not know is just as important as what you do. If a particular mechanic or relationship is not known it’s impossible for that mechanic to influence your actions. When examining games and the manner in which they attempt to convey meaning it is important to understand the manner in which this subjectively understanding of the game mechanics influences player actions and reactions.

Show me how to play.

Getting stuck in a game is far from uncommon. The precise reasons for it are as varied as the games themselves, however all such situations can be grouped into one of two categories; skill based obstacles and logic based obstacles. 

Common to action games, the former situation is where an inability to progress comes from a lack of skill. The goal is obvious and the actions required equally so, however achieving the goal requires actions to be performed with a degree of skill not yet attained. Consider God Of War, standard combat encounters are skilled based, the aim is clearly to defeat any and all enemies and the means of doing so are the standard attacks and special abilities at your disposal; some attack are more successful against particular enemies however they are rarely the only options available. Achieveing the goal, defeating the enemies, is a matter of skillfully wielding the tools at your disposal.

Even the environmental puzzle section at which I became stuck, was a skill based obstacle. That I needed to walk along the beams and avoid the blades was obvious, however I lacked the skill to do so. My frustration was caused by my own inability to perform what I knew to be the correct actions required for progression.

Achieving progression is bound to the acomplishment of certain goals; the completion of any goal is a two stage process. It requires an intent and an action. In order to proceed players need to know what to do, and how to do it. Skill based obstacles are concerned with situations where the goals and the actions required are clear. The challenge exists in the act of performing those actions.

The second type of obstacles are those based on game logic. These occur when either the goal itself, or the means of attaining it are unclear. In either case, without adequate signposting, players are left performing seemingly arbitrary actions in the hope that they might progress. The player’s conceptual model of the game has broken down, they can no longer made valid judgements about which actions will lead to which outcomes. Their ability to communicate their intent to the game has been removed, or at least severed hampered.

Unless and until the correct course of action is identified progression is halted. At this point external information is required in order to keep playing.  The most common form of external information available is the walkthrough. They provide a solution enabling progression without the need to necessarily understand the reasons why you were stuck, or the logic underlying the obstacle. I’ve used walkthroughs myself and understand their appeal, however if we are to create games with meaningful mechanics walkthroughs as they are often written, can be potentially damaging to the play experience. If the meaning of a mechanic can only be appreciated through the act of play, using a standard walkthrough could be damaging. They provide the solution but rarely highlight the logical deductions and assumptions that led to it.

In order to experience the meaning inherent, or imbued, in any mechanic it seems vital for players to learn how that mechanic functions and understand their application. If such a mechanic is difficult to understand that could be intentional and the act of understanding itself is important to an appreciation of the mechanic.

Using a walkthrough would enable players to use a specific mechanic without any requirement to understand it.

For people with a history of playing games, the often obtuse logic behind some game mechanics is understood, often expected. We might laugh as a boss character changes its attack pattern upon entering the “third stage”, willfully ignoring the logic that if such advanced attacks were available the boss should have been using them earlier. As seasoned gamers we have an understanding of the conventions of games and this can often enable us to deduce the solutions to logic based obstacles that really aren’t that logical at all. For gamers without this learned understanding of the illogical logic of games, without this game literacy, even the most straightforward of obstacles can seem insurmountable.

There are various methods used to try to mitigate this problem. The most obvious of which is to attempt to make any obstacles as straightforward as possible; to, in the words of some, “dumb down” by limiting available options and possible actions until only those vital to progression remain. Though I personally don’t share same concerns of some about this trend, I am worried that such attempts to ease progression can be damaging to the ability of game mechanics to convey meaning. If games reach a stage where players are effectively going through the motions, there will be little need to pay attention to the mechanics themselves and so their meaning could be lost.

One interesting possible method to avoid this was presented by Nintendo, and is due to make it’s first appearance in New Super Mario Bros. Wii that of the “digest moving image”or what has been referred to as the Nintendo “Kind Code”. The core of this system seems to be the ability to allow players to, at any point, view a playthrough of a particular section with which they might be struggling. Though they will not be playing the section themselves, by watching it being played it seems likely that they will pay more attention to the mechanics themselves, than they would if they had simply been told the solution.

Learning how to do something is potentially more important that actually doing it, the ability is what is retained not the act itself. Understanding the logic behind a particular problem makes it more likely that players will be able to solve future similar problems.

It’s important to help players create a conceptual model of the logic behind particular mechanics because often that conceptual underpinning is consistent throughout the entire game.

From out of the shadows.

It’s not uncommon for games with a well defined core mechanic, specifically action games, to include sections with different mechanics to break up the pacing and provide variety. When it comes to these “palette cleanser” sections it seems the stealth section is on a par with the turret section as the favourite choice of developers. If a new action game doesn’t feature one it will almost certainly feature the other, if not both. Given my feelings towards Thief II: The Metal Age and similar titles it should come as little surprise that I am fascinated and frustrated in equal measure by stealth sections in games. With very few exceptions stealth sections in games that haven’t been designed specifically around stealth mechanics are poorly executed. Think of any recent game with a stealth section; it was likely passable at best, if not outright unpleasant.

The obvious argument is that simple resource management means any mechanic used only for a single section of a game is going to receive less attention than a mechanic around which the game is focused.  I’m sure this is true and has an affect on the implementation of stealth sections, however I believe there is a specific problem with the mechanics of such sections; they are based not simply on poorly implemented stealth mechanics, but on bad stealth mechanics.

Stealth games are about power and the relationship of power to physical location. Good stealth games make the player a powerful agent in a world designed for them to exercise that power, bad stealth games make the player a weak agent in a world designed to reinforce that weakness. Bad stealth games, and by extension bad stealth sections, confuse being stealthy with hiding. It’s a fine distinction but an important one.

Consider Garrett, protagonist of the Thief series. Outside his cynicism his defining attribute is that he becomes invisible when in a dark area. It might never be explicitly stated but when the Light Gem is completely black Garrett is, for all intents and purposes, invisible. Darkness is a safe zone for Garrett and he has a variety of tools at his disposal with which he can alter the environment to increase the size of that safe zone.

This basic concept, the manipulation of the environment in your favour, is also present in the Splinter Cell series. Sam Fisher shares Garrett’s curious ability to become invisible in the dark, but the nature of the darkness as safe zone is taken further by the inclusion of alternative vision modes that allow him to see as well in darkness as light.

I think it's clear who's the powerful one in this relationship.

Both Garrett and Sam Fisher operate in environments which are, the majority of the time, in darkness. Environments where they are the ones in positions of power. The various guards and other non-player characters in the world might be better armed and more numerous than either protagonist but they lack a lot of their abilities. To them the darkness is a hindrance, to Garrett and Sam Fisher it is home. With access to an enhanced move set and the ability to modify the world around you, playing as either protagonist you have the upper hand. You have the ability to plan your approach and the moment at which you act. Things don’t always go as planned and you often have to improvise to survive but the choice of where and when to initiate action is yours.

Compare this mentality to that manner in which stealth sections in action games are presented. The core mechanics of such games provide the player with the most power when they are heavily armed and operating in open well lit environments. The available verbs are those that make the most of that environment.  When such games enter a stealth section the rules are changed, the previously available verbs, are either entirely removed or drastically curtailed. Non-player characters are now the ones operating from the position of power. Darkness for Garrett provides the ability for concealed movement and safety, for the standard action game protagonist it represents a diminished vocabulary and restricted move set.

This use of a diminished vocabulary in order to encourage stealth gameplay can be seen clearly in the two stealth sections of Fahrenheit (aka Indigo Prophecy). Both sections are a variation on the same theme, with the teenage Lucas trying to gain access to a restricted hanger on the airbase on which he lives. Stripped of the ability to interact with anything beyond that which is required for forward progression there is no choice of where and when to initiate action, the player must react to the game world and respond correctly or fail and be forced to restart the section.

Good stealth mechanics revolve around making the player powerful and giving them the means by which to exercise that power; bad stealth mechanics revolve around making the player weak and requiring them to work to mitigate that weakness.

It’s all in the wrists.

After this year’s E3 it’s clear that every platform developer is looking to add some form of motion control to their system, with complete 1:1 correlation between input and output being the apparent holy grail.

The potential various implementations are fascinating, however I have some usability concerns regarding the concept of motion control itself, specifically the desire to completely map input to output.

In it’s current incarnation the Wii is able to turn my uncoordinated and inelegant flailing in smoothly animated movements that make it look like I actually have a clue how to use a tennis racket or golf club. This ability to amplify and correct my input, making it representative of something much more elegant is vital to the appeal of games; nobody plays a game to look inept. Do we really want our every ill judged and inexperienced motion to be translated directly? Isn’t part of the fun of a game that it is a simulation of reality not a direct emulation of it?

I think this is potentially much greater issue than simply getting embarrassed when we fail to effectively swing a tennis racket. All the current implementations of motion control, including both those now available and those only recently announced, have one thing in common: they fail to take into account the affects of weight and resistance.

Consider how we interact with objects in the real world. If I pick up a golf club, the ways in which it can be used, the manner in which I can move it, is different to that of a tennis racket or a sword. Each of these objects has certain physical constraints limitting which actions are possible; how fast and accurately they can be wielded. I am able to swing both a golf club and a sword but the relative size and weight of each affects the control I have over the speed and power of that swing; the extent to which inertia and air resistance affect them is different.

If I am using some form of motion control device, or even my own hands, to represent a golf club or sword the degree of control I have over my swing is going to be very different. The controller will likely be both lighter and smaller than either the sword or the golf club and so my swings will be much faster. This lower mass will also mean less inertia so I will be able to perform more accurate and precise motions.

Having spent our entire lives interacting with the real world we have all developed an inate understanding of how objects of varying sizes and weights will react when we pick them up, swing them, or otherwise attempt to manipulate them.  This inherent understanding of the physical constraints an object posses is reinforced when we actually attempt to interact with it. It’s unlikely anybody, with experience of both objects, would attempt to swing a sword like a golf club. If they did somehow believe that this was a feasible and practical thing to do the moment they picked up the sword theywould understand why they would be unable to do so effectively, and why attempting to do so would be dangerous.

We understand the physical constraints of the world we live in and expect objects to react in certain ways.

When we are using a gamepad, or other indirect interface device, to wield a sword or a golf club there is a clear level of abstraction between input and output. This abstraction means our preconceptions do not exist, we don’t know the manner in which our inputs are mapped to our outputs, so we have no inherent expectations of how a movement of a thumbstick will relate to the swing of a sword or a golf club.

When we have a motion controller in our hands the level of abstraction between input and output is still present, because after all we are not actually holding a sword, or a golf club. However compared to using a gamepad there is a much greater correlation between input and output which leads us to logically expect any motions we make to be translated accurately. A direct 1:1 mapping of input to output leads to unexpected behaviour. The input we supply leads to the correct output for the motion controller itself as it’s movement is governed by the laws of physics, but the output of the object we are manipulating in the game world could be radically different from the output we are experience from out senses and that which we have been conditioned to expect.

The closer we get to a direct mapping of input to output the more noticable the dichotomy between physical movement and representation movement gets; it’s an uncanny valley for user interfaces.

If you use a motion controller to represent the movement of a sword then one of two things will happen.

1. The sword will swing exactly as fast as the motion controller, which is much faster that a sword possible could in the real world. This will seem strange as our inherent understanding of the physical constraints of a sword will have led us to expect it to move in a specific manner. At the same time we will see the sword moving at a speed and with a degree of accuracy that is utterly unnatural. There is a disconnect between what we instinctively expect the output to be and what we actually experience.
2. The sword will swing as fast as it should in the real world, which given its physical constraints will be much slower than that of the motion controller. This will seem strange as there is a distinct disconnect between our own movements and that of the object we are supposedly wielding.

Motion Controllers are seen as an effective means of lowering the barrier to entry of video games, they are less obviously intimidating than a standard gamepad, their use and means of operating often immediately obvious. The conceit that when I move this object I am holding, then an on screen representation of a sword or golf club moves exactly as the object in my hands does is a very appealing one. What is going to happen when full 1:1 mapping of input to output arrives and players begin to notice the dichotomy between what their senses are telling them is happening and what their conditioning is telling them should be happening?

When we use a gamepad we have to learn form a new conceptual model for how inputs map to outputs, this can be time consuming. When we are using a motion controller we might have the even more challenging task of unlearning a previously formed conceptual model.

I wonder if maybe I’m concerned about nothing, if the human brain is more than capable of assimilating this new mapping into it’s conceptual model of the world and if we will very rapidly lean to cope with the disparity between predicted and actual output?

On a curious side note there is one object for which motion controllers are absolute perfect, for which none of the stated concerns apply:

"An elegant weapon for a more civilized age..."

Since it is a fiction weapon (Look I won’t tell anybody if you don’t), there is no inherent understanding of how it is supposed to move. The blade itself is made of pure light so there is no air resistance, and the only weight that exists is in the hilt.

Playing the Sex card.

For a mature title The Witcher displays an all too believably juvenile attitude towards sexual relationships. At the mechanical level women are objects to be acquired the more the better. That on it’s own is offensive enough without even considering the soft porn cards you are “rewarded” with for having sex with particular women (Particular women in this instance being any with more than one line of dialogue it seems). If this wasn’t troubling enough there’s a situation in the early stages of the game where a woman “rewards” you for rescuing her from being raped by having sex with you…

The casually offensive attitude with which women are treated in The Witcher would be comical if it wasn’t so distasteful, still it does raise an interesting question: Is it ever justifiable for a game to feature an explicitly prejudiced and offensive character? Furthermore if that character is the protagonist is it justifiable for the game to present the world in a prejudicial manner, to include game mechanics that are clearly sexist or racist in nature?

At the risk of trying to justify the sexism on display I feel it needs pointing out that Geralt, the player character in The Witcher, is a predefined character based on a series of books by Polish author Andrzej Sapkowski. The usual room for player expression and interpretation of their character provided by most role playing games is limited. It doesn’t take long to realise that Geralt is a  misogynist, he sees women as disposable pleasure and doesn’t think twice about having sex with somebody and never even talking to them again. Given such a character the conceit of a collection mini-game for sexual relationships seems apt. It is an accurate representation of the way in which Geralt views the world, and women in particular; it’s not pleasant but then Geralt is not pleasant.

Sexism, racism and heteronormative behaviour exists in the world this is an unavoidable fact, to ignore than in our fiction and in our art would be extremely naive. Everybody has had experience with somebody who is sexist or racist and though those people might not be ones we’d choose to call friends, they are not evil people. Unpleasant to be around? Probably. Misguided? Almost certainly. But evil? No.

If we want mature games in the truest sense of the word then at some point they will need to engage with themes of prejudice and intolerance. It stands to reason that such games will need to feature characters who are sexist, racist or otherwise prejudiced and offensive. Demonising such individuals is the easy option and does little to tackle the issues themselves. If we are to include sexist and racist characters in our games, might we even consider making our protagonists sexist and racist? After all in a game we have the closest connection to the character we are playing. With such a protagonist we will need to make use of game mechanics that portray that character’s inherent intolerance and communicate it to the player.

Is it possible to implement such mechanics in a meaningful manner as a narrative tool without causing offense?

I believe, and hope, that it is. Of course it will not be easy. If we are to engage with the themes of sexism, racism and prejudice it will be necessary to ensure that games use such mechanics in an appropriate and intelligently designed manner. This will only be possible once we can ensure that games are not being casually sexist and racist through ignorance or apathy. It will be very difficult to make intentionally provocative and meaningful games dealing with prejudice and privilege when so many have shown they are capable of causing offense unintentionally.

Mechanical Definitions.

At a low enough level all game mechanics are the same, we press a button, move a stick, and something happens. Action and Outcome. Process and Result. Context is what allows us to determine if we are moving a ship through space or a counter across a board Context is provided by the narrative of the game.

If a plot is conceptually a “to do list” of events, then the rules of a game define the games systemic plot, what actions are possible and when. Using this metaphor game mechanics are the constituent, atomic, elements of a game plot, so what are the constituent elements of a narrative plot? Sentences? Dialogue? Games are often compared to film, and a line of dialogue, an individual frame, these are potentially the atomic elements of film?

As a method of conveying meaning, what is the role of any line of dialogue, any scene?

1. To move the story forward.
2. To provide information.
3. To characterise.

Nothing is wasted. Everything that is present should be important, and everything important should be present. If a character walks a certain way, it should provide information, characterise and move the story forward, or at least two of three. Shouldn’t the same hold for any game mechanic? How do we define a specific game mechanic precisely enough to determine if it meets any of these three criteria?

“Shoot this Grunt” is that the mechanic? Or is the mechanic: “Move yourself in the world so that you are Aiming at this character and press the Fire button”?

If an individual mechanic is to be an atomic element then surely the latter is too complex? It is several discrete actions: Move, Aim and Fire. In Halo: Combat Evolved it requires moving two sticks independently then pulling a trigger. It also has to occur at the correct, optimal, time. But then so does a line of dialogue, a particular scene. If they occur at the wrong time they make no sense.

Events in the wrong context hold incorrect meaning.

Is it actually sensible to try and examine game mechanics devoid of context? Should a game mechanic be considered an action within a context? Not: “Press this button to increase this number”, which increases some arbitrary number in the underlying simulation of the game but: “Increasing my Strength”? Should it be even more high level, an abstract: “Improve\Change my Character”?

If game mechanics should carry meaning at which level should that meaning exist? Or does it exist at each level? Is meaning implicit in action or is it, as I’ve discussed previously, tied to context? Context might inform whether we are moving a ship through space or a counter across a board, but does the action itself hold meaning free of this context?

Is it the responsible of the narrative context, to move the story forward, provide information and characterise or is it possible for mechanics to do that separate from their context?

I have a voice…

A few days ago I was asked to guest on a couple of podcasts, I’ll admit to a fair degree of excitement and trepidation. Unfortunately due to some unforeseen technical problems only one of those podcasts was recorded successfully.

On that I’d like to direct everybody to the latest podcast from the fine folks at Critical Distance, I’d also like to take this opportunity to apologise for basically coming in and all but taking over. I’m afraid that on listened to it now I seem to have talked an awful lot, though I’d like to think most of what I said was of interest. I also had quite a bit to say about a certain game.

I’d also like to publicly make a point of thanking Demonic Murray not just for inviting me in the first place, but also for producing the podcast and somehow managing to make us all sound less like rambling idiots. Trust me the amount of work that must have taken is such that I can totally understand if he occasionally forgets which episode it is…

So please everybody take the time to visit Critical Distance, if you haven’t already, and listen to all the podcast episodes there. You’ll be relieved to know the previous ones don’t feature my insane nonsense. Oh yeah, and listen right to the end…

DM-Aerie – Progress Log 2.

After overcoming some issues with UnrealED (Which is fortunately a less frequent occurrence than in previous versions) I’ve been able to spend a few further hours working on DM-Aerie.

As described in my previous log entry I had an overarching theme and had developed a number of individuals locations that would fit with that theme. With these in mind I began sketching out initial floorplans on graph paper. Not to scale, and covered with notations, I used these draft floorplans more as a way to solidify my ideas than as any direct guide for how the level would be constructed. The notations often took the form of questions and it’s only in further iterations that I was able to provide answers.

After filling up two sheets of graph paper with concepts and general layouts I moved onto physically laying out brushes in UnrealED. This allowed me actively move through the space giving me a better sense of scale and the manner in which locations would connect to each other.

(Click Images for larger versions)

Initial Floorplan.

The very first floorplan was extremely basic and didn’t last long, however being able to load the level and move around it in three dimensions really helped to answer some of the questions I had, specifically issues related to the overall size of the level and the type of secondary locations I would need to connect the core locations together.

The floorplan went through a number of revisions over the next hour or so until I had a version that gave a much better idea of the type of layout I was looking for.

Revised Floorplan.

The most obvious differences between this revised floorplan and the original was the inclusion of interior walls to give an idea of how rooms would be separated and the addition of an extra floor in the main building, putting the Control Room and Conference Room on different levels.

The reasoning for this was to add more space both horizontally and vertically. This was also the reason that I added a basement level running underneath the main structure and linking it to the hanger level, thus providing a second route to the hanger and airstrip.

Because the layout is still fairly linear it’s important to try to provide players with different paths between key locations. Though it can be difficult to see from the screenshots, most rooms have at least three separate entry points to ensure there is always an alternative entry or exit available.

Between the main facility and the airstrip there are two routes that can be taken: through the main facility and out the front doors onto the airstrip itself; or through the basement, into the hanger and up onto the airstrip. I’ve made sure that at various point along these two routes it’s possible to change from one route to the other. If players enter combat in either the lower or upper route they are not limited to having to eithe rcontinuing forward or retreat. Depending on where they are they should never be too far away from a point at which they can change to the alternate route.

In addition to allowing players to disengage from combat when needed this type of layout also provides players who know the level the ability to plan ambushes and get the drop on their opponents.

Current Version.

I’ll end with a image showing the current layout of DM-Aerie, and leave discussion of how it reached that stage, and the reasoning behind the changes, for the next Progress Log.

DM-Aerie – Progress Log 1.

Today I started work on a new deathmath level for Unreal Tournament 3, for a number of reasons I’ve decided to maintain a log of it’s development, focusing on creative influences and initial design through the actual process of creation and testing.

 

Initial Design:

I spend a lot of time looking through digital art magazines and similar publications to try and find images to inspire me, it was while paging through the December 2007 issue of ImagineFX that I came across this image by American artist Amir Salehi.

The moment I saw this image I was able to visualise a level concept in my mind, I don’t expect the final DM-Aerie to relate much to this image very closely beyond being a mountain top facility however it was enough to spark an idea. I’ve spent a lot of time using UnrealED so I have a good understanding of the range of materials and static meshes available (My current skillset doesn’t extend to creating my own art assets) and so I knew there were a number of assets available to create a snow covered mountain facility.

Developing the idea I started to visualise a number of mental images of the type of locations such a facility might contain. Conceptually I was envisioning a much smaller facility than the original image depics, in my mind I saw a futuristic take on the Kehlsteinhaus or Eagle’s Nest, Hitler’s mountaintop retreat. I imagined the main facilities built inside the peak of the mountain with a hanger and repair facilities for a private jet of some sort, and a landing strip extending outwards over the edge of the mountain.

Once I’d got a more solid mental image of this initial concept I started to think more specifically about the individual rooms and their layout. At this point I find it useful to look to similar locations in other games, or films, I do this not in an attempt to copy those locations directly instead I use them as a mental temp track to help me better visualise the style of location I’m aiming for.

 

Locations:

The core locations I came up with and the visual references for them are:

Control Room: A roughly circular room located at the front of the structure, similar in layout  the control room seen in the Peragus II  section of Star Wars: Knights of The Old Republic II- The Sith Lords. A split level structure with a number of control banks facing a large window looking out across the mountains. Outside a metal walkway runs underneath the window allowing access to aerials and other sensor equipment built into the mountain itself. A second influence for the visual look of this location was the bridge of the USG Ishimura in Dead Space.

Conference Room: A similar size and shape to the control room but on the opposite side of the facility, above the airstrip. Designed much like any corporate conference room, build around a central table facing a second sweeping window looking out over the mountains and the airstrip. This would be linked to a number of suites making up the living areas of the facility.

Airstrip: An external airstrip carved out of the mountainside and ending in a sheer drop of several hundred feet. Reminiscent of the similar airstrips seen at the start of GoldenEye and Tomorrow Never Dies. The entire exterior of the Aerie would including falling snow and heavy use of volumetric fog. The original idea included an aircraft parked on the airstrip, the engines running, two piercing yellow\orange beacons in the fog. However that was later changed to make way for…

Underground Hanger: An underground hanger and repair facility for the aircraft, built underneath the airstrip itself and featuring a large cargo elevator that when operated would lift the aircraft up onto the airstrip itself. The concept for this location was heavily influenced by the hangers facilities on aircraft carriers, this ship like design would extend to other interior locations, the conceit being that such a construction would be the best way providing protection from the elements.

 

With an overarching concept and these four core locations in mind I had a good idea of the overall structure of the level, enough to start working on a floorplan. I’ll present this initial floorplan, as well as discuss some of the specific level design decisions made during its creation in my next Progress Log.