IceCreamYou | Design

If you don't understand a term, click here.
For a similar discussion, check out Map Design on the UnrealWiki or read Sjosz' thoughts.
I talk about balance a lot in this article. If you would like a discussion about balance itself, click here.

Most of this article is directed at level designers for the Unreal Tournament series, which is fast, multiplayer, and arena-based. However, almost the entire article can be applied to other similar arena games, and much of it is relevant to other game genres as well. Because this article was originally written for UT2004, all of the screenshots are from that game unless otherwise specified.

Sjosz (mapper of CTF-Archanastre and many others) claims of this article: "The read was... nothing really new but interesting enough anyway." I point this out because, if you're a beginning mapper or even an amateur one, it is necessary to work at each element of design that I mention in this article until it becomes second nature. If you have an incredible amount of innate skill, you may realize that you know most of these "rules" without having to be told, but the vast majority of us don't. This article is aimed at those who aren't quite to Sjosz' level in order to clarify, organize, and consolidate all of the random information floating around out there on the internet.

Also, I'd like to add a note that doesn't really fit anywhere else in this article: when creating a map, it is important to note the limitations of computer-controlled players (often known as AI or bots). Often, bots can't pull off difficult, multi-part maneuvers or won't recognize the value of a certain action (like triggering a killzone or a door). Keep in mind your intended audience when you make a map that's likely to have such features: if you want people to play the map offline, it's a good idea to make sure bots can do everything humans can. Avoid close hallways and complicated obstructions.

Good luck, and have fun!

I. Gameplay Elements

A. Flow and Layout Top

Flow is arguably the most important element of a level. It is, loosely, the way a player moves around in a map, collecting items, visiting areas, and interacting with other players and components. Ideally, the flow is such that players do not need to consciously think about where they are going; instead, a player will be guided everywhere in the map by an instinct generated from the map's design. Even better, players should not only be able to glide along various routes, but they should be able to use the map to their advantage: to learn its ins and outs, tricks and traps, and use these against their opponents.

There are several characteristics that lead to good flow. The first and most crucial is layout, or the "blueprint" of the map--where the routes go and what connects where. The second is item placement, and the third is botpathing, and the fourth is audiovisual cues.

Layout can be broken down into several types of "set pieces." There are dead ends, long thin hallways, gigantic open spaces, and rooms with two exits, all of which are often regarded with disdain. Then there are high-connectivity, medium-sized rooms with obstructions and z-axis play, which are usually considered ideal. Are you getting the idea yet? Players must always be able to have multiple options available to them which provide strategy, escape, and challenge. The standard beginner's Room-Corridor-Room layout (RCR) in which each room has exactly two exits, each leading to a long thin hallway, is a terrible layout because a player has only two choices: go forward, or go back. Players cannot be elusive; there is no strategy; and anyone with a flak cannon will win a fight in a hallway. Much better is a room with at least three exits, preferably on different vertical levels, with "stuff" in it to hide behind. The vertical component allows a coveted extra point of strategy: players can drop down unseen on their opponents' heads, making the lower area a riskier location. The various exits allow escape and strategy. And the obstructions and multiple routes provide choice - but a choice that must be casual enough that players can make it subconsciously. As will be discussed in the AudioVisual section, this requires that a player knows without thinking about it where he or she is going.

Generally regarded as unhealthy, these mapping techniques showcase as much skill as it required to create this picture:

In general, wide-open spaces are dangerous because players can be seen and attacked from far away (because of large view distances) and from all directions. Use this element with care--properly administered, and open area can be a great asset, but it can also be a giant sore. Players need a reason to venture into the middle of such an area, which is likely to be the target of many hitscan crosshairs. On the other hand, cramped areas are perfect for spammy weapons and restrict movement more, so these should be used with caution as well. As if your life needed to be any more complicated, there's been something of a revolt in the UT community against what are considered the "standard" maps--medium-sized, self-looping DM maps--so you need to make sure your map is balanced in that respect too, and doesn't rely too much on a standard size to be effective. Dead ends, long thin hallways, and rooms with less than three exits are not always bad either, but limit the number of them that you use and make sure it's apparent when an area is a dead end. These are all high-risk areas, so make sure there's a balance between risk and reward. Also, try to avoid sudden-death areas like lava pits or long falls. These have their place, but they break the flow and they shouldn't be difficult to avoid. Lastly, dead corners (corners with no reason to go there--that serve no gameplay purpose and are empty) can usually be converted into curved walls. If possible try to avoid placing items in corners as well because it's annoying for players to have to go out of their way into a vulnerable and predictable location in order to grab the pickup.

Many exits and an abundance of vertical variation make it harder for your girlfriend to nab you with a flak cannon. Excellent.

Building a strong layout is a multi-step process. The first is the conception; an idea for a room or some sort of gameplay feature might pop into your head. The second is visualization: draw the map on graph paper, by hand. Try to think about how big the level needs to be, how many players it will fit, where the weapons will go, how the map will flow, which areas are weak or risky, and what kind of theme it would support. Optionally, you can then draw the map with a graphics program and try to get its dimensions as exact as possible (though you'll probably end up changing them). Next, build the basic layout in simple Add and Subtract brushes, and run around the map to make sure you've got the scale and flow right. If an area doesn't work well, fix it or dump it. If necessary, put in some botpathing (nothing complicated) and observe some botmatches to see how the bots run around. Consider how much room you'll need for decorations, especially obtrusive ones on the walls and large ones in the middle of a room. Consider also how the visuals might impact the flow: will a flurry of emitter particles obstruct one's view? Will directional arrows point the way? Consider also the z-axis: do you want rooms to have a vertical emphasis or a horizontal one? How will rooms connect between floors? It helps to draw the layout in a graphical program because you can use different layers with varying transparencies to show the different floors. If you're drawing the layout by hand, use different colors or some sort of symbol to indicate a floor change. At this point, it's okay to begin filling in the visuals, which should take much longer. For an experienced amateur Level Designer (often abbreviated "LD") building the layout (once the plan for it is completed, a process that can take a ludicrously long amount of time) might take anywhere from four hours to four days. Creating the rest of the level varies from game to game, but an experienced LD will usually take 2-6 weeks for UT2004 and 4-8 weeks for UT3. For someone still learning, it is not uncommon for the entire process to take as much as six times as long. Unless you're only working a few hours each weekend, you should stop and consider what's going on if you take more than six months to make a map. If you're continually rebuilding parts of it, it may be time to move on to something more productive. Don't get attached to your work just because you've spent time on it; if your time could be better spent, do something else. Mapping takes up too much time and effort to waste it on a futile project. At the same time, try not to start projects and then forget about them or start something else. Most people can't handle more than two projects at once, if that, so try to focus your effort. Write down your ideas and save them until after you're done with what you've been doing.

A quick-and-dirty map layout can often look like bad modern art, but your mom is usually less excited by it.

B. Item Placement Top

For a parallel discussion of item placement by BIOS, click here. Much of this section is relevant to the Unreal Tournament series only, but it's easy to draw parallels.

Crucial to the flow is Item Placement. This tricky category can be summarized in three words: Risk Versus Reward. A dead-end area, dangerous because of the lack of escape routes, might be the perfect location for a powerup because the risk in grabbing it justifies the reward players get if they are successful. A wide-open area might be the perfect place for a sniper, because it's useless in its current location but extremely powerful once the player gets out of the way. And a low-lying area is usually better for other valuable items, since it is more dangerous to be below a foe. These are all just popular conventions which are easily and frequently broken, but they work, and it is crucial that items be perfectly balanced for effective flow. Players have to want to go somewhere (but be able to get there without thinking about it)--but they won't want to go there if there isn't sufficient reward. A player that's loaded up with a Sniper, Flak Cannon, and Shock Rifle is not going to venture out into wide-open, vulnerable space to grab a Bio Rifle. On the other hand, a UDamage in the same position might be worth the loss of a few health points so the player can pull off that extra frag.

Valuable: Worthless:

Weapons are the most obvious balancing factor for maps, but health, ammo, adrenaline, and shields are just as critical. Health and minor Shield pickups prolong life, allowing one to preserve his hard-earned inventory and use it against his less-prepared adversaries. Therefore, health should be placed in easy-to-reach positions, but in places that are open enough to make players consider whether they'll lose more health grabbing the healthpack than staying where they are. Often you'll want to place more than one healthpack in a given location. Shields should be in slightly more vulnerable positions. Health vials, as well, provide some needed energy along ramps and hallways, making those longer routes worth the haul. Adrenaline (UT2004 only) provides a similar function as health vials, but adds an extra twist: when a player garners enough adrenaline, he gets to choose a powerup. This means that control of adrenaline can be crucial to close matches, allowing a flag carrier to escape faster or a deathmatcher to become invisible. Ammo allows players to refresh their weapons, effectively allowing them to continue using their advantage. Quite often ammo for a particular weapon is placed adjacent to the weapon itself, and it is a good idea to place it somewhere nearby so players know where to find it, but games become more interesting when players have to search for their ammo instead of being guaranteed an extra dose. It shouldn't be hard to find, and it's not the most valuable pickup, but if you've ever scored a Rampage you'll know that ammo is necessary to keep you in the running. Many complaints about Beta maps center around ammo being too close to the weapon. A good solution is to put one ammo pack near the weapon but not immediately adjacent, and at least one other pack somewhere else in the map, usually grouped with ammo for other weapons. Ammo can be placed against a wall instead of in the center of a path, but sticking it in a corner is often bad etiquette because corners are uncomfortable positions gameplay-wise and often aren't worth the reward.

Superpickups are the last category of pickups, including the Redeemer, Big Keg O' Health, UDamage, and whatever the more powerful Shield happens to be called (we'll call it by the common abbreviation, 100A). (This category also includes the Painters in UT2004.) It is necessary to understand the difference between the superweapons (Redeemer, UT2004's Painters) and powerups (100A, Big Keg, UDamage): powerups give the player who possesses them an advantage, and can be picked up by multiple players, while superweapons instantly give an entire team the advantage of being able to wipe out the entire other team with one shot. Both should be placed in either remote or dangerous areas to balance the disproportionate reward each superpickup gives a player. This includes separate platforms reachable only by trick-jumps, wide-open areas, or dead ends. Remember that if you place a superpickup in an easily accessible location, action will likely shift towards that area and control of that item will become crucial. On the other hand, depending on the size of the map, it may not be necessary to place any of these at all. If you're wondering how many superpickups you need, consider the frequency with which players will pick one up. Deathmatch maps, being naturally smaller, tend to have fewer superpickups, while UT2004's Onslaught and UT3's Warfare tend to have many more. There aren't any hard rules for the number of superpickups you can use, although there usually isn't more than one or two superweapons even in the biggest maps.

The Redeemer, an ultrapowerful superweapon, should be placed in a difficult-to-reach or very dangerous location unless you consider nuking a valid strategy.

When placing weapons, consider how the location of the weapon will affect how it is used. The example was given earlier of a sniper rifle being placed in an open area; obviously, it is most effective in a high overlook location, but in that case there is hardly any risk in taking it. It must be noted also that some people are against the convention of placing certain weapons near objectives like flags in CTF: nearby link guns can sometimes lead to ultrapowerful link-chains and rocket launchers can create an impossible field of spam. Make players work for that advantage--place weapons in locations that are easy to reach, but where the weapon will not be immediately useful. This means that weapons should be in places where it's easy for players to walk over them without going out of their way. For example, weapons should be in the middle of rooms and halls instead of in small alcoves that require players to divert their attention from playing the game (unless the map was designed for just that to occur, an unusual and risky option; if you do this, it's extremely important that the risk vs. reward be balanced). Know how each of the weapons are most effectively used, and place them appropriately.

Notice the Link Gun's location (lower-right) in proximity to the flag. This is often frowned upon, but the "rule" is viewed with enough complacency that this map, CTF-Archanastre, is often used in competitive matches. It's like staying up late: you know you shouldn't, but everyone does it.

The last note about items is their collective placement. Certain items are (obviously) much more powerful than others, and placing these close together creates a "haven" zone where one powerful player can hang out and keep others from powering up. Control of these areas in a map is crucial to winning but it's also relatively easy, so anything a mapper can do to break up such zones is a good idea. One method is to alternate hitscan and spam weapons; another is to place powerful items at opposite ends of a map. Best is to do both, while ensuring that more valuable items are harder to get. Take into account how long it takes to move directly from one place to another, but also how long a player is actually going to spend in each area. It's especially important not to put powerups next to each other. Consider which weapons complement each other: for example, since the Shock Rifle and Flak Cannon are used in different situations, they should be farther apart than, say, the Flak Cannon and Bio Rifle since the latter two weapons are used essentially the same way.

C. AudioVisual Cues Top

For a discussion specifically of this topic by Chrysaor, click here.

Though often overlooked, a level's visuals play a significant role in how a player moves around the map.

The first and most important item to discuss is Identification. A player must have a sense of place, of knowing where he or she is at all times, and furthermore he must know where he is going. If all the rooms in a map look the same, the player will become confused and not know where to go; but if each room is distinct, players can say "I'm in the fire room" or "I'm in the lobby" and anyone who has played the map will immediately know where they are. Don't vary the theme, but give each major area something visually distinctive. This is critically important in team-based games like TDM and CTF so that teammates can find you. However, it's also important in all gametypes for flow, because a player will be attracted to good-looking areas and will also have a firmer sense of knowing where to go.

These two screenshots from my own CTF-LeftAlive show how the Red and Blue sides of the map can be easily identified. If figuring out where you are is harder than finding your keys, you need to work on your Identification. Click the Thumbnails to enlarge.

The second item is of debatable value: the lines and curves indicated by the visual components of a map can guide players in certain directions. For example, a long hallway with a horizontal texture on the walls will guide players to either end; but a more square room with vertical architecture will cause players to look up. A room that narrows to a point will draw players to that point, and a room with a sloped ceiling will draw players to the shorter side. This is not always true (and it becomes less so as players become more skilled) but it is nonetheless something to think about when constructing your map. Of course, a sign with arrows or directional words serves the same purpose with a more obvious effect.

Notice how the architecture guides the eye to an exit on the far side of the area. Of course, a provocative picture might do a better job. Screenshot from DM-Jaturungkabart from UT1.

Next is the color scheme. It may not seem like the colors you choose for your theme have anything to do with the gameplay, but they actually affect how well players can blend in with their environment. This is especially true of the lighting. Dark lighting and dark texture colors make it difficult to see other players, while light colors and bright lights make it extraordinarily easy to do so. Make sure you find a balance, because otherwise bots will gain quite an advantage. Along the same lines, it is crucial that the architecture of the level does not block or catch players, disrupting flow and causing hassles. Players should be able to slide easily off the architecture, and there should be room to double-jump and dodge without hitting anything in most cases. This ensures the smoothest gameplay possible for players, and not ensuring proper collision can both lower frames-per-second and replay value.

Finally, audio cues are crucial to the way a map is played. They reveal the location of players, adding risk to the map, as well as warn players when they are approaching a trap or dangerous area. The most common examples are a creaking board or a splashing puddle, both of which frequently show up next to powerups. You don't want to over-crowd your map with these, but a few here and there can be a critical touch. Jumppads, teleporters, doors, and lifts absolutely need to have these. Watervolumes have default splash sounds, so you don't need to add anything there.

D. Surprise and Strategy Top

Strategy is the ability to predict where your opponent will move to or appear, and the capability to react to it. Levels promote strategy through the layout. A height variation allows players to jump down on each others' heads, avoid projectiles more easily, and rain projectiles on their opponents. A layout with good connectivity permits players to flee, attack cautiously or aggressively, time moves and pickups, and generally control the map. It also increases the unpredictability--which is good until it reaches a threshold. Basically, there needs to be enough in the map that can create an imbalance between players that will be exploited by the better player. For example, dodge ramps allow quick floor changes, but unskilled players and bots rarely if ever use them. The importance of this is that, as a result of these parts of a map, one player can surprise another. Surprise is by definition unpredictable; therefore, it gives the attacker an overwhelming advantage. It is a level designer's task to provide enough strategic elements that players can tip the balance by their own skill, but not to add so many such elements that the map becomes an impossible hide-and-seek. One possibility that might annoy players is a large number of secret passages or a confusing maze of hallways. Ideas that increase a map's strategy include liftjumps, walljumps, dodge ramps, jumppads, z-axis, connectivity, and a variation of tight (small) and open (large) areas. When you add a component to a map, also consider how much time it takes for a player to pass it; for example, if a player has to crouch to get under a door, that player will be much more vulnerable because he will move slower and will not be able to jump or dodge. Players on lifts are easy targets because their destination is obvious, but a liftjump will send players flying quickly and unpredictably, giving them an edge.

I wrote three key words in the above paragraph that summarize the basis of strategy: unpredictability, imbalance, and surprise. When designing a map layout, keep each of these in mind. (Note, though, that the imbalance should be achieved through a player's hard work, not by a mapper's handing it to him on a silver platter.) Consider how changing view distances or adding a window might affect any of these.

It's worth pointing out that hitscan in UT2004 is generally stronger than spam. It's a flaw in the game that hitscan is so powerful, so pay attention to that when you're considering how your map will be laid out. However, hitscan only works in locations that are open (wide) enough for such accuracy to be realistic. In close-quarters fighting near objectives like CTF flags, spamming the objective with a rocket launcher will deter attackers if not destroy them. Some weapons also have the advantage of lifting players into the air (particularly the Rocket Launcher) and propelling them (especially the Shock Rifle) in the direction of the hit. This is crucial in places where players can easily fall off the map, for instance on a catwalk over a void.

The Red base in my own CTF-LeftAlive is cramped, so using hitscan is like trying to build a computer in the dark. Sometimes it works, and most of the time something blows up.

E. Gimmicks vs. Gameplay Top

One thing that beginning mappers often try to do to the sighs of their more experienced peers is add "new" ideas to the gameplay. The most common "new" idea is to create a map where people play on the walls and ceilings--an idea that is not conventionally possible (in UT2004 anyway), and if it were, it would be above most peoples' skill level. Now, originality and uniqueness isn't bad by any means. However, as Hourences says in his book The Hows and Whys of Level Design, originality is only good while any player can pick up the map and understand what's going on. If an explanation is needed or the map is just too insane, the map has moved from "unique" to "childish" and it will be discarded. These "new" ideas aren't always bad, however. In my map CTF-LeftAlive, I created a new concept for the use of terrain, namely a small height difference that allowed for lots of z-axis play but with many more tactical options than a conventional indoor map because of the openness terrain allows. What is bad is something that inhibits, rather than contributes to gameplay, like doors that randomly shut and block off paths, or jumppads over all the floors (both ideas that have tried and failed). When you're considering what you want to have in your map, remember that convention exists for a reason--it works, and it's fun. Anything you can do is encouraged if it adds something new to the mix, but not if it keeps people from having fun.

One of the most common gimmicks in a map is traps. Like all gameplay additions, there are good traps and bad traps. First of all, note that bots are usually really bad at using traps effectively, but they're good at staying out of them as well. DM-Pressure from UT1 is the classic example of a map with a trap: there was a "pressure chamber" with a valuable pickup in it, but there was a glass window so other players could see anyone trying to grab the powerup and a touch of a button would close the chamber and kill the opponent. This was applauded as a "good" trap because it balanced risk and reward. "Bad" traps are, unfortunately, much more common, and essentially consist of pointless and sudden death. Remember that traps only work if there's a good reason for a player to take the risk of being trapped, and you should never implement a trap that will cause damage to a player for an unknown reason. Always make sure there's a warning - whether visual, aural, or otherwise.

When building a map, consider also the number and size of your traps. Generally, a single trap will serve the purpose, and rarely do multiple traps contribute anything to gameplay. Traps should be noticeable so that all players are aware of them, but not overly large so that they dominate a map (unless that's the intention) or are simply obnoxious. And of course, if you do put a trap in your level, make sure there are the appropriate effects when a player gets caught.

DM-Fractal from UT1 has a trap door which could open when a trigger was shot, dumping unsuspecting players into a pit before they reach the Shield they wanted. Of course, bots love to lemming almost as much as, well, lemmings.

II. AudioVisual Elements

For ideas for themes, click here. For insight into how to come up with ideas, click here.

A. Introduction to Visuals Top

One of the most critical aspects of a map's visual design is making it feel like it's a real place. The map doesn't have to be strictly realistic, but it has to look that way. For example, A six-inch-thick brick wall should not be holding up a humongous warehouse, no matter whether it would be likely in Real Life. It simply looks better to have a wall that is several feet thick, regardless of the plausibility of such a wall in Real Life. The most important principle of realism, however, is making the space feel lived in. It must feel like the place has a purpose, and the best but trickiest way to do this is to fill space.

Probably the best example of creating a world that truly feels lived in is the game Bioshock. If you haven't played it, you should go do that now. Since you're back (or maybe you never left) you can also download the best UT example, Hourences' map Addien-Dwy. This game and map are so excellent visually because there are no long, blank walls; no large empty spaces in rooms; no pointless spaces; and lots of fake exits. As with pretty much everything else in Level Design, these no-no's have their place, but that place is usually the Recycle Bin. In addition, the game and map are not overcrowded; the worlds don't feel as if they were stuffed with random objects, but rather everything in them is congruent, and feels like it is there for a purpose. One could spend hours exploring Rapture, the world of Bioshock, and find everything from creepy audiotapes left behind by former residents to messages written on walls in blood. In Addien, papers litter the ground and books spill off the desk, giving the impression of a struggle and hasty exit. In the dirty world of Bioshock there are no clean spots, no obstructing out-of-place futurism, nothing except what fits the theme. Addien feels genuinely like a house. It has an atmosphere. Someone lived there.

CTF-Addien-Dwy feels like a place--a very nasty one with some rather angry people in it who probably should have been more careful with your papers.

By this time, I hope you understand what I'm getting at. The world you create must be realistic to a point; its space must be filled, but not too much; and above all, everything must be congruent. It's all about balance, just like everything in Level Design.

B. Creating Atmosphere Top

A map's atmosphere is what immerses players in the map. It's what makes a level feel like a place, or rather it is the feeling of the place that the level tries to emulate; a sense of location. The better the atmosphere, the more emotional interest it draws from the player, and the better the experience.

The first part of building an effective atmosphere is to consider the location. What would such a place look like? What sorts of objects would be in it? Why does it exist? How does it work? And, importantly, am I capable of doing this? Answering these questions often involves a Google Images search for reference pictures to reinforce the mental image. Keep a notebook or digital folder of these pictures, and refer to them often. Remember that, while you may have a certain idea that's really cool, there may be a gap between what you're capable of or what works and what you want.

Detail can make or break a map. No one's going to play in a map with no visuals, with the sole exception of DM-BoxCube101-Pro. A map with mediocre visuals will be playable, but won't be well-received even if the gameplay is outstanding. Yet a map with incredible visual detail will get played even if the gameplay flows slower than the dribble down players' chins as they gape in awe. You can achieve such mouth-watering sumptuousness by following the recipe, then baking at 350F for 80 minutes... or you could use a lot of system-slowing environment maps, bump maps, polygons, dynamic lighting, emitters, projectors, moving textures, custom objects, and precise sounds. See, the problem is that good visuals usually means bad frames-per-second (FPS). Everything, as usual, is all about balance. One of the most important aspects of detail is that you keep it consistent. If you have one area that looks amazing and another that looks rushed, the whole map is going to look crappier than if neither area was detailed. Don't go overboard, but adding those little things adds realism and atmosphere. Adding too many of them makes the map feel crowded--you don't need to add each individual blade of grass--but too few of them makes everything bland. Also, try not to mix themes. It will break the visual flow instead of feeling congruent. Each detail should contribute toward a whole, a complete entity which makes up a certain place. One place, not two. On the other hand, a good example of a map in which a mix of themes works really well is AngelMapper's CTF-Phaseshift, but it is rare that this can be pulled off so well.

DM-BoxCube101-Pro has a consistent theme, but is about as aesthetically pleasing as roadkill. DM-1on1-Nagem, however, could be your high school prom queen.

Some of the most important details are trim and movement. Trim should be added nearly everywhere that a texture changes--on corners, around doors and entryways, between walls and floors/ceilings, on stairs, on the sides of catwalks--anywhere that would result in a texture misalginment, to an extent. If you've got trim on every single angle greater than 30 degrees then it will end up looking stupid, but you can put trim nearly everywhere and it will improve the look of your map. Generally, a single trim texture is all that's necessary. Movement, though, is probably the most impressionable part of a map. Everything else is static, so it's the movement (or, in rare cases, the intentional stillness) that really makes the atmosphere. Movement is created through emitters, moving textures, waving banners, fluctuating water, dynamic lights, and movers, each of which has its own uses. They imply that people actually live and work in the world of the game; that instead of the place existing for the sole purpose of running around and shooting people, someone out there actually goes to work there and has an effect on their environment. From plants and their shadows that sway in the wind to moving pillars in a generator to smoke coming out of a far-off chimney, these effects create the illusion of life.

Architecture is also an important factor. A common beginner's mistake is to make rooms that are overly cubic, with blank walls and flat floors and ceilings. Very rarely is this the case in Real Life; for an example, see Sjosz' CTF-Archanastre which uses over 5000 CSG brushes to simulate realistic architecture. Angles, scale, and above all variation are all crucial components of appealing architecture. Consider also the thickness of the walls--for example, Archanastre is set in a stone monastery, so the walls must be appropriately thick--and how the lines created by the architecture guides players towards the horizon. Be careful, because using a lot of brushes often results in BSP holes and difficult texturing, but if you do it correctly there shouldn't be any problems. This isn't a tutorial on components of mapping itself, but a method called "brush sinking" is often useful to fix this and improve blank walls. This last is important as well--try to keep surfaces (not just walls--ceilings and floors too) from being blank and flat. Add paintings, tapestries, built-in columns, trim, alcoves, or anything that can break up the monotony of a blank, flat surface.

Notice the thickness of the walls and columns, the curved architecture, and the way flat surfaces are broken up in CTF-Archanastre. If it helps, consider your favorite actress.

One of the best ways to create atmosphere is by adding potential exits. Closed doors and gates, especially if players can see beyond them into other inaccessible areas, remove the feeling that the player is fighting in a random spot that exists only for that purpose and replaces it with the feeling that the player is fighting in a place, somewhere in a real world. While it doesn't make up for a lack of exits, showing the player that there is an "outside" via a skybox is similarly effective, whether it's through the occasional window or overhead peephole or from open skies. This means that the unplayable area must also look realistic and match the theme; it could be a rolling countryside, a giant lake or ocean, a cityscape, or any variety of skies. Be creative and add as much detail as you can, but remain cautious of the performance cost (lower FPS). Another common atmosphere-enhancer is to add broken pieces of the environment, like tiles, boards, dripping pipes, and concrete.

There are many fake exits in CTF-Archanastre. None are made of silicon.

Though often overlooked, sounds can add a humongous amount of atmosphere to a level. The four most important aspects to consider are amount, volume, source, and choice. Amount and volume go together: it is important that there are not too many sounds in one area or they become dissonance instead of enhancing. The human ear cannot discern more than two to four sounds at once, so try not to have more than three at a time. If you have to have multiple sounds, vary their volume. A dripping pipe can make a good ambient sound, but it's not going to overwhelm a generator. That links in to the source: make sure each sound has a believable place it's coming from. You can have background sounds, like croaking frogs or chirping birds, but only if it fits your theme. "Foreground" sounds like the dripping water needs to come from a visible source of dripping water, and the sound should line up with the drips visually if possible. A good technique for altering sounds to better fit their source is to change the pitch. For example, a high-pitched wind sound could become a low-pitched ghostly whispering. It's important to choose the right sound, because some sounds can be really annoying and others out of place, even if they would fit with the source. Also, try to keep repetition of the same sounds like screams to intervals of several minutes so they're not too repetitive. Don't add in gibbing or shooting sounds because it distracts the player and makes him think there are enemies nearby.

Choice of music is also important. Make sure the tempo of the music fits the pace of your map, and the genre of the music fits your theme. Heavy metal and castles don't mix, but ethereal piano or synthesizer and alien wasteland do. There are a lot of good stock music tracks, but custom ones often add a nice touch. Don't use a track with lyrics though. You may also want to experiment with leaving out the music. Many people turn off music in their settings anyway, and if you want to create a quiet, apprehensive atmosphere in a map that has only occasional clashes (meaning the fighting isn't constant so you could hear the silence between conflicts) leaving out the music might be the way to go. If you do this though, it makes the rest of the sounds that much more important.

Certain actors and instances almost necessitate accompanying sounds. These include jumppads, lifts, doors, steam, puddles, and dripping or streaming water. Commonly, electrical equipment and generators will have a crackle, beep, or low buzz. Background sounds are often natural, including chirping birds, running water, rain, or an electrical hum. The reason I point these out is that the sounds need to be recognizable; the instant anyone hears these sounds, they should recognize where it's coming from. This is especially important from a gameplay perspective, because for instance if a player triggers a sound by walking through a puddle or stepping on a creaking board, enemies should immediately know where they are and why that sound occurred.

Note that sounds are often called SFX, an abbreviation for Sound Effects (say it out loud).

Let off some steam. (CTF-BraveNewWorld)

I keep saying that this element or that element is one of the most important factors in a map. While good lighting can truly make a map outstanding, bad lighting really isn't going to make the entire visual impact horrific in most cases. However, there area a few crucial rules to follow. The first is not to use the default white lighting. It just looks ugly and shows a lack of effort. The second is to pick a congruent color scheme. It should complement the rest of the visuals and consist of three colors: an "ambient" color which makes up most of the lighting, a secondary color to add some variation, and a third color for particular emphasis and more variation. Typical color schemes are blue/yellow/orange, yellow/blue/green, orange/red/purple, green/yellow/blue, and sometimes blue/yellow/red or blue/red/yellow. Third, make sure all lights are sourced, meaning the light must appear to come from a certain location, whether a sun in the skybox or a torch in a hallway. Coronas are often overused, but can be appropriate in many situations if you make sure they're not too big and bright. Last, light/dark contrast, especially via the use of shadows, is a great way to set off a certain area and add a visual appeal to your map. However, nothing and no place in your map should ever be entirely black (not lit), so there should be a very low (about 8) ambient lighting.

Notice how the yellow from flares and otherworldly plants complements the blue ambience in Hourences' DM-Sae. It would almost be soothing if you weren't being shot at.

On the technical side, often multiple lights are needed to achieve the appropriate effect. One or more high-radius, low-intensity lights with low falloff often serve as the initial "glow," while a single low-radius, high-intensity light makes up the main source. Sometimes, low-intensity lights must be added in certain areas to illuminate the shadows there. If an area is too bright, reduce the light intensity or add some shadows using projectors. Projectors are an invaluable tool in a map because they allow detail that would be otherwise unavailable: precise shadows can be cast on terrain and Static Meshes, color from tinted windows can be accurately portrayed, blood decals can be inserted on to the floor, and more.

III. Final Notes Top

This section only applies to non-linear objective-oriented gametypes, namely Capture the Flag and UT2004's Bombing Run.

The general complaint with Asymmetry is that it can never be truly balanced. This is true, to an extent. What asymmetry really means is that one side is going to be easier to exploit for certain people than others. The perfect example is my map CTF-LeftAlive. Its gameplay was highly praised (and audiovisuals as well once I eventually got the theme nailed down after 6 betas) but it is completely asymmetrical. The red base is cramped and so it's dominated by spammier weapons, while the blue base is more open and so hitscan-oriented. This caters to different strengths, and it becomes an intriguing experiment to put hitscan whores on the red team and flak monkeys on the blue team.

To balance this difference, the Rocket Launcher on the red side is placed farther away from the flag, and the Sniper Rifle is placed farther away from the Blue flag. Various other map-specific factors worked into it, but the point is this: if you're making an asymmetrical map, you'd better make darn sure it's balanced. If either team has an advantage that the other team doesn't have, the map has no gameplay. CTF-LeftAlive gives each team advantages and doesn't have any exploitative spots, so it happens to work, but for each asymmetrical map that does work, there are a dozen well-made maps that don't. So before you make an asymmetrical map, you'd better know what you're doing.

Vehicles in Unreal Tournament and often in other games are very powerful weapons when used as intended, and terrible obstructions when not so. When placing vehicles, keep in mind both the increased speed they give to players and the strength of their weapons. Vehicles can run people over; they can fly or drive; and they can shoot weapons that are usually more powerful than handheld ones. All of these are very powerful characteristics, and so they should be placed with care. Gametypes that use vehicles will usually have larger maps, which presents a problem for players on foot (except in UT3, where players have hoverboards) so make sure that there are either enough vehicles for everybody, there's easy ways for players on foot to get around, or the vehicles are intended for restricted use or for a single player. A good example of this last is UT3's VCTF-Containment, where the ultrapowerful Necris Darkwalker is contained in a tiny area in the middle of the arena, but can shoot at a large part of the map. A bad use of vehicles is the Leviathan in one version of UT3's WAR-Torlan; the Leviathan is a gargantuan vehicle, and it frequently gets stuck in the base, blocking humans and confusing the bots.

UT3's Darkwalker (left) and Leviathan (right) cause hell. Sometimes, you can even hurt the other team.

When building a map, keep in mind how many players there will be in it. Beyond the obvious idea that more players means a larger arena, it is crucial to realize that more players means less items available. If there is a single weapon in a base that disappears the first time it's grabbed and doesn't reappear for 30 seconds, the rest of the entire team is out of luck. If there is a single vehicle and a long walk to an objective, players will despise the inaction on the way there. Particularly in Deathmatch maps, where having the best items is essential, a LD must ensure that there are enough supplies within close reach for each player to be able to compete without feeling overwhelmed. A mediocre player who gets the first kill should not be able to lock down the map just because he has a better weapon. Also, consider how many people are likely to be in each room. Try to keep the flow balanced so all of the action isn't focused in one tiny space.

There really are no absolute principles in map design.
--IronMonkey

You shouldn't be asking yourself "what" you can do with the engine, but get ideas first and then ask "how". Otherwise you're limiting your ideas to a simple set of tools.
--Hazel.H

While these two quotes really sum up quite a lot about level design, I'd like to point out that it's almost always necessary to throw out some ideas--not just cool ideas you have before you start that you don't have the skill to do or that aren't possible within the engine, but also those ideas which you've worked away at for months or even years. If at any point during testing you discover that a part of your map isn't fun, I can't stress enough the importance of throwing it out. Don't attach yourself too much to your work or hold on to something just because you've spent a lot of time on it. If your work would be better off without it, get rid of it.

Also, don't make players wait, and don't make them die needlessly. If a player dies, he should know why and it should be in an attempt to get a reward. Players also shouldn't be made to run down long hallways to a meager reward or fall long distances into a void before dying. All of these situations break the flow, and should be avoided as much as possible. There are times when they are useful--for example, a long fall (during which players can't fight or participate in the game) might be sufficient risk to balance going after a powerup in a precarious position.

Wrapping up, I'd like to talk a little about cliché and originality. Hourences makes a strong case for cliché in his book The Hows and Whys of Level Design, essentially arguing that it is most important for players to know immediately what they are dealing with, without having to think about it. And he's right, as is evidenced by his map DM-Rankin, the most played map in UT2004. However, some of the best community maps are so great simply because they broke the rules. AngelMapper (Reconstruct, Phaseshift) and AngelHeart (Goatswood) are famous for doing just that with great results. I broke rules with my own CTF-LeftAlive. If you do something original and unique, though, you absolutely must make sure that it is both an asset to gameplay and not so unique that players can't immediately sync with it. The very best mappers make original maps that players can still relate closely to. It's quite difficult to find a--yes, I'm going to say it again--balance between originality and what is actually fun and absorbing. My favorite examples can be found here.

Finally, if you're serious about mapping, I recommend keeping a notebook with you to quickly jot down any ideas you might have before you forget them. Also, I know this page (and mapping in general) seems intimidating with all the stuff there is to learn, but keep at it--once you get to a certain point, it becomes like second nature, and there's a large community out there willing to help you on your way. You do need to consider how much time you have to dedicate to mapping though. A lot of time and effort goes in to making something worth showing off to your friends, and much, much more into something worth showing off to the community. You may have different reasons for making maps, but above all, have fun!