Search public documentation:


Interested in the Unreal Engine?
Visit the Unreal Technology site.

Looking for jobs and company info?
Check out the Epic games site.

Questions about support via UDN?
Contact the UDN Staff

UE3 Home > Level Editing > Multiplayer Map Theory (Gears of War)
UE3 Home > Level Designer > Multiplayer Map Theory (Gears of War)

Multiplayer Map Theory (Gears of War)


We’re excited about the community interest in making multiplayer maps for Gears of War once the editor becomes available. We thought we’d share what we’ve found works and doesn’t work when designing multiplayer maps for Gears of War. This isn’t a technical “how to use Unreal Ed” doc, but rather a high level design doc specifically for Gears multiplayer maps.

We weren’t sure if we were even going to include MP in Gears simply because it is very easy to make a bad Gears map, and it took us a long time to figure out the “rules” that make MP maps worth playing. We don’t want the community to have to “reinvent the wheel”, and hopefully this doc will save both designers and players from a bunch of frustration.

Every “rule” below can be (and has been) broken, and indeed doing so will often create something interesting or unique. Just be aware of the basics and decide consciously that you are going outside the lines rather then stumble blindly ahead. I would caution a designer against disregarding more than one or two of these in a given map until they understand the repercussions of the various facets discussed below. The ‘fun factor’ of a Gears map can be a house of cards, so be careful when you bump it!

So without further ado, here they are in no particular order:


The most important gameplay factors in Gears are the narrow field of view and restricted rotation rates. It’s like being inside the Batman suit from the original Tim Burton movie. While technically limiting the player, this is also a major reason why Gears feels intimate, as it allows the enemy on screen to be substantially larger and more visible, and it generates tension and vulnerability.

With this in mind, it becomes very important that both teams have a “front” or common direction an enemy is likely to come from. Players need to be able to line up on some cover with your squad and concentrate fire. A random, sprawling arena with no clear directional structures results in enemies coming from any direction, and cover use becomes a burden instead of a benefit because it generally only works in one direction.

Since the player is essentially looking through a cardboard tube, they need to know where to look, because scanning the environment is difficult and will blow the pacing of your map.

The shapes and positioning of your cover can play a big part in defining a front. Set up obvious defense areas and give them a facing that makes their intended use clear to anyone who sees it. Use walls or impassable areas to funnel players through the regions you've defined as combat areas.

A clear example of this is the map “Escalation”, where teams start at the top and bottom of a giant linear staircase. From either spawn position, the player immediately knows which direction the threat will be approaching from. There’s room for lateral movement and flanking along the front, but the general facing is always there.


The Holy Grail moments we look for in Gears playtest sessions come when a flank intentionally happens, it’s fairly earned, and it’s effective and rewarding to execute. The player’s visibility as discussed above plays a huge role in making this happen, especially the bit about being “fair” and not random. When players die, they need to feel like it was through a decision of their own making, and not the designer’s cheap trick or oversight. Players are just fine with dying if they're learning from their experience, or as a result of smart play by opponents, and not just feeling like a victim.

Picture a level made of many tight rooms all connected with random doors. This is a breeding ground for what is absolutely the most frustrating scenario you can give your players. You’re in cover engaged with an enemy, and out of nowhere someone steps out of the doorway behind you and blasts you with the shotgun point blank. You didn’t have a chance to see him coming, you didn’t have a chance to react, and our lethality means that you have little chance of surviving… you were a victim of a random car crash and your decision making had little to do with the outcome. You’ll giggle about this once if you’re the guy with the shotgun, but the guy you shot will most likely never play your map again, or possibly never play Gears again (choosing instead to spend his time ranting about how overpowered the shotgun is). Do us a favor and avoid this, will ya?

But, how do you avoid this? If you take nothing else from this document, take this. A player who is paying attention should always have a chance to see an enemy attempting to flank them, and have a chance to react. Instead of using standard walls for most of your level’s structure, consider using low impassable areas, crevices, or at least put many gaps in your walls to allow an observant player a chance to see what’s going on in the next area.

Flanking someone should involve more than a split second roadie run or evade to the side, it should be a deliberate action and ideally require a bit of risk or exposure to pull off (picture a high ground flanking area, but with little cover or no good escape route).

“Gridlock” is the clearest example of this concept. With the abundance of low cover, lines of sight are running all over the map, allowing a good player to constantly be aware of enemy positions and potential flanks. “Mausoleum” deviated from this substantially, and as a result players can often feel vulnerable from the rear and flanks as unseen enemies scatter through the map.


Players need to feel safe in cover. They need to be able to recognize useful cover at a glance before moving to it, and cover needs to behave predictably because players don’t want to experiment in the middle of a firefight. When cover doesn’t fill these needs, we call it “Fuzzy Cover”. Examples of fuzzy cover are foliage, chain link fences, railing where bullets could pass through the holes, short cover with sloped sides that result in parts of the player being exposed, pillars you can take cover on that are too narrow to actually protect you from fire, small alcoves the player can’t actually fit into, etc. Cover needs to be viable for protection, or scaled down so as to remove it as a safe option from the player’s mind. If you’re hell bent on making a fancy grating or railing, find a way to put something behind it so it is still viable cover for players, but you still have your visuals intact.

Fuzzy cover can also come in the form of solid cover that is positioned in a way that someone can easily be sniped out from behind it. For example if it’s likely that an enemy will be firing on a cover piece from a particular angle, arrange the cover as perpendicular to that position as possible, making it as clear as possible which cover is intended to be used in each likely combat situation.

“TrainStation” did this well, with most cover coming from simple geometric structures that a player can clearly recognize as potential cover. “Mansion” contains some fuzzy cover on the interior areas, where columns and railings often result in players believing they’re safe, when they’re really not.


The reachable path area, or “flow” of a Gears map is not nearly as large and complex as you might see in a game like Unreal Tournament. But this can make the flow and general layout of a Gears map that much more important.

Players should be able to “get” your map within a match, or maybe two. We don’t have the kind of mobility many games have, so don’t expect the player to explore your map thoroughly before they realize how you expect them to play it. Often, a Gears map when viewed from above will be as simple as a figure 8 or an H shape. Plan your flow out in advance, and don’t overcomplicate this step. Gears isn’t about zipping around a circuitous map timing powerup respawns. Most of the real “layout” of a Gears map happens on a smaller combat scale within the branches of your overall flow.

Something we do have in common with levels from other shooters though, is the concept of using lures to move the action around the map. Two to four super weapons placed strategically in a map will definitely affect how the map plays, and this can be used to aid long term replayability of a map. A lure could also be a key cover structure, or a powerful flank position, or a scripted object like a button that triggers an event.

Also, a good combat arena can often work from several angles. A map such as Gridlock will often see a firefight rotate in orientation, occupying the same space, but with action happening on a different axis.

Lastly, look at your map and think about the verticality of it. With our selectively limited methods of moving up or down these height changes, you can use these differences in depth to force a flow around your map without resorting to tall walls. But there are some big ramifications to Z-axis differences on combat. Having high ground on someone will often negate low cover and also increases visibility. Higher ground gives you a combat advantage and you need to build in a tradeoff unless you plan on them being all powerful from their position. Also, you’ll have to adjust cover height and thickness to deviate slightly from the normal standards when trying to fire over a low-cover barrier at an enemy below.

“Shallow” used lures and flow to great effect. The sniper rifles on the side bridges and explosive weapons in the central areas keep players moving around the various paths and make that movement meaningful. “Gridlock” is a more open arena feel with less defined traditional flow, but lures are still used heavily to turn the open arena action into something more tactical. “Rooftops” suffered from the lack of central flanking opportunities, and never quite came together because of this.


Cover is the meat and potatoes of any Gears map. While variation and experimentation is important, here are a few good general guidelines.

In general, low cover is better than tall cover. Picture a wall in the middle of your combat area; you have only the corners to interact with. You can move between the extents of the surface and even manually crouch at the edges, but generally your interactions are limited to the ends of the wall. With walls you’re also greatly limiting player visibility and separating the player from all the action going on over the wall.

If you lower the wall you’ve drastically increased the options for the player. They are now aware of things going on over the wall and can be on the lookout for flanks and enemy movement. They can pop up and shoot from anywhere along the wall, giving them far more choices in firing positions. They can stay crouched and feel sneaky as they maneuver for a better shot. And of course they can mantle over the wall. As soon as they start moving along a high wall they are detached and simply traversing the map, but with a low wall they’re still fully involved in the match.

Don’t overcomplicate your specific cover pieces. Gears' cover system allows for some pretty creative cover node layouts; take it easy on this stuff. It may seem novel to build a castle wall with alternating high and low cover, but trust me, the amount of control confusion and transition animations this generates quickly becomes obnoxious to players trying to use your fancy wall.

Well placed cover should facilitate the “platform game” type movement of cover slipping and swat-turning from cover to cover. Try to move through your map as a test while minimizing your time out of cover and see if you can “jungle swing” through your map. Also try roadie running through it for times when you’ll need to traverse the map quickly, and look for sticky points.

Lastly, avoid cover crowding. Cover should never feel like something obtrusive that you're stumbling around. Cover is placed to facilitate ranged combat, but overpopulate an area and you actually have the opposite effect, giving people the ability to close on each other with impunity, devolving the game into shotguns and chainsaw dancing. You need a "no man's land" between cover positions, or the cover itself loses its importance. Those open areas devoid of cover are dramatic dashes just waiting to happen. They give players interesting decisions to make and also increase the importance of mobility and roadie runs.

“Depot” has about the right balance of cover and movement. Most clusters of cover have stretches of vulnerability between them, which facilitates ranged combat and clever flanking. “Mausoleum” is an example of a map with crowded cover. Dense tombstone clusters fit the flavor of the map, but movement through key combat areas often feels clunky and restricted.


Bigger is not always better. Before you design a map with 5 separate large combat areas, you need to realize the value of a tighter environment. There are two aspects of scale here: literal size of map, and gameplay space.

Literal scale first. We made core gameplay tradeoffs to get the enemies big on your screen (slower movement speed, weapon effectiveness over distance, etc…) to avoid shooting at 4 pixel tall enemies from a half mile away. You should take this into account with your map and keep combat reasonably close. Gears is simply not designed for sniping at enemies on the other side of a desert – it’s fun in some games, but it’s not likely to be fun in Gears regardless of your preferences. Keep the combat distances at ranges close enough that an enemy may be a threat if they decide to charge in, but far enough away that you should be able to stop a charging chainsaw monkey before he could reach you. You should ideally be able to see hit impacts on an enemy clearly from one cover area to another.

Then there’s gameplay space. Your players need to be playing in the same map and not branched off doing their own thing. A parking garage might be an example of too much gameplay space. Players may technically be close as they’re engaging each other, but you don’t want to have to search 5 floors for an enemy to shoot, especially when most gametypes have dead players spectating, watching you hunt for the last enemy. Pick a single primary combat region, and if there’s an outlying area of your map that isn’t relevant to what’s going on in that central area, consider deleting it… you’re probably only fragmenting your combat instead of trying to get everyone involved in one common experience.

You want all your players to be relevant to each other, you want them close enough to influence the action, and you don’t want players running an extended side mission to get a grenade you cleverly placed 3 blocks away.

Almost all of our maps had a central combat area measuring around 4000 unreal units wide. Some maps, such as “RavenDown” used a much tighter scale to give the map a unique feel, but such a deviation will dominate the design of your play area.


While it might not seem that important to the bulk of the combat in a map, the initial approach to a map from the spawn area has three effects on the map as a whole.

Of primary importance is having a line of visibility to areas enemies are likely to enter the map from. Even though this view is far enough away that combat is probably ineffective, it’s incredibly useful to be able to tell “Hey, 3 guys went high, 1 guy’s going for the Boomshot! It’s like seeing the “play” unfold that the enemy team is calling and allows you to adjust your tactics for the round.

The initial approach can establish a good deal of the feel of the map and offer potential path choices to the teams. Players will generally take this opportunity to take a few shots at the opposing team if they can see them, and even though it’s not going to kill anyone from so far out, it greatly enhances the feeling of drama as the round opens and everyone is roadie running forward under distant fire.

It also serves as a nice breather between rounds. If you were the last one to die in the previous round, you’ll probably appreciate the 3-5 seconds to collect your thoughts and think about what to try next while you’re running into the next fight.

Most of our maps start with a lengthy area for roadie running until enemies are encountered. Maps like “Gridlock”, “WarMachine”, and “Depot” have such long view distances that they reward the observant players who can see which direction the enemies scatter to once they leave the spawn areas.


How will people remember your map? You want people to play it again, so what makes your map stand out from others? You need a hook, a theme, a gimmick, something identifiable to players.

This doesn’t have to be a heavy-handed overriding visual element such as a map on the extended hand of a giant monkey statue. Nor does the gimmick have to dominate the gameplay of the map. People aren’t likely to enjoy running around a dense minefield. The layout and gameplay need to stand on their own merit, but the hook gives the map something to cling on to.

This could be a central visual landmark such as a series of arches stretching across the map, or perhaps it’s really windy and some debris rolls by periodically amongst the audio cues for wind gusts, or maybe it is a gameplay gimmick such as a ticking bomb in the center of the map.

Whatever you’ve done to tie your map into the player’s consciousness, bring that into the name of the map as well. MP-Guordiosa, or MP-Dianima, or MP-Twjfslaek… those don’t mean anything. The may be neat and remind you of your first dog or your favorite D&D character, but to the players who download maps and try to keep them organized in their heads they mean nothing and earn you no love. That map name is prime advertising real estate. Make sure it sells your level, provides useful info, and is simple and memorable. No one ever says “let’s play that Gears map with the big mansion in it… what was that called?”

The giant train passing through the “TrainStation” map is a great example of a gimmick that doesn’t dominate the gameplay and yet gives the map a unique and memorable quality. “ClockTower”, while in theory getting its name from the structure in the center of the map, didn’t prove as memorable as it could have been. In retrospect it wasn’t likely players would be looking upward to notice the structure.


Unreal Tournament 2003 was when we first started to really pay attention to visual clutter issues. We wanted super detailed environments, but we weren’t careful about what we wished for... in no time we realized the difficulty in perceiving enemies or other important gameplay elements against intricate backgrounds. With Gears we started finding a balance, but it's almost entirely up to the judgement and restraint of the level designer to facilitate this. It's very easy to go overboard on mesh details in your world; you have to resist that urge. It's getting to the point where you can place as much meshwork as you want in a map for free; but what you're not suffering from on performance, you're paying for in playability. Here are some general tips.

-Use clearly different textures to contrast floor surfaces and walls, so a player can see a "floor plan" as they look around. Ideally cover should be textured to contrast as well.

-Use depth fog to help clarify the level's depth complexity (even an unlit level can be navigated with fog alone).

-Normal mapped simple flat surfaces still look fantastic. It's OK for a wall to just be a wall sometimes. Every surface doesn't have to be slathered in pipes, rubble, or random visual noise.

-Use lighting to guide players through your map. When you have an area with a specific exit like a door or arch, move to the far side of the room, squint at the screen and ask yourself if you can tell where the exit is. If the screen is a grey sea of muddy noise, use a contrasting light source in the exit to catch the player's eye. Contrary to that, if something catches your eye that isn't an exit or important feature, tone it back so as not to misdirect the player.

-When actively editing your map, place dummy character models around the map to use as a scale reference as well as a guide to see how the characters "pop" in your areas, and see how the light is affecting them. Place character-only lights in areas where enemies could use more clarity.

-Be careful with overly stylized post-processing settings. While they do a fantastic job of tying your scene together and uniting it visually, they can create a real challenge for the player as they try to distinguish friend from foe.

-Again, avoid placing any ancillary meshwork that might be confused as "fuzzy cover".


So there you go. These are the pillars of our cover-based level design to this day. These aren’t meant to limit your creativity -- deviating from these will be what gives your map a unique feel and that should absolutely be encouraged. Just keep these concepts in mind as you experiment, and use these as your “control” group.

I think I speak for everyone at Epic when I say we can’t wait to see what you guys can do with the tools once they’re available. You guys are going to have so much fun working with this stuff, we all did.

(Big thanks to Mike Capps, Cliff Bleszinski, Jim Brown, Dave Nash, Dave Ewing, and Dave Spalinski for their contributions to this doc, as well as all the content guys who helped us learn these lessons)


You don't necessarily have to be a great player to make a great map; most of the LDs here will get trashed playing against hardcore guys online at this point. But you should understand what makes the game fun to both new players and experienced players. Anyone designing layouts of maps here enjoy playing the game a great deal regardless of skill level, and the vast majority of people doing artwork in the maps are very fond of the game as well.

It's a safe bet that most people making Gears maps are actually trying to make maps that play well, in the hopes that others will download and enjoy it.

If you don't like the way Gears plays, you can still use the editor to do all kinds of cool stuff, especially with Kismet as a tool to prototype new game ideas. If you're purely interested in visuals, the editor can be a great sandbox as well, and if you get the hang of it you could easily land a job with a few gorgeous maps at many UE3 licensees who (trust me) are always look for good visual meshers.