April 20, 2006
Question Responses
well here are some responses to the questions I drew up on MultiPlayer Level Design (I'll break them down by question):
1) What are 5 important features that make a great Multi-player FPS level?
------------------Email 1
1. Defensible chokepoints that aren't insurmountable
2. Intelligent spawn locations
3. Clear strategic locations
4. Mix of locations such that no single weapon is dominant
5. Obvious staging points
---------------Email 2
1) Theme-...even the most familiar settings whether it's counter-terror or WW2 can be made interesting. SWAT vs Terrorists in a warehouse or middle-eastern row of huts is boring. SWAT vs. Terrorists in an eccentric billionaires mansion or a natural history museum would be cool. Do your homework!!! Research your setting.
2) Try to prevent exploitation for square one-...unless the LD has some say in weapon balancing or networking code, his level is really only the place he can keep exploitation in check. Not to sound cynical, but if there is any chance that players can exploit the flaws of a map and grief other players right off the server, they will. I think some mappers build their levels with their heart in the right place and have a lot of ideas for ways to make their map interesting, but overlook the sniping nests or other map features that the player can abuse. Unintended map tricks can be cool as long as they don't afford a griefer with a chance to jerk other players around with impunity.
Keep these things in mind from the moment the player spawns into the map. Nothing seems to piss off other players like being spawn camped by a bottom feeder looking to pad his score. In my RtCW maps, I tried to spawn players in areas that were inaccessible to enemy players. If you are making an objective based map (plant a bomb, mess with a computer, steal an item, etc) then you probably shouldn't place the defending teams spawn point too close to the objective otherwise it may tip the balance too steeply in their favor. If the defenders spawn point isn't near the objective or along a main artery of map flow, then it stands to reason that any attackers near the spawn are there to camp for some easy points and should be easily dispatched. Allowing spawning players to drop down into the map, spawn behind cover, or near a gun emplacements goes a long way to giving them the chance to get going on the right foot.
3) Flow-...keep in mind your player loads and scale the map accordingly (don't make a sprawling battlefield for a 2x2 or 4x4 game). Once you've got an idea how many people you're going to have running around you can start to lay out the main arteries of player traffic. Choke points can be good for funneling the action but can also result in a stalemate. I've found that you rarely want to have more than 2 routes (1 main/ 1 alternate) between key positions. Too many routes can result in a map that is mazey and confusing. Too many routes can also result in fragmented teams and a lack of cooperation. Also, one rule that always seems to apply is that dead-ends are a big No No. Dead-ends in games like Enemy Territory were ok because they were simply obstacles that could be overcome if the team cooperated (so they weren't true dead-ends).
4) Visual Variety-...while consistancy is important, you don't want the map to all look the same. Make sure that you are not decorating purely for the sake of aesthetics. Unique mesh work, lighting, and texturing can not only make a level prettier, it can also provide clear landmarks and indicate path flow.
5) Efficiency-...high frame rates are extremely important, especially in multiplayer. Variety is great but haphazardly placing meshes, using too many textures, or too many lights can easily result in cluttered looking maps that overtax system resources. Next-gen tools allow for terrific looking maps but they also allow the builder to max out his texture memory and poly counts quicker than ever. There are many more things to take into consideration when placing environment models. Engines render meshes differently, but multiple lights or shader effects can result in a model being rendered multiple times per frame. For example...A 500 poly model can really be a 2000+ poly model if its being illuminated by 3 or more lights. This must be rendered each frame. A 500 poly model rendered at 30 frames per second can result in a 60,000 polys per second. I invite a programmer to correct me since I'm sure I'm not stating things as accurately as possible but you get The Idea.
-----------------------Email 3
Supports strategy - Strategies in these types of games are simple, generally a strategy for the attackers is what route most of the team is going to take and which bomb site they are going to plant at and the strategy for the defenders is which bomb site they are going to primarily defend (since they usually have to defend both). Supporting strategy means forcing the players on the team to commit to a strategy (ie a route) by using limited routes and chokepoints to prevent/discourage players from switching to a different route mid battle. This creates a battle line at the chokepoint and allows players behind the battleline to react to the battle by coming in as reserves (ie joining the battle), holding secondary lines of defence behind the battle line, or attempting to flank the battle. Without limited routes and chokepoints the battleline spreads out to cover the entire middle of the map and you cant effectively react to it because their is no definate strategy to react to.
Is balanced - Not nessarily that each team has an equal chance of winning (although that is nice) but that each player has an equal chance to get kills. Dust2 is the obvious classic example: on the team strategy level, each of the CTs are forced to commit to defending either A or B, which splits the team, allowing an organized T team to strike with its entire team vs half the defending team, which neutralizes the defenders advantage. Dust1 is very hard for the Ts to attack but its also very hard for the CTs to attack the T spawn, so its equal in the sense that players on each team have an equal chance of getting kills by defending their territories around the main chokepoint.
Cover supports the micro level of tactics - Dust2 is also balanced in cover, attackers coming down long A have just a few corners for cover on that road but defenders have basically the same limited set of choices. Attackers coming from short A have a low wall and some large boxes and defenders have the same. (agruably the routes into B both favor the defenders over the attackers explaining why the most common Terrorist strategy is to attack A.)
Supports variety of tactics - Tactics is the term I'm using to differenciate between things like close range combat and long range combat. Militia supports close range tactics in the sewers and in the house itself and long range tactics (ie sniping) in the approach to the house and in the backyard of the house. Aztec has the water route which supports long range sniping and the bridge route which supports close range firepower.
Is possible/easy to learn - the most popular CS maps tend to be fairly simple, with a "knowable" set of options for where defenders can be hiding. This allows players to guess where enemies are based on map knowledge and more importantly not feel cheated when they guess wrong. If there are too many choices then it is pointless to guess and the only real option players have is to "get lucky" this puts players in a powerless situation which is no fun.
----------------Email 4
1) Framerate/Performance
2) Not exploitable
3) At least 2 primary conflict points (usually 2 is sufficient)
4) Deliberately placed Landmarks (includes colors)
5) Reasonably even and natural distribution of flow (cut all dead areas that see virtually no action)
6) Interesting setting that provides a backdrop that "feels good" to be moving around in.
I can't do 5()@*#*()
--------------Email 5
a. It places the weight of success on the team and their usage of the map and it's resources, not on the map itself; ie: the map is balanced.
b. Frame rate. If it doesn't run in the wild, no one will play it.
c. Easy to play out of the gate, with depth to support alternative strategies.
d. Meshes with the game mechanics to enable multiple valid strategies per team.
e. Leverages/forces team skills/weapons.
============================
2) Are there any features that are important but for some reason always seem to be left out?
-------------------Email 2
See above for spawn protection. I'm always annoyed at how rarely the player is given any sort of protection by the mapper (things like temporarily invincibility is a Design/Code issue) and is easy prey for spawn campers.
-------------------Email 3
---No. There is no magical x factor that is missing from the most popular maps of CS
-------------------EMail 4
Play-testing your map for a few weeks with a variety of people before releasing. Playtesting is a feature.
------------------Email 5
No, every map is a tradeoff. Frame rate is the only factor which is map independent.
-----------------
=================
3) What is your basic strategy when playing these types of levels? Do you tend to charge the enemy? Or hang back (Camp) and wait at a strategic location while your teammates soften up the enemy?
--------------------Email 2
That really depends on the game. BF2 and Enemy Territory are both my favorites. I either play as the sneaky covert saboteur who's stealing uniforms and planting bombs or I play as the engineer in a tank and fix other vehicles. Even when I have a tank, I never charge in. I treat it like a 70 ton rifle that fire depleted uranium shells. :) I'll sometimes do drive-bys with a tank, but I never charge into the thick of fight like you might in deathmatch or in other less tactical games.
-------------------Email 3
---Not relevant, your stategy should be dictated by the flow of the match and the play styles of your team and your opponents and you should be able to use a variety of strategies/tactics to succeed.
------------------Email 4
Depending on the gametype hopefully a player can pick either one and be reasonably successful. Especially in team based games, a person who literally does nothing but defend against people that break through the front line is completely viable and important, even if they get less kills than everyone else. The satisfaction can come from just contributing to the team that won. If you stopped them from capturing the critical flag in Day of Defeat giving your front line the chance to move in and take theirs, you will feel like you won the game even though you only got one kill.
-------------------Email 5
Depends on the teams playing.
I've been in many pickup games where the first couple of rounds are complete unmitigated disasters one way or the other until the team learns what their teammates are doing, and/or what the enemy is doing.
A team that's played together before will usually have several strategies, and will know which one to use based on on where they see the other team's players, or preference.
============================
4) What do you feel is the main reasons that these types of levels have remained so popular? (i.e. DE_Dust for for Counter-Strike, and other
maps that tend to have 24/7 servers?)
---------------------Email 2
I think a lot of these maps are popular because they had simple objectives (plant bomb/ defend target) and offered some alternative routes instead of funneling players into one meatgrinder. I personally think CS and DE_Dust have been far surpassed by other games in recent years, but they were unique for it's time and have a built in audience that's huge. I know I was certainly hooked on CS for a good year. Also, never underestimate the popularity potential of free games. :)
------------------Email 3
In game modes like CS, map knowledge is critical and players who dont know the map will be at a big disadvantage vs those who do, since players who dont know the maps will rush into chokepoints and quickly die, or will be unable to use the information they are getting about the battle (ie "need help at A - where the fuck is A?"). Thus players will do better and have more fun (since players like to do well in a game) in maps they know. So there is very little incentive for players to want to play new maps other than boredom which can take a long time before it becomes a serious factor in map choice.
Dust and the other maps are also simple enough that players can learn them quickly.
--------------------Email 4
Momentum. Once a game gets enough players the popular maps almost *never* change. Early in a game's career certain maps stand out (for the above listed quality reasons) but from there they are just what people know and understand and they aren't playing it for the map itself so much as because they then only have to focus on improving their personal strategy and skill levels.
However some extremely early maps which really aren't all that great will still stay popular simply since people have been playing it from the start and they simply don't want something new. So long as they aren't exploitable they will ride the tails of the really great maps.
------------------Email 5
The map is balanced to support the opposing goals of both teams, runs well, is easy to learn but has depth for continued learning and play, and meshes well with game mechanics to force actual team support and play.
No one blames the map when they lose, no one praises the map when they win.
========================
5) What is your most memorable moment when playing this type of FPS level? (This moment doesn't have to be just from the games mentioned
above, if you have something from Halo or other console games please feel free to elaborate)
------------------------Email 1
Bringing in a Galaxy way high, dumping troops into a base, getting shot down, and somehow limping to safety, bailing, and proceeding on foot, all in PlanetSide. That game was SO close to being great.
My favorite MP team based online level is probably St. Mere Eglise in CoD2, but there were quite a few CTF maps for Quake and Quake 2 that I played a lot but can't remember. Good times.
-------------------------Email 2
Such a tough question. Don't make me choose! hmm...
Enemy Territory...Playing as the Covert-Ops character lead to some genuinely thrilling gameplay moments in the Oiltank and Beachhead maps. Stabbing unwary enemy soldiers, stealing their uniforms, spotting mines, and sneaking a comrade engineer through restricted doors, and keeping him covered with an almost overheated Sten was was incredibly tense. INSANE.
BF2...it (was)'s all about the DeathBus:
Working like crazy to repair a BlackHawk while 2 miniguns scream away like power tools drilling thousands of rounds into an enemy cpature point was intoxicating.
-------------------EMail 3
When I use an unusual/out of the way path to out flank my enemies. For example in Dust2 I'm guarding A and the terrorists plant at B and I use the back routes to attack B from the same direction as the terrorist spawn.
-----------------Email 4
2fort4 in classic Team Fortress for Quake 1. Quite possibly the simplest and most basic teamplay setup you could ever create and I played it for easily over 1000 hours. No one moment could really stand out, like Zied said... anytime I come up with a clever or unusual tactic that actually works (as opposed to just being lucky) is very satisfying. Lots of those being a heavy machine gunner in Day of Defeat.
=================
6) If you have the time it would be helpful to my research, if you would detail your work process for designing a Level. (If it is not something you care to share I completely understand, and thank you again for taking time to answer these questions).
-------------Email 4
Sadly I haven't made a map for a few years now. :( And I've also not been a project that got popular enough to have a map become a "classic".
That said, the order has almost always been:
- Think about settings (either ones I've wanted to do or search online for something that seems interesting).
- Research photos and articles about the setting to develop key areas and a visual look.
- Plan out on paper.
- Build basic shell and run around in it non-stop to make sure the distances feel good and the connections make sense.
- Quick playtest with someone else.
- Visual pass.
- Playtest again.
- Polish pass on weapons and visuals that detract.
- Playtest a lot.
- Polish some more.
- Cry as you release it to the public.
That's simplistic. Framerate is always a concern from step 1 till release.
BTW, regarding map momentum. As an example when Day of Defeat Steam came out... I instantly loaded up the new versions of the old maps I had played. I didn't even think about it at the time, but it was just what I naturally went to. I don't even know what the new maps are as I just look for servers that are running the maps I already know. There might be some brilliant new maps in there but they aren't getting the playtime because lots of people do what I did as well.
==============
Random Room
---------Email 5
I would only add that, in my view, the best MP maps are "expandable" as the player becomes more skillful. By this I mean, new routes are opened as the player exploits higher level movement, such as strafe jumping, rocket jumping, upcrate running with momentum jumping (heh), or the dodging and wall jumping in UT2k4. Maps like DM-Rankin, from the Ut2k4 demo, are a good example of this sort of thing -- it features aluminum corrugate ramps (just leaning against the wall) that one can use to jump up, but it's rather difficult for the noob. So they become paths not open to everyone -- thus encouraging a higher level of play and mobility. A more recent example, perhaps, is the CTF map Dead Wood featured in the last FEAR patch -- jumping up on the mountain tops is a tricky manuever not many can pull off, but it opens up a new strategy.
--
From Spencer Boomhower (independent contractor)
4) What do you feel is the main reasons that these types of levels have remained so popular? (i.e. DE_Dust for Counter-Strike, and other
maps that tend to have 24/7 servers?)
I think one reason these stay popular, and one I don't see mentioned too often, is lighting. Specifically, these levels are bright, vivid, and visually uncluttered. It's very appealing to be able to easily see the enemy players as sharply defined silhouettes against brightly lit walls. I'm thinking specifically of (in CounterStrike) Dust, Italy, the original, pre-Source Aztec, and whatever that British castle one was. These levels also have a fair amount of breathing room.
This is in contrast to dark, serpentine levels with lots of nooks and crannies for people to hide in. There's nothing wrong with cramped, dark levels (as long as they're not too dark), but they give a slight advantage to experienced players who know to check behind that one crate around that one corner. So they're simply less inviting to new players.
I could see it being a significant number of newbie players who keep these 24/7 levels from emptying out. They, being the prey, are drawn more to the bright open savannahs, and less to the cold dark caves :)
It's worth pointing out a that Dust isn't open, bright, and clean to the exclusion of all else. It still has a bit of darkness for contrast. But you can see these areas coming and know, "ah, now I'm going into a cramped, dark hall... best be extra careful."
I think there's something similar that goes on in BF2, where the more popular 24/7 levels are bright, daylit levels. But the one nighttime level I can think of (with the string of islands) doesn't get much play. I don't play a lot of BF2, though, so I might be off-base there.
--
What are 5 important features that make a great Multi-player FPS level?
It's hard to pick a top five but here are a few that I think about.
Flow - few or no dead ends; multiple exits (more than two) from most locations; always channeling the player back to key locations; etc.
Well-tested item and powerup placement
Plenty of places for opportunity fire
Movement smoothness - nonvisible ramp planes over stairs; nonvisible planes to smooth inner corners; etc.
Interesting, if not breathtaking, locations
I like TW's point, btw. I'd be thinking about that too.
detail your work process for designing a Level
Here's a general plan. I don't know how useful too much more detail would be...
Thinking about the game type (deathmatch, teamplay, whatever) and the theme (which will usually be provided for you), choose a scenario. Most of these are well-known and ready for your particular twist.
Draw a quick "Zork map" style flow chart, then draw a slightly more detailed version to refer to but don't go too far.
Since one of the most important things to do is to get testers running around in the level asap, rough out the basic level with no details. It doesn't have to look pretty yet.
Drop spawn points, most of which will be obvious or already planned, a few weapons and powerups just to get things rolling, and invite the testers in.
Listen to their comments, get in there and run through the level with them, and make changes to the basic layout as appropriate.
Get your basic lighting scheme happening. Goal points and things like teleporter locations should be added at this time as well, at least their initial locations.
Test again.
Place/adjust the locations of/tweak items and powerups. In fact, be ready to adjust everything including spawn point locations (however the layout should be mostly fine by this time), and test again (and again). Begin to really finesse the level.
Lastly, finish visually detailing the level. Test again, make sure that everything feels smooth, looks exceptional and plays well, and you're out of there.
