Contents: 0. Introduction 1. Pace 1.1. Size 1.2. Chokepoints 1.3. Geometry 2. Invidivual classes 2.1. Scout 2.2. Solly 2.3. Demo 2.4. Medic/Heavy/Sniper 3. Other concepts 3.1. Rollouts/Midfights 3.2. Hiding spots 3.2. Learning curve 4. Conclusion 0. Introduction There are only few mappers with experience in competitive TF2, so I am writing this to help those who are interested in making a map for the competitive scene. A lot of the information in here is applicable to public play as well, so it might still be of interest for everyone. It will be 100% focussed on gameplay, and it will mention Badlands a lot, as it is the quintessential competitive map. I am gonna start with observations about how maps play, and what affects gameplay in competitive play, and will try to work that into useable suggestions for maps. I will assume knowledge of the maps (most of them are official+obscure and gullywash) and of course the game, but will try to make my points understandable without any knowledge of competitive play. A lot of these can be seen in/explained with public play as well anyway, since a good map is a good map, no matter what. It's just that in 6vs6 smaller issues within the map will have a larger effect. (while other issues might have a smaller effect). This article covers a bit of it, but is not focussed on competitive play, and I want to go deeper (cue inception meme): http://forums.tf2maps.net/showthread.php?t=13399 And if you have absolutely no idea of 6vs6, consider reading this first: http://forums.steampowered.com/forums/showthread.php?t=792794 Furthermore here is a list of dos and don'ts for a comp map: http://www.nodraw.net/2009/12/mapping-for-a-competitive-audience/ The discussion in this thread should be quite insightful as well 1. Pace Starting with something that is quite hard to understand, pace. The pace of a map in a competitive environment is very important. When there is a problem with the pace of a map in a 6vs6 environment, it is pretty much always that the map plays too slow. So what affects the pace of a map? 1.1. Size At the most basic level, it is the size. The longer it takes players to reach the objective, the slower the map is, pretty straight forward. A map must not be too big. Granary is pretty much at the limit of how big a map should be and therefore how far travel distances should be for 6vs6. A map that is too small doesn't work either, since it will be a messy spamfest without much tactical depth (think Junction). Notice that by big, I mainly refer to the travel distances and therefore the time it gets from A to B here. Longmap is long 1.1. Chokepoints Something a bit more complex are chokepoints. Chokepoints are one of the main deciding factor of which position a team is going to hold. At the simplest level, the narrower and longer a choke is, the easier it is to spam and therefore harder it is to push through it. There are a lot more deciding factors though. First of all the question, if the choke point has a roof. Most chokepoints in the competitive map pool have roofs above them. The most popular example of an open one is the one on Badlands leading from mid to spire. If there is no roof, it is a lot easier for soldiers or the demoman to jump through. Additionaly there can't be sticky traps above the doorway, eliminating one spot players have to worry about and check when trying to push without uber. The Badlands choke, prime example for a roofless chokepoint. Staying on the choke at Badlands, the height difference on either side is an additional factor. On the chokepoint on badlands, the team holding mid has the height advantage, making it quite a bit harder to push from spire to mid than the other way around. With the middle choke from 2nd to mid on obscure, players are at a height disadvantage on either side. So while that chokepoint itself is quite wide, the rather large height advantage a team has to overcome makes it harder to push through in both directions. Doors are also a way of slowing down pushes. First of all, they are prime positions for sticky traps, so players have to be careful when opening them. They also block sightlines when closed (who would have guessed). This can help fight some nasty sniper sightlines, so the sniper needs someone to open the door for him. It also means that players have to approach the door to see if someone is behind is. With less information about enemy positions, making the best decision is harder. Good examples of door useage are Granary 2nd and Badlands lobby. Good useage of doors 1.3. Geometry With chokepoints out of the way, there are other factors that change the pace of a map. I want to distinguish the map geometry from the chokepoints issue. One point I want to ramble on about first is the size of the combat areas. Seeing that narrow chokepoints slow a map down, one could make the assumption that large open areas make it play faster. That's wrong though. What it comes down to in areas like the yard between 2nd and mid on Granary, or the 2nd point on obscure, the defending team can force the enemy uber quite early, while hanging back a bit. To get everyone into the fight, the attacking medic will have to split his uber a lot, resulting in a much shorter uber and making it a losing battle. But the type of choke leading into the area also has a large effect on this though. On badlands, while the area at spire, from the stonebridge to the choke, is quite a large open area itself, it is quite easy for the attackers to get in and out. This makes it harder to force the uber early, so the attacking team can more easily get in 'for free.' A large open area also means, that snipers will be really strong. Playing against a really good sniper will make you more anxious to push, especially when you can't properly pressure him or hide from him. The other deciding factor in map geometry are height advantages. This is quite straightforward and shouldn't need much explaining. Giving either team decisive height advantages favors that team in the battle and can make points too hard to attack or defend. How strong a certain position is is dependent on a couple of factors besides height advantage and chokes as well. How easy it is to get up there? The harder for every class it is to get there, the better for the player up there. This can be different for the different classes as well. How easy is it to avoid damage there? For example the Badlands balcony gives less of an advantage, because it is not wide enough to avoid splash damage and a soldier can explode rockets on the wall behind it. Props and smaller geometry can be used in a number of ways to affect gameplay, but they don't affect the pace of the map that much. They might do a bit, when allowing for strong sticky traps. 1.4. Item Placement Item placement has importance on the pace as well. Ammo placement isn't as important for the pace of the map itself (you could argue either way, too much ammo promotes spam, not enough ammo can delay pushes), but ammo can be a huge annoyance for players. (Most?) Players won't actualy mind having a lot of ammo around, since that eliminates one constant worry and let's you focus on the enemies. Engineers also aren't really a concern in 6vs6. Ammo should be easy to get to. (think of a random position on Badlands (or your map); now think of the nearest ammo pack and how to get there; repeat). In my eyes there shouldn't be too much ammo on the map, and you should mostly spread small packs, while having a few medium packs available in central locations. Health has a bigger effect on the pace than ammo. Not enough health on the map, and players on the flank will have to fall back to the medic and stress his finite heal rate even more, resulting in a slower map with less combat. If there is too much health, there are less chances of getting a pick on a player, since players can replenish their health way faster. Depending on the placement, different classes will benefit from a healthpack. A medic will pick up more easy to get to healthpack, while a soldier will pick up health packs on higher locations more often. Two small healthkits mainly used by soldiers and demomen after jumping A healthpack for on the fly pickup Small healthkits in certain locations will help mitigate some of the damage players take from spam or from jumping. These generally are placed somewhere, where they can be picked up on the way through, or right after jumping. Without a few of them, a map will feel more tedious, as you have to get back to your medic or a medium healthkit more frequently. They won't decide gameplay too much and can add a nice touch to an area. Use them . Good medium health placements Medium healthkits have a lot more impact than their small counterpart, so they require more caution in placement. There shouldn't be a lot of them in a map, but those that are on the map should be in central locations that are easy to get to. At the same time, the health should not be in a location where players would stay for long. Contrary to the small ones, a position where a player will not run through it on his path is better here, since that prevents players from accidently picking it up when their medic needs it. 1.5. Cap times Cap times can be a huge aspect in wether a push can be succesful or not. Furthermore, the cap time dictates how much the point effects the combat in the area. A shorter cap time gives the point more imporance during a fight, since the defending team will have to act much faster, when the attackers start the cap. A really short cap time, like Badlands last, will keep the defenders on their toes, since even at x1 it will cap in 2 seconds. That means players have to be careful to not let any entrance unguarded, or keep a player (or stickies) in a tactically unfavorable position to block the cap. On Badlands simply threatening to run on the point can force the enemies to jump down, which then leaves the rest of the attacking team free to take the highground. A point that caps much slower, like Gravel pit A and B (60 seconds at x1), will be much less important during the actual fight. As a result of that, the combat on the Gravel pit points usualy focusses more on the geometry around the point, for example the roof on B. A huge factor on 5cp maps are backcaps. The faster a cap point caps, the more worried a team has to be about players running around or hiding somewhere when trying to push. If the cap times are roughly equal, the amount of routes and hiding spots available are a larger factor than the cap time here though. It's still worth thinking about imo. Cap times are a part of the map that has to be well chosen and then fine tuned in gameplay. They of course have to fit the geometry and type (last/mid/etc.) of the point. Best advice I can give you is to take a look at the cap times of the popular maps and decide what will fit your map. 1.6. Spawn times Last (and least to an extent) are spawn times. I would not recommend fiddling a lot with the spawn times. That doesn't mean you should absolutely not change the spawn times, but you have to be sure of what you want to achieve with it. First of all, the effects of minor changes (+-1 second) are hard to notice in actual gameplay and larger changes can really unbalance the game. So what does changing the spawn times actualy do? If you reduce the time, it means there is less of a penalty for a death. In theory, this rewards aggression, since the punishment of losing a player is not as harsh. But what it really does is punish the attacking team, since they have the longer reinforcement route. If you have ever played a 32 player instant spawn server and tried to break a defense, you know what I am talking about. Giving a team that owns 4 cps a spawn time advantage in a 5cp map is fine, since the team defending last has a much much shorter reinforcement route. In attack/defense and koth maps, playing with the spawn timers a bit can be worth it though. For koth, the attacking team often gets the timer advantage, in order to help comebacks and make the matches more even. From my hardcore competitive point of view that is bullshit. If a team is outclassed, they should just get rolled and not get helped by the map. But that overlooks the defensive strength of the TF2 classes, especially the demoma. On attack/defense maps, given the nature of the stopwatch mode, attackers should win most of the time. Giving the attackers spawn advantages can turn the map into the attackers favor, but in my eyes this should be done by other means. In the end it is fine to give the attackers on a/d and koth a spawn time advantage, I personally would try to find a way that is not spawn times. 2. Classes After taking a look of what affects the pace of a map, let's take a look at what a map needs to cater to the individual classes. I will focus on the classes mainly used in competitive play of course. Engineers are rarely used, so the map does not need to do much against them. At the most, try to eliminate very strong sentry spots on last points. Catering to spies, have a lot of ammo and places to sneak through/hide. Pyros, if used, are mainly just used for airblast. Not much of a concern, but especialy annoying in tight hallways. In the ideal case, a map will give every class good opportunities, but it is not always possible to plan them. I will give examples for that in the individual explanations. 2.1. Scout Scouts are played for their mobility and their deathmatch potential. I will treat those as two different aspects that a map can cater to. To let the scout make use of their mobility, large travel distances and a lot of different routes will help them. Having too many routes on your map will make it too easy for scouts to get around to get picks or start backcaps. Having too many will make the same process too hard for scouts, and scouts will be less useful. Making shortcuts specifically for scouts can also be a good idea. Those jumps can be either on purpose (see the dropdown on badlands, I assume it's on purpose), or by accident. If you take a look at this jump on granary middle. It is highly unlikely that Valve actualy implemented that jump for Scouts to use, so it is an example for something nice by accident. It is tricky to do, but can have large pay-offs, so it is a nice addition to the map. Another example for a good jump is this jump on Badlands spire. The deathmatch potential a scout has as a very fast class with small hitbox, double jump dodging and a good shotgun can be catered to by the map as well. A large area with a lot of smaller obstructions is a joy for a scout to play on. For example the middle point of coldfront is quite a good area for scouts to deathmatch, just jumping on and around the little rocks and fences a scout can be quite hard to hit. The opposite of that, tight corridors can be quite hard on a scout, since the mobility is of less use and soldiers will have a way easier time hitting the scout with splash damage. Flat open areas aren't that good for scouts either, since the scout will be easier to hit and easier to spot. 2.2. Soldier Soldier is the class that makes the most out of a height advantage. While height advantages are not the be all and end all (even on spire ), solly gameplay will often revolve around gaining and using height advantages. Badlands has great examples of different heights to use. Smaller ones, like the trains on mid (or even smaller ones), are great for helping to control the area. Bigger height differences like on the last point will make the soldier dominate everything that is below him. At a certain point, like the height that spire gives, the fall-off damage (and dodgeability of the slow rockets) will decrease the damage potential a solly shooting rockets down, actualy weakening the advantage of the height a bit. With the splash of rockets and his damage potential up close, a soldier can do quite well in tight hallways. Defensively a soldier can get a lot out of spammable chokes. The main thing to make a map interesting for a soldier is vertical space though. The on demand mobility of rocket jumping is a key component to the class and should be useable in the map. Spire is pretty obvious, but Granary Middle or Gravel pit B and C all make for good examples of this. Something that is really annoying for a Soldier is small geometry that will prevent splash damage from the rockets. These can be very small ledges that rockets explode on and don't do splash to nearby players (the rails on Badlands mid for example). They can also be holes in the ground like the ones on Badlands middle or on spire. To fix these, you can use bullet clips. Doing so can create counter intuitive effects of where rockets explode, but the rockets that randomly do 0 damage to players right next to them are way worse. 2.3. Demoman With demoman, as with a medic as well, you don't really need to worry about the viability, since the class is too strong to not run it. One thing to consider is the availabilty of sticky traps. Sticky traps are mostly used at chokepoints, and in hallways, because the player traffic is predictable there. The locations for sticky traps in the Valve maps mostly happen from detailing and therefore probably by accident. Let's take a look at some popular traps on badlands: There is a sticky trap with 5 stickies hidden on every picture. Go figure. Notice how in the last two, the stickes are almost invisible. The other traps are either invisible from different angles, or can be hidden more carefully, making them invisible as well. While props are generally unclipped in order to ease the movement through the map, you have to be aware that unclipped props are potential places for hiding stickies. With this in mind, it can on the one hand be used on purpose, in order to give a demoman another option. It can also be avoided, by having the props collisions turned on, and using a playerclip in order to increase flow of the map. Something else a demoman can make use of are very long jumps (for example onto spire from badlands choke). Those require that there is enough space to jump though. Chokes without a roof help this cause, but other than that, planning for the demoman to do long jumps shouldn't be much of a concern. 2.4. Medic/Heavy/Sniper (or the rest) Medics are kind of hard to specifically cater to, since their functionality is quite independent from the map. General principles still apply of course, like the pace of the map. As for the heavy, while you could make different points about different geometry in the map favoring him, the main factor for his viability are walking distances. On most maps, the poor guy is just too slow to be able to keep up with the pace of a 6vs6 game. As for 5cp maps, gullywash stands out here, since the walking distances are short and the amount of flanking paths are quite low, which makes running a heavy on gullywash that much more viable. Snipers benefit from large sightlines and few flanks. Making snipers too viable on a map can ruin it for the players that are not the sniper. For example Viaducts main problem is that snipers are really strong. As for 5CP maps, Coldfront, Obscure and Freight are examples of maps that are quite strong for snipers. Large sightlines are ok, if there are alternate routes and there is cover from the sniper. Having a map that is entirely bad for sniper is a bad idea as well, since that map will be so constricted, the other classes won't have much fun on it either. 3. Other concepts In this section I will try to cover other important concepts, that I have not or only slightly covered yet. 3.1. Rollout/Midfight The rollout at the start of the round and the resulting fight for the middle point mainly concern 5 cp and koth maps. It applies to a lesser extent to ctf maps as well, and it can be possible to apply some concepts to attack/defend maps. The key class in the rollout is the demoman, followed by the soldiers. The demoman will, on most maps, have the hardest jumps and get to the middle point the fastest. The soldiers have to do their jumps too, but they are, apart from gunboats rollouts, usualy easier and a bit less important to do perfectly. A good rollout should be very hard to do consistent and perfectly. It should hower not have a too harsh punishment for missing a jump. On Badlands, if a demoman misses the jump on the stairs, which can happen quite often, he can jump to the balcony on house instead. You can't do that much to really plan the rollout. While it should be a consideration, the dynamic between the control points and the design of the control point areas is much more important than the rollout. Plus if you design the map with a specific rollout in mind, I guarantee you that players will prove you wrong and do something different from what you planned . So don't break your head over this, but keep it in mind for adapting your map. 3.2. Hiding spots Hiding spots are a bit of a double edged sword. On the one hand, they add something to learn for the map. They can help players be (or feel) more creative, and are a mean of adding specific map skill. On the other hand, checking a whole lot of hiding spots can be tedious for a player, just for the small chance that somewhere might be some idiot hiding, hoping that his spot doesn't get checked. There are three different types of hiding spots I would like to distinguish. First of all let's take the simplest one, corners. Corners can be an odd choice as a hiding spot, but if players don't expect anyone hiding there, they will probably not check. Here are a few examples: Some typical corner hiding spots A solly 'hiding' on the last one might not seem like the cleverest idea, but when players are focussed on the exit of that room towards spire, they will not see the solly, who can then easily kill them. Even if spotted the soldier can still deal damage there of course. You can find many of those spots, and they often rely on the enemies being focussed on someone else, or just not checking, because of laziness or because they are in a hurry. I wouldn't recommend having spots like that everywhere, which will make the map less clean cut and annoying to play, but you should have a decent amount for players to be able to get creative. Another type of hiding spots are high up spots, where a player will sit on a small ledge, which are mainly used by soldiers, who can rocket jump up there. These are harder to check, and can involve tricky jumps to get there. Let's take a look at a few examples as well: A handful of selected high up hiding spots. Soldiers mostly use these spots when their team is in a bad position. If most of the team is down, the hiding spot can be used to kill the enemy medic. The same thing applies when the enemy team has uber, a soldier hiding can be a good way of forcing it, while the rest of the team stays safe. Having too many of these hiding spots will be fun for soldiers who like to play sneakyly, but annoying for everyone else. Those are a good addition to a soldiers arsenal though. I wouldn't recommend planning with them too much, you will have some eventually and just have to decide wether they stay in or not. If you have eliminated them all, you might want to consider clipping a prop again here and there, or removing a player clip, to allow for some of these. The last type of hiding spot are just hard to see spots. These can be unclipped props players can (partially) hide in, or just dark areas. And again, examples: Where is waldo the scout? There are far less of these spots, because Team Fortress 2 is not a stealth based game (spies have invisibility, they don't need shadows ). They work a bit like the other spots, except that they rely on players not looking closely enough. 3.3. Learning curve The general game design concept of easy to learn, hard to master, applies to maps as well. The layout of a map should not be overly confusing with a lot of twists and turns. Players should be able to find their way around pretty easily and to identify the objective quickly. While this is especially true for public play, competitive play is no different here. If you want people to play your map, it can't be too complicated to understand, or they won't bother. The hard to master part is difficult to tackle, and way more important for competitive play than public play. What is quite possible to do is to make sure that there are a good amount of easy and harder jumps for different classes to do. I covered some of them in previous chapters. Another example for a jump adding something to the learning curve, while not being a barrier on learning the map (can easily play it well without the jump) is this example from gravelpit. Some other examples include Turbine, Viaduct (sliding up the rocks on middle), Badlands last (switching sides with the ledge on the control point) and Granary last (jumping onto the highpipe from the rooms above 2nd). Another thing you have an influence on are the different types of hiding spots and sticky traps. Of course there will always be some that you haven't planned and the players will find, but you will have the decision wether to remove those or let them stay in. Hiding spots can be an annoying part of learning a map though, since players will learn about them after getting killed by someone hiding there. Regarding the viability of different routes, places to hold positions from, methods of attacking, viability of different classes and the number of tactical choices, there is only so much you can predict, even if you have experience at a high level. Most of it can only be really seen in actually playing the map. 4. Conclusion I've talked about the different qualities a map for competitive play should have, and tried to give suggestions of what to do regarding these factors. On some of the factors, there is only so much you can do without a lot of playtesting, while others can be seen and adjusted. A couple of qualities are needed. -fun diverse gameplay -balance between the classes -easy to learn layout with enough depth for tactical variety -a proper learning curve, with harder to execute jumps, possibilities in rollouts, etc. -good optimization, framedrops are really annoying when trying to play the best possible game Not all competitive players think alike of course. While some like me don't care about the looks of a map at all just the actual gameplay (you can jump on it, or it blocks that sightline is enough of a reason for me personally to put something somewhere). Many will care for detail and immersion in a map, so those are important factors as well. In the end it comes down to testing and adjusting the map accordingly a lot, since what can be done by theory fortressing is pretty little actualy. Getting the map playtested can be an annoyance as well. First thing you can do is post the map on the major forums of the competitive scene (etf2l for europe for example ). You can try to get it into pickup channels, and ask for help in them. If you know a lot of people you can try to get mixes going. One last tip I can give you is to be selective with the criticism your map gets. Try to find out why someone said what he said and then think about a solution.