Archive for May, 2010

The Unoffical Rulebook to TF2 Mapping.

May 29, 2010 Leave a comment

Recently Ive been frequenting a custom maps server. Being on the server really shows me how much people who could really make some awesome maps really get bogged down with some stupid little thing that should have never been in the map, or some issue with the map that they should have caught immediately. So I decided to Put forward my “unoffical rules of TF2 mapping”  What this is is a list of rules that I have observed when looking through valve maps as well as things when making custom maps. These are things from simple layout mistakes that cause stalemates or easy caps to literal no no that may look fine but can cause players real bad problems.

First off the #1 rule, This isn’t from valve but its from me. walk your maps. Before even letting a friend see them start up tf2 and run your map. and explore. Run up every staircase, through every door, cap every point(for both teams). Do everything look everywhere, while doing this keep a notebook handy to jot down any problems you see or improvements you could do. A few minutes of wandering can save you big time further down. Last week as I was working on King of the Train 2 I found 3 items just from wandering throughout the map. That’s about 4-5hours in total just wondering around and testing everything.

keep Objectives and spawns separate
This sounds simple but many people really make this mistake. Never Put an Active Objective near an active spawn. This means If its ctf, keep the intel room and the spawn separate. If cp Never put the currently contested control points right outside of spawn. ones that would be locked when the spawn is active are fine however make sure to keep the resupply cabinet away from the doorway. Just remember, you may be playing your map on just plain tf2 but There are servers with 100% crits, moon grav, instant respawn, crazy weapons, ect. And If the control point is right outside of spawn an attacking team could be facing an endless stream of enemies.

Never have only one route!
Go to any valve map and find a spot that the only way past is through this one spot. This is because tf2 can turn into stalemate really fast. so everyone going to the same place means there’s everyone on one team has to die for one to push forward. This is not even putting into account the good old engineer. A smart engineer can turn a poorly designed map into a nearly guaranteed win for his team.

Skybox means high room.
once again a simple concept that many have trouble following. If the skybox is above the player, make sure they have plenty of room to sticky/rocket jump. Enough said.

A good mapper Can Make models for his maps. A great one doesn’t need to.
one thing that will always sell your map is the download size. If your map looks great, is fun and balanced, and small to download then you have no where to go but up. download rate is key when making custom maps. So how do you keep the size down? recycle! first off use the valve models as much as you can. they don’t have the book shelf you wanted? make it a cabinet instead. Secondly Is using func_detail models. Still need that bookcase? Make a texture with a bunch of books on it and model one in hammer. (Interesting fact. In red spawn on 3rd round of dustbowl there’s a film projector seen inside an eye candy room. However if you try to find that projector in the model viewer it doesnt exist! The projector is in fact a func_detail model, however it looks good enough to be a full detailed prop!) finally recycle the textures, If you see a texture you want but you cant use it for some reason(its a hud texture, or a model skin) then make a material that makes it a texture instead! You cut down on size and time. (In my pad map you can see 20 foot tall bonk drinks. see the third picture These are really just cylindrical brushes with a material that uses the bonk drink’s model texture.

Smooth out the edges
One of the biggest annoyances on a map is when a player is backpedaling trying desperately to say just out of a pyro’s flamethrower when suddenly a small random object stuck to a wall causing him to instantly stop just enough for the pyro to overtake and kill the player. This problem can happen anywhere. Therefore you must make sure the player doesn’t stick anywhere. Now before you start breaking out the player clip colored shrink wrap, This doesn’t mean covering your map in player clips. rather just thinking like the player. If there’s a small conduit that leads up to a speaker high overhead. then I shouldn’t run into it. Similarity a unspoken game rule pops in here. small corners don’t exist in games. any high traffic location that has a small outcropping should be playercliped so if a player is using a wall as a guide he curves around the outcropping instead of just stopping at the corner. a good rule of thumb is if the player model is  wider than the outcropping then it doesn’t need to to be clipped.

Expect Idiots.
There out there just waiting for you to let your guard down so they can ruin your map. Every time valve releases a new map slews of people go to work trying to find every little spot that valve forgot to player clip or made incorrectly, or created a stuck area. then they use those to grief and mess around with. Remember this? Always assume there going to exploit something. use thick playerclip fields and make sure if it looks exploit able it isn’t. Last thing you need on your map is another gravelpit.

beware of closed areaportals.
To be honest I don’t use closed area-portals at all. there far to dangerous for their worth. And valve seems to agree with tf2. I have yet to find any closed areaportals on any valve tf2 map. closed areaportals are for single player when its only one monkey in the cage. in multiplayer you got 24-32. Much harder orchestrate properly opening and closing areaportals.  this does not mean you shouldn’t use open areaportals to help viss learn what to render. just don’t close them.

Keep it simple stupid
OK your making you super awesome sweet map that has a Super awesome super complicated game mode that…STOP! Kiss method. This is a saying used in the Film, gaming, web, and any other multimedia industry. When working on something try to keep it simple. This applies to tf2 as well. Don’t make things too complicated. you’ll cause too much work for yourself and you’ll confuse the player.  This doesn’t mean not to make something look detailed. just not down to the inch. Now some may be thinking “but valve does this complicated intricate and highly detailed maps, why cant I?” First They got-cha… second they dont. Take a close look at a valve map next time, there in fact very simple in design. they often use textures, props, and overlays to break up the simple geometry and make it look more complicated than it really is.

the necessities of tf2
Tf2 has a habit of things. Things that are in every valve map guaranteed. such things as in a respawn room there is a func_respawn, and a resupply cabinet. This means in every respawn there SHOULD be a func_respawn and a resupply cabinet. Similarity every spawn door has a blocker to prevent enemies from getting through.  Secondly If there’s a cap point, there’s always an hazard outline that dictates the range of the cap point. Similarly if you break these rules players will assume they are still in place and this can cause problems.

two houses of mapping.
there are two ways you can make your map. These are standard maps and fun map. Standard maps are maps that are designed according the development scheme of tf2. Fun maps throw the scheme to the wind and wing it. Fun maps can ignore or bend rules without causing the player much strain. So how do you tell the difference between a fun or a Standard map? the basic difference usually comes from how you want players to play the map. For instance a currently popular map Koth_lolcano. The map is full of big LOL’s OMFG, ROFL, and wonderful Russian singing(he he he). However for all of its craziness Its still a Standard map! Even Lava coming up and killing everyone on the control point doesn’t change it. the idea is that people are still intended to capture the kill and hold it for the allowed time.  Unless this changes the map stays a standard map. King of the train? that’s a fun map. Its a single cp Push map that is on a train that crashes into the enemy base upon capture. When making a map think about where you want the players to go and what you want them to do.

two Worlds of mapping.
It is true how to map looks makes up for about 70% for when players like your maps. However the other 30% is The world that you cant see. the one the engine uses to know where the player should go and what each item should do. In order to make a good map you need to make sure both worlds are working perfectly together.

Beware of being shiny!
As I said 70% of people judge your maps on looks alone. However, this doesn’t mean you should just a map that looks good. A map that looks good will have a good run but it will be soon forgotten when game play problems become apparent. Once everyone knows how to break the map your “shiny map” gets stuffed in the trash on nearly every server. Think Hoodoo. I have always said the only reason that map was ever selected was to show off valve’s particle editor with the particle effect at the end.

the gamer is always right
If you are playing your map and someone complains about something. FIX IT! yes, Many times they complain for stupid things that would break the game and those comments should be left where they stand. However If its something that issnt a major game change or something that is simmilar to your own opinions then change the map. remember you make maps for everyone else just as much as you make them for yourself.

ques save time and headache.
Giving the player a hint where to go will always make your maps run smoothly, not to mention reducing the problem of people wondering where they shouldn’t be at the time. There are visual ques(most common) audio ques, and text ques(when visual and audio ones wont work.) Visual ques are signs, lights, or objects that make players have better view of ways that they should go. (go watch the valve commentary if you dont know what ques are.) Simmilary you should have all player spawns pointed towards the exit to spawn.

beware of linkage!
the theme of Team fortress 2 should always take prescience in any map you make for the game. Ceritan models, sounds, and events are linked in a player’s mind with important game play elements. For instance the win music played when a team wins. When a player hears it they may assume they won.  although a quick look at the hud shows otherwise they may become irradiated at the deception. when placing something always think “will anyone be confused because of this?”

Give the classes what they need.
You cant predict 100% where everyone will go and what may result in your map. But you should at  least give the player some place to start. These are what I call zones, spots that a specific class will excel. The most important zones are Engineer, Sniper, scout and soldier/demo most other classes dont have any “spots” they would exell or are less dependant on geometry. Engineer’s need a spot where their sentry can cover every direction. This doesn’t mean that its invincible rather just no class can just get to a spot out of combat and hit the sentry from range. Snipers need long corridors and spots that take them out of the main combat keep them safe from spies. Scouts need spots they can double jump onto giving them the ability to quickly navigate a map. and Soldier/demos need spots where they can get up high and rain down death.

When in doubt ask valve
yeah I doubt you have a direct line to gabe.. however you have the next best thing. Their maps vmf’s . Recent updates in the sdk gave tf2 mappers the source files for several tf2 maps, including dustbowl, lumberyard, badwater, goldrush, and many more. If you ever wonder how you should make something think if its been done in a valve map then see how they did it. Similarity if its in a custom or a valve map that the source wasn’t released in. there are programs online that can decompile maps so you can see what they did as well just remember decompiled maps are not 100% copies of the source files. There are bugs in them. NEVER recompile a decompiled map.

watch the commentaries
Like i said further up the commentaries for valve games really have some gems in them that give you a sense of how to map for game-play rather just to map. Watch all the commentaries for all the valve games. Even if its on a different game many things can be still done in tf2.

That’s all the rules for now. I’ll try to update this as and try to make this a really good guide on what to and not to do in tf2 mapping… If you have any questions feel free to comment!

Skeletons in the closet.

May 13, 2010 Leave a comment

Yes im working on Kott2 I’m busy hammering down the last few problems and getting everything looking somewhat presentable for the first balance testing release of my map.

Ive always tried to be creative in my maps, and as many of my testers know Ill try to push the limits of my mapping. However pushing the limits can often cause them to break, very badly… And such i present to you my failures.. these are maps that never quite made it to release, Many of them are just dev textures, while others are fully textured proped and rigged, however every map Has something wrong with it that prevented it to be released..


This was before when I was deep into Neotokyo mapping and wanted to make the traditional Cherry blossoms blowing in the wind thing. However due to the fact Neotokyo was modded in the old version of source It didn’t have access to custom particles and thus my efforts where for naught. That’s when I decided to make a Tf2 map. (one above) I didnt want to really map that big of a map. Mostly since I wanted it to be quality over size. Due to the fact King of the Hill was not implemented into tf2 yet i was forced to make it a Arena map(a game mode I have always despised).  This eventually wore down my initiative to make the map and stalled it before I could do any more than get a rough layout… I may still restart this map later but for now its on cold storage..


CF_docks or Control Fuel docks, was a custom game mode and map I designed. The map is fully textured and rigged. The game mode was this: 3 control points  and 2 fuel depots, one next to each teams spawn. When a team captured a point the fuel tank would fill, More points faster filling. once full the team wins(yes its domination).  The map itself was deigned to be a “muli-map” and idea I spawned from a previous game I played. There are CTF, CP, CF, and arena versions of Docks. my idea was to release them as a map pack. same map played 4 different ways. Balancing was a nightmare and I had to start making special objects and sets for each game mode, Something that drew away from the original concept. This eventual put the map on cold storage.. Due to its ability to be used for nearly every symmetrical map game mode It may just see the light as some other game mode.


The concept of this map came from two things, 1)was the want to make a canyon map(both teams fighting across a canyon.  2) was a part of the meet the spy video (this scene) Where they have an underground base from the cave entrance. And so I began work on Canyon. adding a back door on the upper level similar to the one in the scene. This coupled with a large defend-able  entrance. The problem came inside the base that went through iteration after iteration. each being either too complex, to hard/easy to defend. or very hard to model. Eventually I gave up and put the map down to cold storage..


Ahh, infamous bayfeild. This is the first map I ever made. The maps balance is great, game play is great, everything is great, the map is fully ready to compile… If it would compile… for some reason this map has been unable to compile after the beta version. I know if i go back to it now i could probably find the problem. However The map has lost its “theme” I’m my opinion and Ive move onto other things. However this map has in a way continued on.. Next time your in King of the Train take a look in to the skybox. this map is in it!


Do a barrel roll! no really! this was an interesting idea i had. a slew a barrels falls from the top of a hill the objective of the teams is to get up the mountain and cap the point at the top. hit a barrel and YOU DIE! The map started with problems. the barrels would get caught and start stacking up and causing massive lag. Even after I added a timeout to the barrel the barrels still lagged the server considerably. Secondly once a player reached the top It was nearly impossible to get him off due to the fact that a player would have to fight him and the barrels at the same time. Finally i wanted to make special barrel types, Exploding barrels, Iron unbreakable barrels, barrels that break with health inside, and more. However this needed me to model special barrels. Something I didn’t know how to do at the time. So the map eventually was thrown into the freezer.


I SOOO want this map to work correctly.. However this is my classic case of pushing the envelope too far. everything in the map is either destroyable or moves with damage! the idea was to have 4 warehouse buildings with multiple barriers blocking each teams, The teams would have to destroy walls and barriers to reach the enemy team! even more there are 4 pillars inside the buildings that if destroyed take the entire building down on everyone inside!! Problem is 1, crush damage doesn’t work in tf2. so I would have to lace a kill field inside every object if i wanted to make the crashing buildings work correctly. Also, when rendered with shadows, the entire building gets a checkerboard pattern from the separated breakable objects.  surface to say two serious problems with the map that would need to be fixed by a valve patch if this map would see daylight.


streak (like Winning streak) was designed to be a fast paced map with teams battling between two close bases. whats more ever so often a train would go right through the map forcing team to use the lower or upper routes. as you may notice the map is fully textured, rigged, and lighted. However, there where two problems. one was func_train hates turning upwards/downwards. It would sharply turn the train cars  that made the train look very odd after it left the map. Secondly was MASSIVE game play hole. once a scout had the intel and was out the side door he could easily get to his base by jumping over the gorge in the picture. and with the layout of the map  no one who saw the scout grab it could even get out of their base before the scout was back on his side and running into his base for the cap. I wanted to move the intel into a basement however This left even more problems and drew attention to the lower floors where I didn’t want it.


This is the 3rd map i did, and a attempt at a custom game mode I found.  The game mode is Siege. Red would attack the blue base, First busting down the front door, then destroying 3 “relay stations” then finally the core. After the core is destroyed the base explodes. The map had a solid design.. If it was about 50% smaller. being that this was a early map I still didn’t have proportion correct and several rooms as WAYY to big or small. So I would have to totally recreate the map if i wanted to go any further with it. One thing i did take off of it was my “hidden sentry” prefab, I don’t have a place for it yet… but ill find one someday…


Its tug of war, basically both teams push the cart to the enemy base, Once the cart reaches it it explodes destroying a room in front of the spawn. The map never was finished due to game play problems. basically the game was either a slaughter or a stalemate.


This was when I thought “if well has 2 trains” I wonder how fun it would be if there where more trains?  this map has 10 trains! each going through the map at various spots, many times spots where combat is likely to happen! the problem of the map was length. It was easy to die and took a while to get back to the battle. It eventually lost interest. This map still lives on in spirit, many pieces of the map are lifted for newer maps. Most recently I used the Lighting that was in a part of this map for a eye-candy room in King of the Train 2.


There is nothing wrong with this map, no seriously there isn’t! The map started from a piece of valve concept art for lumberyard(shown here). The map was going to be 6point 3 round cp map similar to dust bowl. The problem was simple…time. This map came shortly after release of Summit. and I realized making large maps is Tiring. I sill haven’t fully recovered form the effort from summit. Eventually I may release this.. but not for a while….

Categories: Hammer, Team Fortress 2
%d bloggers like this: