The wicked scales of the beast

  • If you're asking a question make sure to set the thread type to be a question!

Deodorant

L6: Sharp Member
Oct 31, 2011
263
214
In my many unsuccessful attempts thus far to create a moderately enjoyable map, I think one of if not the biggest problem has been getting the scale right. Basically everything I've made has been criticized for feeling either too small or too large, and in one case both simultaneously. Whenever I map I try to actively think about and question my intuitive understanding of scale, and to compare what I make to things from official maps, both by taking actual measurements in Hammer and by just walking around in-game and seeing what 'feels' right. Whenever I submit a new map or version of a map it feels like I've probably finally gotten it right this time, and yet when it's deemed overscaled or underscaled it always feels really obvious in hindsight.

I figure it's about time to ask for adult supervision.

How do you guys think about scale? Did your intuitive understanding of how big to make things come naturally over time, did you consciously work on it or did you have it from the start? Are there any good methods or rules of thumbs to getting it right? If you have any thoughts about scale whatsoever, blurt them out and I'll lap it up with unconditional awe and obedience.
 

Fruity Snacks

Creator of blackholes & memes. Destroyer of forums
aa
Sep 5, 2010
6,394
5,571
It you think it's too big, it probably is.

The one thing I use, and highly recommend for people to do, is use a prop_static of the heavy.mdl. It's the "largest" char model so if looks big with him standing around, it probably is.

A lot of other stuff just kind of came up through design.

  • You want doorways to be wide enough for a player to walk through on one side, and a spy to sneak through undetected (So, the standard doorframes are excellent).
  • Hallways need to be wide enough so that you can easily jump around, but not so wide that you can completely avoid confrontation.
  • Ramps shouldn't be really steep. If it's a steep ramp, it should be stairs.
 

Idolon

they/them
aa
Feb 7, 2008
2,105
6,106
One thing that I've found is that certain layouts simply cannot be scaled correctly, no matter what. This is because scale has to work on a micro and macro sense. Even if all of the different elements of your layout (rooms, hallways, stairs, ledges) are scaled to something thats pretty comfortable, the overall size of the map might just flat out not work. The best way to fix this is to see what other elements you could insert between to get scale working on a macro level.

This also works vice-versa, you could have decently scaled caps/spawnrooms, but just not enough layout stuff happening to fill in the space between (this is when things get obviously overscaled).
 

kinggambit

L1: Registered
Oct 13, 2013
17
2
What's your workflow like? Planning ahead is the best way to combat disproportionate scaling from my experience (usually sketching out some concepts and an overview layout and block out phase).

Blocking out the map in hammer is probably the more influential factor in establishing scale.

I block out very generally (a building would just be one big primitive brush). brushing in windows and doors isn't really helping unless you want to check certain positions (i.e. if there's a window area for snipers to be peak from).

Then I scatter a couple of player_info entities every now and then in areas you expect players to be. I assess the scale constantly using the player entities.

I would block the entire map out using those two methods.
 
Last edited:

mint onion

L3: Member
Jul 18, 2013
101
45
i struggle with this as well. i have a developing understanding.

i currently associate "scaling" with two things. the first being the visual relationships between props vs. props and props vs. player. how big is a doorway compared to a window? how big is a doorway compared to the player? this is difficult to pin-down in a cartoon universe.

creating and enforcing spacial rules to play spaces will create a sense of unity and structural language. ensuring your walls are always the same thickness, ceilings are the same height, steps are the same height, etc. in my mind, consistency creates a universal balance.

the second association, is basically feng shui or level flow. this, in my opinion, is at the core of everything. the relationship between player vs. game, or the game's inherent mechanical behaviors. in my opinion, the execution of this relationship, by creating the space that houses them, determines the true value and success of a level. a level's goal is to fulfill the game.

under normal conditions for example, a player is a certain height, can move a certain speed, can jump a certain height. a player will take fall damage when they drop a certain distance, weapon fire becomes less effective at certain ranges, and splash damage has specific radii. these mechanical volumes and thresholds define what appropriate playspaces are, which then define the appropriate use of structural language, which then defines the size of a doorframe.

all of this is just talk and thought and i have yet to apply and polish these ideas but that's my take.
 
Mar 20, 2012
391
806
I generally use Fr0Z3n's method, although I like to use a T3 sentry/sentry nest for my scaling.

However, it's not fool proof and the problem is that scale really depends on how much action that particular area receives. The more action, the more you need to compensate for mobility, positioning and team strategy. Too small and you create a choke that, combined with other chokes, causes the game to halt. Too large, and you either create an impossible to defend position or a sniper's haven.

So, it's pretty rare to hit scale right off the bat. Often my major scale changes come after the first playthrough when I review my notes, individually look at each route and ask: "Is this a viable route to push through? Who is using this route? What are my options if I want to get from A to B?"

Ultimately, like most of the issues within mapping for a game like TF2, the consequence of improper scaling is that you add on a new layer of imbalances to class relationships.
 
Apr 14, 2013
663
343
I usually tend to make things underscaled, and I think the easiest way to deal with it is to use visgroups, and then if something is too small you can use the visgroups to move the area that is scaled correctly. then, you can build more things around it or increase the scale of that area's surroundings. I guess this could work for overscaling too. The dev-engi's model from the Ultimate Resource Pack is also useful.
 

Deodorant

L6: Sharp Member
Oct 31, 2011
263
214
Come to think of it, I feel like there's a discrepancy between how big things are experienced in first person and how big they are compared to third person playermodels. 64-unit heights, for example, reach to the chests/necks of most classes, but I unconsciously think of them as around waist height. If I open an official map and put playermodel props in it and start thinking about it, everything feels way bigger than it does in-game.

Is it just me?