How do I generate ideas for unique layouts in CSGO Wingman maps?

[fishingbrick]

Now, I know that this may seem specific, but I'm genuinely having a difficult time making a top-down layout to get me started with my first mapping project. The best that I can come up is 3 lines like Rialto, and I'm seeing unique layouts all over the place in the workshop. Is there a formula to it or something? I'm just completely stuck on how to come up with a layout.

[NikiOo]

Well, it's actually quite simple.

Ideally, you want to minimize the filler areas and maximize areas of type alpha and omega. You do this by learning how to think in 3 dimensions and how to generate interesting areas, while working within the constraints of nature and human architecture. Here is the formula for what makes a good area:

Goodness of narrative comes down to story-telling, which is a very expansive topic in and of itself so I will focus on goodness of color and goodness of shape. So we'll use a simplified formula for goodness of area, shown below:

Formula (1) is the lego approach to area engineering, where you deal with large presets, since you're not yet experienced enough to work on area microarchitecture. Now I will show you the wizard approach. You won't read about this in LD text-books.

Let H be the set of architectural components used by humans and let N denote the set of architectural components used by nature (God / evolution). Then we'll use Z to denote the available language of architectural expression. Z is the proper set of the union of H and N. Naturally, you cannot use the entire language of expression at your disposal so let's say the language you're going to use is some subset of Z, which we shall call L.

Now in order to learn how to make a good area, you need a way to calculate the goodness quantifier of formula (5). Without that you won't have the much needed feedback-response cycle and you will not improve. So we need to measure koppaI from an already built level. To do this, you take all elements of L and distribute them in a 3-d vector space, based on their position of use in the level. If you want, you can use two separate vector spaces - one for the color subcomponent of every component, and one for its shape. But beware, the natural distribution of the reasonable color spectrum for most components is pretty tight.

Next thing you do, you color code the scattered points based on the spatial locality of the components in the real world. In other words, components that appear together are assigned similar colors. Then you want to reduce the amount of overall entropy in the color distribution. One way to do this, without sacrificing variety is to have clusters of low entropy (which appear as splatters of uniform color) in distinct areas of your map, while the connecting areas should be kept low entropy (without many micro-clusters) so they can compensate for the contrasting transition between the larger areas.

For reference, in the image above, a high entropy area would look something like (c), while the rest are relatively low entropy.

Some easy ways to reduce color entropy in your level are:

• using a very limited color palette
• using painterly textures
• using photo textures taken from the same real world location

Some easy ways to reduce shape entropy in your level are:

• using references
• not mixing architectural styles from different historical periods
• use large pieces for cover, remove small pieces that only add clutter
• reducing signs of human life (it is very difficult to prevent signs of human life from skyrocketing your entropy meter)
• hierarchical procedure (starting from a silhouette blockout and gradually filling in details)

Keep in mind, entropy can also be too low. This can happen if you put too much effort into silhouette and visual harmony and don't use any references. You can end up compensating with bad entropy. Bad entropy is when details you add increase visual disharmony very rapidly. Ideally, you want your learning process to lead you to a place where your entropy is similar to the entropy in areas you find visually appealing in real life or virtual environments, and you also want the level of detail to approach the level of detail in those environments.

In formula (6) I show an example optimization model for a realistic style area. You can use this to derive a model for fantastic or stylized areas. You want entropy and detail to move in parallel with one another. The ideal scenario would look something like this:

Keep in mind that adding narrative elements such as signs of human life and natural decay can increase entropy quite alot. For further reading on sings of human life look into How buildings learn by Stewart Brand.

If you get good at adding detail without massively increasing entropy, you will have more confidence to experiment with unique layouts.

Then there is another formula, called the challenge complexity formula. Player enjoyment from playing a map tends to correlate inversely with frustration. Frustration comes from low correlation between expected challenge and actual challenge.

One way to model challenge is as a scalar, representing net difficulty (7).

Another way you can model challenge, shown above, is as a set of micro-games that the player will have to play during the match, determined by the context of each encounter. The phrase "micro-games played in wingman maps on official servers" can be replaced with micro-games anticipated by the user or micro-games user enjoys to play in counter-strike. For this analysis, I assume that enjoyment is maximal on wingman maps on official servers, presuming those maps are examples of prime level design. This is more true for defuse maps than it is for wingman.

Micro-games constitute things like how you peek certain corners, how you play around with cover and position in certain contexts, how you use your utility, and all sorts of other actions you take to become more unpredictable and to put yourself in a favourable situation. The possible situations you may find yourself in are determined by the architecture of the level.

You don't want to introduce too many unique ideas into your level, because it will reflect badly on the micro-games ratio coefficient.

You can just plug those formulas into hammer and watch the magic happen!

Edited June 15, 2021 by NikiOo

[fishingbrick]

That's the longest joke I've ever seen and I applaud you for that. All jokes aside, I've been working on a top-down layout for about 4 hours, and I've only come up with an extremely simple 3 lane layout that I don't think will play very well.

[NikiOo]

Why do you think a 3 lane layout is not interesting? Most defuse maps follow a very similar four square pattern. That doesn't prevent them from doing a lot of original things within individual areas.

Also, I think you'll have better luck coming up with interesting ideas if you play around with it in the editor rather than on paper.

Here's something you can do - take a cool real life location that you would like to make in the game (this will motivate you) and think about how you can design a wingman layout around this location. You will find yourself solving interesting architectural problems and even if you end up with a three-lane layout still, it will be an interesting three-lane layout because of the constraints of the theme. You may not even notice that you have a three lane layout just like cache is pretending to not be a four-square layout while it is exactly that.

Edited June 18, 2021 by NikiOo

[fishingbrick]

15 minutes ago, NikiOo said:

Why do you think a 3 lane layout is not interesting? Most defuse maps follow a very similar four square pattern. That doesn't prevent them from doing a lot of original things within individual areas.

Also, I think you'll have better luck coming up with interesting ideas if you play around with it in the editor rather than on paper.

Here's something you can do - take a cool real life location that you would like to make in the game (this will motivate you) and think about how you can design a wingman layout around this location. You will find yourself solving interesting architectural problems and even if you end up with a three-lane layout still, it will be an interesting three-lane layout because of the constraints of the theme. You may not even notice that you have a three lane layout just like cache is pretending to not be a four-square layout while it is exactly that.

I took the liberty of doing this and found an interesting location. Now to port it into a layout. To me, coming up with the ideas is more time consuming and frustrating than actually making the map. One of the big problems I have is finding interesting locations in google maps or earth. The problem being that you have the whole world to work with.

[blackdog]

Follow on social accounts that post architecture, like there’s one I remember called Abandoned Places, then you can find many Pinterest boards managed by designers where #reference #inspiration shots are posted.

I have been collecting scraps of paper from tourist leaflets and magazines as well. I look at couple shots from different angles and I try to figure out what’s the real layout and if it suits a map, or how two locations in the same town would connect and make for interesting gameplay.

 You could also start by actually copying someone else layout and then in the process you’d end up changing things. Or think of a reskin.

You could also create small areas, what would suit for choke points, sites, etc, create a bunch and then try and combine them like Lego.