Solving a puzzle can sometimes be substantially difficult. Designing one can be even more challenging. This guide aims to explain the fundamental pieces of what makes a puzzle clever and satisfying, as well as what should be avoided when designing one. Covered first is the integration of a logical problem, followed by more specific sections such as information clarity and antitrapping. To begin, my personal definition of a puzzle...

A puzzle is defined in my own terms as:
A logical, nonlinear reusage of components to complete a satisfying objective.

It sounds complicated, but fret not. Good puzzles can be broken down and analysed to see why they work. Puzzles, first of all, must be largely based on logic and problem-solving. Trial and error, memory, roller coasters, mazes, bullet hell and physical dexterity do NOT apply. But what makes up the logic of a puzzle? Well, the answer is the order of steps and reuse of puzzle elements. The player doesn’t want a linear set of steps to accomplish, because that merely becomes a chore. To make a puzzle nonlinear, puzzle elements must be reused and interact with each other in different, intriguing ways as the puzzle progresses.


Let's use this Wheatley puzzle from the campaign as an example.
Imagine if the puzzle's exit was at ground level and the cube started on the same side as the player does. The player’s steps would consist of:

1. Push the turret away with the funnel
   
2. Get the cube
   
3. Set the cube on the button to reverse the funnel
   
4. Transport yourself to the exit using the reversed funnel

How vapid would that be! The cube is immediately obtained and the funnel only serves to
rid the turret and transport the player to the exit. The way the puzzle is actually designed means the funnel and its polarity is used multiple times. However, take note how many steps are added not by adding elements, but changing the layout of the elements and geometry. Thus, in the actual puzzle, the player is required to:

1. Push the turret away with the funnel
   
2. Reverse the funnel with the floor button
   
3. Suck the cube up with the funnel
   
4. Un-reverse the funnel by stepping off the floor button
   
5. Transport the cube to the player’s side
   
6. Set the cube on the button to reverse the funnel
   
7. Transport yourself to the exit side using the reversed funnel
   
8. Push the cube off the button with the funnel
   
9. Use the un-reversed funnel to transport yourself vertically to the exit

Much better! The funnel is used to rid the turret, transport a cube, transport the player, and remove the cube off of its own button. The button to reverse the funnel’s polarity is used once by the player and again by the cube, making it not only used multiple times, but also in a couple of different ways. This is accomplished with a very concise design, consisting of only the turret, a button and a funnel. This is a simple example, and not all campaign levels fully embrace nonlinear design, but the idea of similar nonlinear steps is the core of puzzle design.


Another example is this other Wheatley puzzle from the campaign.
The puzzle uses the funnel as not only a way to gain height to fling to the frankencube, but it also serves as a way to transport the cube to the ceiling button, as well as creating a barrier to the redirection cube. The angled surface to reach the frankencube is also multipurposed. The player has to fling out of it but also fling into it later. The laser is blocked with one cube, then redirected with the other. As you can see, it's all about reusing elements in different ways.

Nonlinear design works best when elements are reused a few steps apart from each other. When elements are reused in immediate succession (for example, pressing a button and being required to immediately press it again), the process still remains a little too straightforward if the element is discarded after those two uses. However, single-use elements and elements reused in succession are acceptable if they enforce other elements to be given more reusage (for example, an exit button forces a setup to get a cube to it). The most satisfying puzzles reincorporate a core concept used in a previous state of the puzzle, perhaps with a twist, that invokes a sense of satisfaction.

Information clarity is integral to making your puzzle's presentation approachable. Everything should be clear to the player's eye as soon as they walk into a chamber, otherwise they will feel lost, overwhelmed or confused. As such, there are an abundance of techniques a designer can use to highlight important information and direct attention to less noticeable areas.



Circuit uses lighting contrast to highlight important sections and is well-lit throughout.
A paramount part of information clarity is providing your puzzle with sufficient lighting. A dark chamber instantly strains the player’s attention and makes it taxing to solve your puzzle. If you also use the Better Extended Editor or Hammer, there are additional lighting possibilities you can use. For example, in my most recent maps, I contrast the cold, neutral environment with warm spots where important locations, elements and portal surfaces are positioned. However, it is important to note that the Better Extended Editor or Hammer is not required for lighting or a good puzzle. Standard Puzzlemaker's lighting options are perfectly sufficient. The Better Extended Editor is simply an extension of the possibilities, not a replacement.

There is also a prominent mistake that requires an insistent clarification: DARKNESS IS NOT A PUZZLE ELEMENT. People make the mistake of designing levels around finding light sources to progress. Darkness as an intentional design choice only adds artificial difficulty and unfair gameplay; don't do it. Geometrical navigation should be clear; it must be the logical steps of a puzzle that players need to discover how to navigate.



Obstruction uses branching antlines to keep connection indication tidy.
All indicator lights should be as neat as possible to easily distinguish what does what. A messy web of indicator lights is immediately deterring and overwhelming to a player. It can also be a good idea to keep elements and their connections in a relatively close sight to each other, but this may be subject to the kind of layout required for the puzzle. Connection signage is also an option if antlines are not viable, but using connection signage should be a last resort. It generally isn't ideal because it is visually disconnected from its corresponding item.



Anfractuous has a clear end goal observable at the chamber's entrance.
The player's goal, aptly the exit door, should always be visible, so they have an opportunity to create a plan to reach it based on its conditions and location. Without a clear goal, the player can become lost and confused. It can also be good to make smaller goals more visible, such as obtaining a cube or reaching a platform. Mini goals can especially be of help in larger, more difficult puzzles to give the player a sense of direction. They'll gain satisfaction from these drip-fed accomplishments and are less likely to feel overwhelmed.



String lets the player see every piece of information in one open area.
It can also be very helpful to attempt to design your puzzle’s layout as being quite open, reasonably sized and with elements arranged neatly. This way, the player can see most of the puzzle’s information as soon as they walk in. They may have an easier time dissecting the way the puzzle is constructed so they can formulate a plan.



A Familiar Taste uses only a faith plate and blue gel, but reuses them to their potential.
Nobody wants a convoluted mess of puzzle elements, it just appears obscene! A good puzzle uses as few elements as possible with as much reusage. Not only will few puzzle elements make your level appear more approachable, but it also makes the puzzle more clever when just a handful can enforce so many steps. It especially helps give the player that feeling of confidence, followed by the realisation that there's more to the puzzle than they originally concieved.



Trapdoor uses a vertical laser to direct the player's view to the ceiling.
One method to guide the player’s gaze is to use element projections such as excursion funnels, light bridges, laser beams and high-energy pellets. For example, in my puzzle Trapdoor, there are two integral surfaces up high on the ceiling. The player would typically take some time to notice them with no guidance, but the way the puzzle’s laser shoots vertically, it directs the player’s attention upwards. From there, they are likely to see those two surfaces.

Antitrapping is an extremely important consideration in a puzzle. A player should NEVER be forced to kill themselves or manually restart a puzzle from the game’s menu. A lack of consideration for antitrapping is lazy design.

Ultimately, it leaves the player without room for experimentation and adds an arbitrary blockade to a possible plan they have. When a player makes a mistake, they should be able to feel that they have merely made an incorrect plan that they must adjust, not that they have failed. Hence, a trapping situation is essentially a moment where the player can fail their plan. It also breaks immersion and flow of a puzzle’s logic when the player suddenly finds themselves falling into inescapable situations or constantly worrying about avoiding them.



Trapdoor has a return path that is also a part of the puzzle itself.
Antitrapping should ideally be unnoticeable to the player. Changing the layout, design, geometry, element setup and logic of a puzzle should be the first priority to removing trapping situations. Return paths (such as a one-way staircase or drop) and mechanical resets (such as a respawn button for a cube dropper in the event that the cube cannot be retrieved) are the next potential way to implement antitrapping.

However, while return paths and mechanical resets are effective ways to prevent trapping situations, they also add either extraneous geometry or elements to the puzzle that may only serve as a reset rather than a necessary component to the puzzle’s solution. It is ideal, in this case, to attempt to naturally incorporate return paths and mechanical resets into the puzzle’s solution itself. A deeper elaboration of incorporating mechanical-assisted fixes can be found in the Polish and Tweaking section.

In a puzzle, a red herring is an extraneous component, whether it be mechanical or geometric, that serves as only an arbitrary method to distract the player from their objective. Because of this, red herrings should NEVER be included in a puzzle. It can essentially be thought of as a game of chess, but in which one player has an additional piece that they cannot move.



Red herrings serve no functional purpose in a puzzle’s solution, which cheats the player’s problem-solving abilities with a cheap distraction. All parts of a puzzle should be necessary to the solution and should make the player think about:

1. What order to use components in
   
2. How components can interact with each other
   
3. Consequences of certain decisions
   
4. NOT what components are relevant information

A player’s incorrect assumption on a puzzle’s solution should be a fault of incorrectly identifying the path of points one, two and three. This ensures the player is given a fair logical puzzle and all mistakes are a fault of their own thought processes. To put it simply, the solution to a puzzle should be hidden in plain sight but no extraneous elements should attempt to enforce this.

It is important for a puzzle to function with preferably one intended solution. Beating a chamber that is designed to always be one step ahead until the very end is extremely satisfying. As such, no puzzle elements should be left unused or irrelevant. Slight differentiations in the intended solution are acceptable as long as they do not break the core of the puzzle.

Do NOT reward the player for finding design flaws. A common mistake in puzzle design tends to revolve around “rewarding” the player for creativity in breaking a puzzle. There should never be an instance where this is an acceptable mentality to puzzle design. Opting for letting the player cheat makes your puzzle design appear lazy and unpolished, and ultimately reaps any player’s satisfaction of beating it in the manner you intended. Think of it in this way: Is the player clever for finding a design flaw, or is the creator lazy for dismissing a design flaw? The latter is the correct answer. Optional challenges in puzzles, however, are acceptable. There is no wrong in constructing a side challenge so long as it is distinct and does not break the main puzzle’s logic or fairness.



When polishing a puzzle, two methods of patching up holes arise: Seamless and Mechanical. These methods of polish often arise when attempting to resolve a trapping situation, however they can also arise when attempting to prevent an unintended solution.



A Familiar Taste uses a single surface in a compartment to prevent a trapping situation.
Seamless patching relies on managing space, the natural geometry and layout of a puzzle in order to fix a design flaw. This is the best method to solving a design problem, as it adds no extraneous puzzle components to assist in fixing it. Alternatively, some component’s properties can be altered to more naturally fix the problem.

For example, timers can be extended or decreased for the player’s convenience or to restrict potential shortcuts. With the Better Extended Editor for Portal 2, there are also possibilities to add delays to elements' activations or deactivations and changing the speed of excursion funnels. Another alternative involves using return paths to fix trapping situations, however this verges on being mechanical patching (which is less preferred) since it adds/carves extraneous geometry in the puzzle. That is, unless the return path is naturally incorporated into the solution of the puzzle. It must also be noted that the alteration of element’s properties must make sense and not deviate significantly to where they do not follow consistent or expected behaviours.



Trapdoor originally had two buttons on the upper and middle levels, of which one only served as a reset button. It was changed to a single timed button connected to the cube dropper.
Mechanical patching is used when other components assist in resolving a design flaw or complication. For example, an aerial faith plate may not reliably launch a cube onto a button, so an excursion funnel is placed to reliably rest the cube on the button. Mechanical patching is not preferable as it adds extraneous components to the puzzle that, whilst are integral to the playability of the puzzle, are not necessarily integral pieces of the puzzle itself.

Because of this, it is highly encouraged that the geometry, layout and space of the puzzle are altered first in an attempt to rectify a design complication before resorting to mechanical patching. Of course, alike seamless patching, there is also the possibility to attempt to include the assisting element’s usage into the solution so it does not remain extraneous. As an example, my map Circuit uses a funnel to guide a cube to the exit button, but the player also uses it to hop over a glass fence in the beginning of the puzzle.

And so concludes Fumbly Bumbly's Puzzle Design Principles. If it looked like a complicated process, that's because... it is. This is the final, and most important, lesson to learn. A true work of art takes rigorous amounts of time and energy to accomplish. An artist needs to learn to rethink, redesign and even throw away their projects. Many puzzles of mine have not seen the light of the workshop because they simply weren't good enough. Sometimes the ability to let go of a project can, strangely, be the best way to progress. I hope this point is made clear.

A handful of my puzzles that never met the workshop.

Thank you for reading, and thank you to all the fantastic puzzle designers out there who have inspired me to make some of my own. Provided below are some very good guides about puzzle design as well as a collection to some of the best maps in the workshop.



Game Maker's Toolkit:
https://youtu.be/zsjC6fa_YBg

Demon Arisen's How To Make Great Test Chambers series:
https://youtu.be/snZAq2GMOMU
https://youtu.be/fflYyOTjBUA
https://youtu.be/29VPInw1IFQ
https://youtu.be/iuv83MgxI0w

Mikeastro's Portal 2 Documentaries:
https://youtu.be/L9BD004V9bI
https://youtu.be/1pZQGZ_iAp8
https://youtu.be/uyk38-nIlqI
https://youtu.be/y6ac80Kl_cM

Rector's Guide to Making Better Maps:
https://steamhost.cn/steamcommunity_com/sharedfiles/filedetails/?id=642040698

HugoBDesigner's Guide to Monoportal Mapping:
https://steamhost.cn/steamcommunity_com/sharedfiles/filedetails/?id=1300713355

RedSilencer's Introduction to the Workshop collection:
https://steamhost.cn/steamcommunity_com/sharedfiles/filedetails/?id=470046703

The Best of the Best workshop collection:
https://steamhost.cn/steamcommunity_com/sharedfiles/filedetails/?id=1590432047