Puzzle game design

A collection of articles, talks and videos (and maybe eventually, my thoughts) about puzzle game design. I expect this to eventually grow into more coherent note but for now, I’m just making quick notes on references.

• What Makes a Good Puzzle? - YouTube by Game Maker’s Toolkit is a great primer on good puzzle design principles of assumptions, catches and revelations with good video games examples.
  • The Mechanics
    • The rules and limitations of the mechanics form the basis of the puzzle
    • It’s important that the goal of the puzzle is clear for the player. The challenge should arise from figuring out how to get there, not what it is
  • The Catch
    • A good puzzle is built around a catch: a logical contradiction where two things are seemingly in direct conflict with one another
  • The Revelation
    • Solving a puzzle is a revelation and a discovery that makes the player feel good when they succeed to figure out the problem.
    • Mark shares a really good example from Lara Croft GO: the game has tiles that break after you go through them the second time. In one of the puzzles, this mechanic is first used to subvert your assumptions by showing how this can be used to deal with enemies, only to make that the incorrect solution in the second part of the puzzle.
    • Many times a great puzzle is one where it seems like there’s a simple and obvious solution but once you try it, it doesn’t work and it forces you to first get rid of that straight-forward assumption and then rebuild your mental model to figure out how to solve it.
      • Mark also makes the point that this is not only done to trick the puzzle solver but to also give them a starting point — perhaps in figuring out how elements of the puzzle like controls, restrictions and keys & locks work.
  • The presentation
    • Puzzle should only have as many elements as are necessary to make it work. More elements can be distracting or lead the player to the wrong things.
  • The curve
    • Puzzle difficulty should gradually go up and build on top of what the player has learned before
• How Baba Is You Makes Brain Busting Puzzles by Game Maker’s Toolkit discusses the great Baba is You puzzle game and how its designer Arvi Teikari approached designing puzzles for the game in a reverse engineering manner, starting from the solution or mechanic and walking backwards to design the puzzle.
  • Part of good puzzle’s design comes from preventing the player from doing things rather than enabling them to do them
  • A good puzzle game teaches you concepts, mechanics and solutions bit by bit so that the player doesn’t feel overwhelmed by the more difficult puzzles too early. On the other hand, puzzles need to get more difficult fast enough to maintain a good difficulty curve and prevent player from becoming bored.
  • Baba is You doesn’t have a tutorial as such. It doesn’t tell you “push this block here to win” but rather starts with very simple puzzles where you end up figuring out the mechanics and solutions because there’s nothing else you can do. Solving a puzzle by using the mechanics is a way more effective way of teaching problem solving than telling the player how to solve them.
• Empuzzlement is a discussion that took place in IndieCade 2013 with Jonathan Blow (Braid, The Witness), Marc ten Bosch (Miegakure) and droqen (Starseed Pilgrim, The End of Gameplay). They talk about what makes good puzzles and why.
• Level Design Workshop: Solving Puzzle Design is a talk by Jolie Menzel from GDC 2016. She’s focusing on puzzle within the level design perspective
• The Art of Puzzle Design | How Game Designers Explore Ideas and Themes with Puzzles and Problems - YouTube
• Open-Ended Puzzle Design at Zachtronics - YouTube
  • “In this 2019 GDC talk, podcaster Drew Messinger-Michaels talks to Zachtronics founder and creative director Zach Barth about the studio’s puzzle design process, from the initial foundation to the basic mechanics, to the way story is integrated.”
• In How to Design a Puzzle Game In 5 Steps - YouTube Jonas Tyroller shows how he designs puzzles through prototyping. I find his process of starting with random rules without declaring or deciding any objectives very interesting.
  • His thesis is that in a good puzzle
    • actions have predictable consequences
    • the rules are usually not complicated
    • the challenge is often to find a sequence of actions
    • includes all information needed to solve it
    • is only as complicated as it needs to be
    • and is only as big as it needs to be
    • looks nice
    • toys with your expectations
    • teaches you something new about the game system
    • usually goes through many iterations (during design phase)
    • are playtested a lot (by other people than you)
• Introduction to Puzzle Design - YouTube
  • Great talk about what Brett Taylor has learned about designing puzzles through developing Linelight.
  • He starts with the concept of “working memory” and how different concepts and the amount of different options take a certain amount of memory to operate and how for a beginner player, that amount is usually much larger than for an experienced player.
  • He talks about “flow state” and how achieving that requires the difficulty/challenge to be in a sweet spot between too difficult (creates anxiety) and too easy (creates boredom)
  • Puzzle characteristics
    • Hand-crafted vs procedurally generated
      • Hand-crafted = finite amount of puzzles, few skills repeated, more novel solutions, authored content, fresh and varied
      • Procedural = infinite amount, skills constantly repeated, relatively not novel, puzzles born from system, predictable experience
      • This is a spectrum rather than a binary
      • This talk is mainly about hand-crafted puzzles
    • Noise
      • the most important part of the talk according to him
      • anything that uses player memory without contributing to the solution of the puzzle
      • he mainly talks about how noise is a bad or undesired thing but also shows some examples of how noise can be used well to make an obvious solution more obscured to increase difficulty
      • can be used to prevent the player from accidentally solving a puzzle
      • finding the right amount of noise is a puzzle in itself
      • red herrings are often used noise in puzzles to lead player’s first assumption to the wrong direction
  • 7 Lessons
    1. Simplify, simplify, simplify
       • Reduce noise to create cleaner, tighter puzzles
       • Start by describing the solution, then remove everything you didn’t mention.
       • If you want to include noise, start with none and then add intentionally and cautiously
    2. Cut pointless levels
       • For each level, ask why it’s there
       • Possible level purposes:
         • Teach a mechanic
         • Reinforce a skill
         • Unique identifiable moment
         • Break in pacing
    3. You’re the expert player
       • You know your game and its controls better than most players ever will
       • Try to finish your level by playing with your feet only. If it’s too hard, make it easier.
       • Knowing the solution but not being able to execute it sucks
    4. Keep action and puzzles separate
       • For single-solution puzzles
       • Action adds randomness and distractions when puzzles require patterns and consistency
       • Player should know if a level requires action or puzzle-solving skills.
    5. Make solutions unambiquous
       • Red flags:
         • “Was that right…?”
         • “Oh, I solved it?”
         • Players attempting impossible actions
       • A puzzle should be solvable by thinking before action rather than trial and error
       • If the player is not meant to reach a platform or a location in time, make it obvious it’s not the solution.
         • For platformers, make every jump obviously possible or impossible.
    6. Player trust
       • If players assume no agility needed and you break that trust once, they’ll wonder if every puzzle requires agility
       • Player perception is all that’s real
    7. Exhaustive design vs fun
       • Fun = exhibiting mastery, skill or wit
       • Exhaustive design = one of every possible thing
       • sometimes these are incompatible and you need to choose
       • fun for designer != fun for player
  • How Brett makes puzzles
    • He starts with mechanics and the levels/puzzles emerge from interactions of mechanics
    • create mechanics that
      • are diverse
      • could interact in many ways
      • excite you
    • Experimentation → great stuff
    • Goal: have few mechanics that yield many puzzles
    • Start with just a few mechanics and try to exhaust the possibilities
    • “Solution sentences”
      • Describe the solution in a sentence and find the simplest puzzle to implement that solution
      • Elegance is inevitable as simplicity is intrinsic to elegance
      • Each sentence is like a magic trick’s secret
        • If the player knows the sentence going into the puzzle, it’s trivial. But they don’t so there’s magic.
• How These Puzzles Deceive You - YouTube
  • This video shows a few examples of puzzles that initially deceive you by using design techniques that make it easy for the player to initially make the wrong assumption and act on it.
  • They call it “double take” puzzle design

• Puzzle Game Magic Secrets by Brett Taylor

• In The Two Types Of Puzzle Games, Artindi talks about two layers of novelty
  • First layer is the unique mechanics that found the foundation of the game
  • Second layer is how these mechanics are explored and used in different ways
• Patrick of Patrick’s Parabox gave a talk in GDC about system-centric puzzle design: System-Centric Puzzle Design in ‘Patrick’s Parabox’ and I would have watched it if Patrick’s Parabox didn’t burn my brain every time I try to think about it.
• How I made the world’s Trippiest Portal 2 map
  • Igrium talks about his Portal 2 map — The Intruder — and some design choices he made and others he thinks he would now do differently. The video spoils all the puzzles so if you want to experience it, play it first.
  • The video is also not purely about puzzle design but a lot of details on how he implemented many technical aspects of the mod but there are a couple of really good insights for puzzle design.
  • “In my opinion, less elements that are used in interesting ways go a lot further than more elements that just do one thing”
  • He talks about using some of his special effects to also give player a heads up and time to prep for catching a flying cube
  • In the same puzzle bit, he talks about how some players were able to come up with an alternative solution but because of that, they didn’t learn an important things (catching cube) that was required in a later part.
  • One of the interesting things in this video is how he explains a bunch of ways people tried to solve the puzzle in ways not intended and how he finds it really difficult to catch and circumvent all those
• For Portal 2, there’s a community made collection called No Elements that expertly showcase how you can build very exciting puzzles with very limited amount of mechanics.