DESIGNER-STUDY · 2026-08-03
Inside Scott Kim's Philosophy — making the right answer belong to the solver
Forty years out from a two-line definition: a puzzle is fun, and it has a right answer
Intro — the man who insisted there is a right answer
Scott Kim (born 1955) is an American puzzle author and game designer. Outside the US he is probably best known as "the ambigram guy" — the designer of words that read as something else when rotated or mirrored. But he has a second career. Since going independent in 1990 he has written thousands of puzzles for Discover and Games magazines, and designed the puzzle modes of commercial puzzle games including Bejeweled 2 (Neurology Now, Sept/Oct 2009 ↗; Jeremy Gibson Bond, Introduction to Game Design, Prototyping, and Development, ch. 12 ↗).
The reason to revisit him now is that nearly every puzzle author this site has covered stands downstream of one event, knowingly or not: the all-day workshop "The Art of Puzzle Game Design" that Kim ran with Alexey Pajitnov at GDC 1999 and 2000. The handout from those days became the shared vocabulary for "what a puzzle is" for the next quarter century.
This is not a survey of his work. It reads across what he has written and said in public to find what he argues consistently. Every quotation here comes from a source I read in full; my own interpretation is kept to the final section. I made a pot of hot bancha before starting.
Background — came in through music, turned by Tetris
He earned a BA in music from Stanford in 1977 and, in 1988, a self-designed PhD in Computers and Graphic Design (Neurology Now, 2009 ↗). In a 1985 interview he described where he actually learned computing: "In 1975, I started a sequence of computer-music courses taught in the Stanford music department, and what was important, besides its being a good class, was that it was taught at the Stanford Artificial Intelligence Lab," calling the lab "my breeding ground" (Programmers at Work, 1985 interview ↗). Douglas Hofstadter of Gödel, Escher, Bach was at the same lab.
His life as a puzzle author started earlier. His first invented puzzle — a letter of the alphabet folded once, which you must identify — was written up by Martin Gardner in Games magazine (Scott Kim, "What is a Puzzle?" ↗). Isaac Asimov called him "the Escher of the Alphabet" (Neurology Now, 2009 ↗).
The decision to become a game designer came from playing Tetris in the late 1980s: "Tetris took a classic geometric puzzle and adapted it brilliantly to make a highly original computer game—the first computer puzzle game. I thought, that's the type of game I want to make" (Neurology Now, 2009 ↗). Ten years later he would be sharing a stage with the man who made it.
Philosophy — "A puzzle is fun, and it has a right answer"
His definition is two lines long: "1. A puzzle is fun, 2. and it has a right answer." He presents it as "my favorite definition," arrived at in conversation with the puzzle collector Stan Isaacs, and annotates it: "Part 1 of the definition says that puzzles are a form of play. Part 2 distinguishes puzzles from other forms of plays, such as games and toys" ("What is a Puzzle?", The Games Cafe ↗). Two decades later he opened his EG 2008 talk from the same definition (TED / EG 2008, "The art of puzzles" ↗).
What matters is that the two lines are not symmetrical. "It has a right answer" can be checked objectively; "it is fun" cannot. He says so himself: "like beauty, fun is in the eye of the beholder. What may be fun for one person may be torture for another." So he defines the designer's first duty accordingly: "my first job as a puzzle designer is to tailor my puzzles to the tastes of my audience" (ibid.).
He does not lament this asymmetry between subjective fun and objective correctness. He draws his most characteristic sentence out of it instead: "It tickles me to think that for every problem in the world, no matter how tedious, there is someone who would leap at the chance to figure it out" (ibid.) — written in the middle of an argument about whether washing the dishes is a chore or a game, which depends entirely on whom you ask.
Obsessions — build the toy first, never make it hard with controls
The number of practical rules he repeats is remarkably small. First: "to design a good puzzle, first build a good toy," with the reason attached — "The player should have fun just manipulating the puzzle, even before reaching a solution." Borrowing Chris Crawford's four categories (games, puzzles, toys, stories), he places the toy as the layer beneath the puzzle ("What is a Puzzle?" ↗). His example: you can enjoy rotating blocks in Tetris before you know what the goal is.
Second: never confuse where the difficulty comes from. His example is Rubik's Cube — "On computer, Rubik's Cube feels terrible because it is awkward to manipulate a three-dimensional object with two dimensional controls." The lesson is one line: "be sure the player is challenged by the puzzle itself, not the controls for the puzzle" (ibid.). Written over twenty years ago, and it still lands squarely on today's touchscreen puzzle apps.
Third: sequencing. As Jeremy Gibson Bond later organised the GDC 1999 slides in his textbook, Kim's process runs in eight steps (inspiration, simplification, construction set, rules, puzzles, testing, sequence, presentation), and the sequencing principle is this: introduce every new concept in isolation in its most elementary form, then raise difficulty within that concept alone, then finally mix it with concepts the player already knows (Bond, ch. 12, citing GDC 1999 slide 97 ↗). That the construction set comes before making any puzzles is very much this man.
And running through all of it, a taste for perceptual shifts. The folded letter, figure-ground reversals, ambigrams — the answer is in plain sight but invisible until you switch interpretations. "To solve this puzzle you must change how you interpret the picture. Personally, I enjoy puzzles that involve such perceptual shifts" ("What is a Puzzle?" ↗).
Failure and response — it works in the classroom, and loses the moment class ends
He has said little in public about failures inside the games industry. Within what I could read, the failure he narrates repeatedly comes from a different battlefield: mathematics education. In a February 2024 post he describes precisely how he loses. "First I beg teachers to give me time to play math games with students. I bring in games like Rush Hour, Prime Climb, or Set to class. Everyone plays and has a good time." So far, a success. "But then everything goes back to 'normal'. I've just shown them 'fun' math. But in their minds, teachers and students wall that off as an aberrant experience, like a once-a-year field trip" ("Joyful Mathematics," Feb 10, 2024 ↗).
This is not a one-off defeat. "Every math reform I've lived through has been royally botched," he writes, naming Common Core as an example ("5 barriers to spreading joyful mathematics," Feb 10, 2024 ↗). He also cites his hero Martin Gardner, who wrote in 1998 that forty years of trying to move educators had produced glacial movement: "to hear him despair at his inability to influence educators is sobering" (ibid.).
His response was not a frontal assault. "My personal belief is that conventional math education is so badly broken that it cannot be incrementally repaired." So what then? "I personally do not want to engage in curriculum standard wars. Instead, I prefer the carrot over the stick. I advocate establishing an attractive alternative that shows everyone a better way" ("Joyful Mathematics," 2024 ↗). His model is Sesame Street — a generation won over by being entertaining rather than by being mandated.
There is one record of a failure at the workbench too. Bond's textbook reports that even after many years of experience, Kim still finds that puzzles he expects to be simple turn out surprisingly hard, and ones he expects to be hard get solved instantly (Bond, ch. 12 ↗). That is presumably why "testing" stands as its own step among his eight.
Dilemmas — how to put the thrill of competition back into a solitary act
One tension he explicitly names as his own. It follows from the premise that puzzles, unlike games, are solitary: "One of the main challenges for me as a game designer is to find ways to introduce some of the thrill of competition back into puzzle games" ("What is a Puzzle?" ↗). He wrote that in response to a colleague's column praising board games as social occasions — long before online leaderboards and daily-puzzle sharing were ordinary.
There is a second point where he does not use the word dilemma, but a tension sits between two statements. Around 2000 he wrote: "play starts by suspending the rules of everyday life, giving us permission to do things that are not practical. Folded letters certainly don't have any practical value" (ibid.). By 2009, preparing a brain-training book with the neuroscientist Richard Restak, he was saying: "If you find puzzles a waste of time, think of them as mental exercise, like jogging or doing yoga. Puzzles aren't just fun, they're good for you too" (Neurology Now, 2009 ↗). The man who made impracticality a condition of play is now selling puzzles on their usefulness.
I do not read this as a reversal so much as a change in ordering. The position he reaches in 2024 — "you can kill interest in any subject by just teaching the mechanics without the meaning" ("Joyful Mathematics," 2024 ↗) — does not deny usefulness. It says that usefulness placed first is fatal. It is an argument about sequence, and it has the same shape as his design rule: build the toy first.
Third, the fit between tailoring to an audience and being original. He writes that his first job is to tailor to his audience, yet he has also said he has no interest in working inside somebody else's matrix (per a summary of his EG 2008 talk ↗ in the AMS Graduate Student Blog, 2016 ↗). His solution at Discover was not compromise but division: split one puzzle into several questions ranging from very easy to very hard ("What is a Puzzle?" ↗).
Influences — the names he gives himself
For the definition he credits the puzzle collector Stan Isaacs. For the taxonomy he explicitly borrows Chris Crawford's 1984 book The Art of Computer Game Design, which he calls "still the most articulate book on the subject." On the spirit of play he calls Bernard DeKoven's The Well Played Game "a wonderful book" and summarises its key idea: "in a well-played game, players are willing to alter the rules to keep the game fun for everyone" (all from "What is a Puzzle?" ↗).
Of Martin Gardner he writes plainly: "Martin Gardner was a hero to me and many other budding mathematicians. He opened the door to the joys of mathematics for more people than anyone else in history" ("5 barriers," 2024 ↗). On the texture of his own work he self-reports, describing his puzzles as "in the spirit of Tetris and M.C. Escher—visually stimulating, thought provoking, broadly appealing, and highly original" (Neurology Now, 2009 ↗).
And then the creed about notation. In 1985, before he had gone independent as a puzzle author, he said: "It is important to realize that all notations, whether music, or language, or computer languages, are just made up. They are symbols that can be changed. There is a choice. The ability to change notation empowers human beings" (Programmers at Work, 1985 ↗). Forty years later, the reason he is angry at school mathematics — that it teaches only the reproduction of symbols — is that sentence turned inside out.
Kizuki's reading
What follows is my own interpretation. I read Scott Kim as a designer who transfers ownership of the right answer. Part 2 of his definition — a puzzle has a right answer — and the school mathematics he spent his later years fighting are, on the surface, handling the same object. Both have right answers. What differs is whose they are. In school the right answer is already somebody else's, and the student reproduces it. In a puzzle the right answer arises on the solver's side, as a discovery. Same correctness, opposite direction of ownership. That, I suspect, is why every one of his design rules turns out to be about the order in which things are handed over. Hand over the toy before the goal. Hand over each new concept alone. Hand over meaning before usefulness. Even the word "empowers" in his 1985 line — "the ability to change notation empowers human beings" — reads to me as being not about handling symbols skilfully but about repossessing them. What he has spent a career designing is less the puzzle than the instant in which a right answer stops being someone else's and becomes yours.
Closing — where to start
The shortest way in is to read "What is a Puzzle?" — five pages that contain almost his whole argument. Then watch the thirteen-minute EG 2008 talk, where his inventions, from folded letters to figure-ground reversals, are shown in motion. If you want to play something, Bejeweled 2, whose puzzle mode he designed, is the easiest to find.
For adjacent reading: Alexey Pajitnov, who shared the stage with him (designer-pajitnov ↗); Frank Lantz, still working on giving abstraction a physical weight (designer-lantz ↗); and Zach Gage, who remakes games people already know (designer-gage ↗). The Soma Cube, one of the objects he carries into classrooms, has its own piece here (The Soma Cube, 1933 ↗).
Sources
Sources consulted for this article (all read in full):
・Scott Kim's blog, "Joyful Mathematics — What it is and how to spread it," February 10, 2024 ↗
・TED / EG 2008, "Scott Kim: The art of puzzles" (talk by Kim) ↗
Reactions (no login)
Anonymous • one of each per visitor per day
関連シリーズ
The Designer's Philosophy第37回 / 全37回
