There’s No Such Thing as an Educational Game

What makes a game teach — and why points and badges don't

Children learn by playing. Nobody finds this surprising anymore — you can see it in preschool curricula, in educational toys, in apps that teach counting by counting little sheep. What’s harder to admit is that adults learn exactly the same way; they just rarely have the courage to notice. This intuition — that play and learning aren’t opposites but deeply intertwined — has even acquired its own name and its own moment of fashion: gamification. With mixed results. Slapping game mechanics onto content that has nothing game-like about it rarely works — and that’s a good starting point for asking what a game actually is, and why it sometimes teaches and sometimes only pretends to.

What Is a Game

For the purposes of this text, I’ll work from my own definition. A game, as I see it, is something that satisfies four conditions simultaneously.

First, a game is a model of a fragment of reality that you take part in. Not all of reality — always some slice of it, simplified and sharpened. And not a model to be watched from outside — you have to act inside it. Without the first of these, it’s hard to distinguish a game from pure improvisation, where nothing represents anything else; without the second — from a film, which is watched from outside.

Second, a game has clearly defined rules. This condition filters out things that are close to games but aren’t quite — a casual social gathering, playing with figurines without an established goal. Something valuable happens there too, but it isn’t a game in the sense I want to use here.

Third, a game is a safe environment for trying things out. Its consequences are symbolic, agreed-upon — which sets it apart from gambling, sports betting, or competitions with real financial or life stakes. Mathematical game theory deals with those situations too, but in my view they belong to a different category: there the stake is a real resource; here it’s only a symbolic one.

Fourth, a game is undertaken voluntarily. It must be something one wants to do without being compelled — something that is its own reason, not merely a means to an end.

These four conditions didn’t emerge in a vacuum. Johan Huizinga in Homo Ludens saw play as a free activity, separated from everyday life by time and place, governed by its own order, and free from material interest — which maps well onto my second, third, and fourth conditions. Bernard Suits took a different approach: he defined a game as a voluntary attempt to overcome unnecessary obstacles, where rules are accepted precisely because they create those obstacles. An elegant formulation, but one focused on structure — it doesn’t ask what those obstacles are actually supposed to serve.

And that’s where I want to mark a difference clearly. Neither of these definitions says explicitly that a game is a model of something. Yet that, to me, is the heart of the matter — and the bridge to the rest of this text.

The second difference is subtler but equally important. In Suits, rules are an obstacle the player voluntarily imposes on themselves — something that makes things harder, and by doing so gives the attempt its meaning. I look at rules differently: as a source of safety. A rule that is clear and stable gives predictability — I know what to expect, I know the world of the game won’t shift beneath me mid-attempt. That’s why a game can be a safe space for testing limits, strategies, and even oneself.

An interesting edge case for this definition is the traditional tabletop role-playing game — not its computer adaptations, but the live form with a Game Master. There, rules tend to be less formalized, sometimes negotiated mid-session, depending on the system. And yet the game doesn’t lose its predictability, because that function is taken over by a person — the Game Master acts as a living arbiter who keeps the rules from dissolving into anything-goes. Which is good evidence that what matters isn’t the letter of written rules, but whether someone — or something — guarantees their consistent application.

Games as Models of Reality

Let’s return to what sets my definition apart — and what I believe is the key to understanding the educational potential of games. A game is a model of a fragment of reality that you take part in.

What does that mean in practice? That in designing a game, someone chose a slice of the world — a mechanism, a relationship, a tension — and built a simplified representation of it in which one can act and bear the consequences of those actions, even if only symbolic ones.

Let’s start with examples everyone knows.

Hide and seek is a model of a simple, physical skill: hide and find. The rules are minimal, the model is literal — and children practice spatial awareness, alertness, and patience in it.

Chess is a model of armed conflict — simplified to its limits, but faithful in what matters most: each piece has different capabilities, every move has consequences, and the winner is whoever thinks more moves ahead. The game models resource management, anticipation, and reading an opponent’s intentions. Chess in one form or another has been played for fifteen hundred years — and all that time it has taught the same things.

Bridge is a model of acting on incomplete information, and from two sides at once. Each player sees only their own cards and has to reach an understanding with a partner through bidding — a sequence of coded calls from which the partner works out what the other holds, even though saying so outright is against the rules. That models something extraordinarily difficult in ordinary life: acting together without being able to state what one knows. The other side is taking chances. Judging the odds is only the starting point: how the missing cards are likely to lie, what it means that a particular card was never played. But hands differ enormously, and some carry an opportunity that will not come again — and then a sound but cautious game is already a failure, because it wasted what was there to take. One has to sense that this is the moment. If everyone could see everything, both the bidding and the risk would vanish.

Mafia — a social game in which each player is randomly assigned the role of either a citizen or a mafioso — is a completely different kind of model. Here it’s not about strategy or cards, but about observation, bluffing, and social deduction. Who is lying? Who is hiding something in their behavior? How do you keep a straight face when you know more than everyone else? Among Us translates the same model to a screen — and shows that this isn’t limited to any one medium or generation.

These are obvious examples — each models something visible at a glance. But modeling can reach much deeper, into places you wouldn’t expect at first.

The simplest example I’ve written about separately, in my piece on the card game War. There the model is trivially simple — pure chance, zero decisions — and yet the game represents something, and through that representation teaches something. Here I want to show how far this modeling can go.

Phoenix Point is a strategy game about an alien invasion in which humanity is split into three factions, while the player runs an independent organization standing apart from them. The factions cannot handle the aliens on their own, and most people on Earth live within them — too steep a drop in world population ends the game in defeat. Winning is therefore not about defeating the factions but about bringing them through the invasion intact. And left to themselves, they start attacking one another, and that internal war can cost humanity more than the aliens do. The game tracks relations between every party, and every so often it also puts a short decision in front of the player, off to the side of the main game: a description of a situation, a few options, an immediate effect on those relations. Some options improve a faction’s standing with the player’s organization; others open the way for relations between the factions themselves to improve. The model prices those two choices differently than intuition suggests: factions already fighting each other cannot be pulled apart by anything, while the player’s organization is strong enough to survive being resented.

Playing, I realized that such a dilemma existed here at all. Only afterwards did I notice that at the right level of abstraction the same arrangement fits raising children. It comes down to something close to managing relationships: who helps whom in which situation, who can be relied on, who owes whom something, and how that forms a bond between the children that a parent cannot build for them.

Daredragon Fellowship is a board game about a party of daredevils fighting a dragon, or really a model of extreme interdependence and of sacrifice for others — the fight itself is only the space in which that has somewhere to happen. The interdependence is built into what one is not allowed to do alone: cards are never drawn for oneself, one can shield someone else but not oneself, and getting up off the ground also takes another player. The dragon will not be beaten by the strongest daredevil, but by a party that works as one. Sacrifice, in turn, has a measurable price. Whoever takes hits holds fewer cards and so loses options — and yet one can deliberately draw the dragon’s attention, or take a blow that would have landed on someone else. Nobody has an assigned function either: all the characters have identical capabilities, so every such play is somebody’s decision rather than a role being performed.

The last thing happens not in the fight but at the table. Games where everyone plays on the same side have a characteristic problem: at a table where one person is more forceful than the rest, joint deliberation turns into issuing instructions, and everyone else carries out someone else’s ideas. Here talking is entirely unrestricted and the advice may be perfectly sound, but the cards in hand stay face down. Privacy is the tool for exercising one’s own autonomy: as long as nobody sees my cards, the choice is mine, even when someone at the table “knows better”. “I’ll decide what I play” then has the rules behind it, not just the character of whoever says it.

Chants of Sennaar is a game in which the player’s task is to decipher the languages of successive alien civilizations — not by memorizing vocabulary, but from within: by observing context, gestures, and the life of a community, and, to a degree, by taking part in it. Along the way, the game reveals something that is hard to feel firsthand: that language is not a neutral tool for transmitting information, but a reflection of the culture and perspective of the community that speaks it. Immersed in one’s own language, this is hard to notice, because one’s own way of talking about the world simply looks like a description of the world. Meeting a foreign culture from the outside will not reveal it either, as long as one stays at the level of translating words — behind words lie different ways of experiencing reality.

In the end the game asks the player to bring about communication between communities that have lived apart the whole time, each on its own floor of the tower and with its own language. The player is the only translator there — and only then does it become clear that the task is not substituting one word for another, but translating one way of experiencing the world into another. When it works, the communities finally come closer to one another — and their relief carries over to the player, who has watched that isolation up close for the whole game and now sees it coming to an end.

Baba Is You is a puzzle game, or really a model of coming to know a reality. The rules lie right there on the board: phrases like “BABA IS YOU” and “WALL IS STOP” are objects that can be pushed around, and changing a rule is an ordinary move — pushing the words of “WALL IS STOP” apart makes walls stop stopping. That extends to who you are: you can control several things at once, and when nothing is “you” any more, the level is lost.

But that much is visible from the first level. The harder part is what isn’t on the board: there is no tutorial and no manual, so the picture of how this world works is one the player builds alone, out of whatever has worked so far. Every such conclusion puts the world in order, but some of them are traps — not false rules, but generalizations drawn from a few levels and then held as the limits of the possible. The greatest obstacle is then often not the complexity of the puzzle, but one of those assumptions, made without noticing and sitting in the player’s head between them and the solution. It is probably the most abstract thing one can learn through a game.

Why It Works

One might ask: why does learning through a game reach places instruction doesn’t? Not that it’s always better — a textbook conveys facts faster, and a lecture can cover ground no game will. But something happens in a game that doesn’t happen in either.

The answer is built into the definition of a game — each of its four elements does something specific.

Start with voluntariness. A game is played without anyone requiring it. That sounds trivial, but it has a serious consequence: the player is willing to invest effort they would avoid in any other form. Tag and running on a treadmill engage the same muscles and require similar physical exertion. But children run longer, faster, and more willingly in tag — not because they’re being tricked, but because the form of the game makes the effort worth it.

The safe environment works in the opposite direction — it doesn’t raise motivation, but lowers the cost of trying. If consequences are symbolic, you can make mistakes, experiment, try strategies that might not work. In real life, a mistake often costs something — time, money, relationships. In a game it costs only another attempt. This makes players willing to enter situations they would avoid in real life.

Clearly defined rules make the model concrete and keep it from being overwhelming in its complexity. Reality is unmanageable — it has too many variables to draw clean conclusions from. A game selects a fragment, simplifies it, and holds to those simplifications consistently. This lets the player understand dependencies that would be invisible in the full complexity of the world. Cause and effect is legible: I did this, that happened.

And finally — the model of reality. This is, I think, the most important element. A game doesn’t only engage and simplify — it shows why the skills being practiced and the knowledge being acquired exist in the first place. It embeds them in a context that gives them meaning.

Terra Invicta is a sprawling strategy game about repelling an alien invasion, set within a realistically rendered Solar System. Lagrange libration points — locations where the gravitational pull of two large bodies, say the Sun and Earth, balances out so that an object can hold a stable position there without constantly burning fuel — aren’t a curiosity to memorize in this game. They’re locations where it makes sense to place a space station, because physics makes getting there cheap and staying there easy. The game shows practical consequences: why these points matter, both in the game world and the real one. The SOHO probe, observing the Sun from a libration point between the Sun and Earth, is no longer an abstraction — it’s a logical consequence of something I already understand.

Tag works the same way. Running in circles on a treadmill is effort without explanation. In tag, it’s immediately clear why running is useful — to catch someone or get away. The game doesn’t just motivate you to run; it shows you why running matters.

This also explains why the model and the willingness to play so often travel together. A model is not a guarantee of appeal — there are games with a faithful model that get played once, of one’s own free will, and never come off the shelf again. But it is one of the sources of appeal, and the one that produces this particular kind of engagement: the player wants to keep going because something in the model actually depends on what they do.

There is one more thing that crosses the boundary of the model. The situation is pretended, but the emotions that arise inside it are not. Resentment toward someone who allied against me at the worst possible moment is resentment toward that person, not toward their token — and it doesn’t disappear when the box goes back on the shelf.

The rule, though, is games in which the players’ roles are in conflict, and so the emotions are hard ones: resentment, anger, satisfaction at someone else’s loss. Learning to bear such emotions and come out of them whole is worth something, and a game offers safer conditions for it than life does. But the intensity varies. A game that models the essence of injustice — at a concentration one does not meet day to day — may for that reason disqualify itself for a more sensitive person. These are not two different properties of a game but one. Practising how to bear difficult emotions is possible because those emotions are real — and that is exactly why a game can be unbearable for someone.

More rarely, someone reaches for this mechanism deliberately and in the opposite direction — building mechanics so that they produce situations of mutual aid. That is closer to a therapeutic effect than to teaching. Either way, the rules of a game determine what may be done inside the model; they have no authority over what remains afterwards between people.

Why “Educational” Games Often Don’t Work

If games have such educational potential, a question follows naturally: why do games that explicitly call themselves educational so often fail? Why is the word “educational” in a game’s name so often a warning sign for children rather than an invitation?

Wyldfyre from Ninjago: Dragons Rising only knew games like that. “Why would Jay be so crazy about a video game? They’re super boring. All that grammar and spelling. […] «Carrying the one.» «Using proper manners.» Blech!”

The answer is simple, though rarely stated outright: most educational games aren’t games in the sense I’ve described. They’re exercises in disguise. I’ll go on calling them games, because that’s what the market calls them — but that’s a genre label, not a qualification.

The classic example is games in which tasks — arithmetic, geography questions, grammar fill-in-the-blanks — fall from the top of the screen, and the player must solve them before they reach the bottom. The mechanic is simple and engaging. But the connection between the mechanic and the content is purely external: you could replace arithmetic with any other task and the game would work exactly the same. Addition isn’t part of any model — it’s just a gate to the next step.

The same applies to an entire genre of board game quiz-racers, where answering a factual question unlocks a move on the track. The question about the capital of France changes nothing in the game world — it has no effect on any dependency, no consequences beyond “you may advance.” The factual layer and the mechanical layer exist side by side, never intersecting.

The model is missing. And without a model there’s no meaning — only exercise. And that is where the boredom comes from: a child rejects such a game not because someone forgot to make it appealing, but because they sense that nothing in it really depends on what they do. Boredom is the symptom; the missing model is the cause.

This is the exact opposite of what makes games teach. My son learned to add before he started school — not because he had a game that taught him addition, but because he had Karak. It’s a game about an expedition into a dungeon, where the hero faces one monster after another: you roll two dice, add the points from the weapons and scrolls you’ve found, and compare the total against the monster’s strength. A higher total wins; a lower one costs a life point. He added because winning the fight mattered to him, not because a correct answer earned him a reward.

This is where it’s worth naming the phenomenon behind most failed educational games: gamification, shallowly understood. The idea is simple — since children enjoy playing, maybe it’s enough to wrap learning in game mechanics? Add points, levels, badges, progress bars? The problem is that it’s not game mechanics that make games teach — it’s the model of reality the game contains. Sticking points onto an exercise doesn’t create a model. It creates an exercise with points.

Wyldfyre, who until then had known nothing but exercises in disguise, found this out by accident in an arcade, at a machine where you shot at meteors. “What? This has nothing to do with bedroom organization. There’s video games where you can blast stuff? Wait… can video games be exciting?!”

Teaching Is Not Endorsement

If a game teaches through its model, then it teaches whatever it is a model of. That sounds like a truism, but it has an uncomfortable consequence: the mechanism described above works the same way regardless of what the model represents. A game in which treating people instrumentally is the effective strategy teaches that just as efficiently as bridge teaches cooperation. It isn’t the case that good games teach and bad ones don’t — bad ones teach too, and that is precisely the problem.

Sometimes the model is only a wrapper. The game has rules, a world, decisions — but those aren’t its real mechanism. Underneath works a reward loop tuned to keep you playing, or the pressure to purchase, or scarcity manufactured by the publisher and sold as value. Such a game teaches something too. Just not necessarily what’s printed on the box.

This is why “games teach” is not praise for games, any more than “books shape people” is praise for all books. It’s a description of a mechanism, not a recommendation. Whether a game is worth the time — yours or a child’s — is a separate question, and it needs a separate answer: what does this particular game teach, and do I want it teaching that?

Games as a Medium

Games teach — which, as we’ve seen, is not the same as saying games are good. This isn’t an argument against games. It’s an argument for taking them seriously, the way we take other media seriously.

Because games are a medium. Or rather — a family of media. The definition I’ve been working with covers very different things: video games and board games, sports and tabletop RPGs, card games and children’s street games. What unites them is interactivity — the player doesn’t observe, they act. And that’s precisely what their static counterparts don’t offer: a film can be watched, a series can be followed, a book can be read. But a game has to be played.

That difference has consequences. A game can teach physical skill, social competence — from handling defeat to thinking in terms of shared interest — strategic thinking, and attitudes toward rules and values, and along the way, knowledge from fields the player didn’t know they were interested in. It can also teach things we would rather it didn’t — by the same mechanism. Not every game teaches everything. But every game that is a real game — a model you take part in, with clear rules, a safe space for trying things out, and undertaken voluntarily — teaches something. Not always what we would have wanted — but always something.

Which is why the category in the title is empty. An “educational game” is either just a game — because every game teaches — or an exercise pretending to be one. The name promises something in between. There is no such thing.

written August 2026