Lesson 5 · Part 2: The Player
Just Hard Enough
How games keep players at the edge of their ability, and why well-designed failure is most of what a game is made of.
Just Hard Enough
It’s nearly one in the morning, and you’ve just failed for the eleventh time. Maybe it’s a jigsaw with a sky made of four hundred blue pieces, or a crossword with one corner that won’t crack. You meant to stop an hour ago. But every attempt got you a little further, and every time you could say exactly what went wrong. Your hand is already reaching for the next try.
Written down, that hour sounds miserable: eleven failures in a row, and nobody made you do it. From the inside, it was the best part of your day. This lesson is about why, and about how designers make it happen on purpose.
The edge of what you can do
Think of Tetris, the puzzle game where blocks fall and you fit them into rows. At the start the blocks drift down slowly and you have all the time in the world. Later they fall so fast you’re placing them on instinct. Somewhere in between there’s a speed where you’re just about keeping up, and that’s where an hour goes by without you noticing.
Psychologists call that state flow: being absorbed in something right at the edge of what you can do. Flow has a wall on each side. Make the task too easy and the player can see everything coming, and they get bored. Make it too hard and they can’t connect what they did with what happened. The results start to feel random, and they get frustrated. Between the two walls is a narrow channel where the outcome is in doubt, and it’s up to the player to settle it.
Racing drivers have a rule of thumb about this: if you never crash, you aren’t trying hard enough. It isn’t science, but it carries over to games. A player who never fails is nowhere near their edge.
Most of playing is failing
Follow that through and you reach an odd fact. A player in flow is failing regularly. In most games, people spend most of their time failing at something, so a game is only as good as its failures.
Celeste is a platform game about climbing a mountain, and you’ll fall off it hundreds of times. Each fall puts you back at the start of the same screen within a second, and you go again. The game even counts your deaths and tells you to be proud of the number, because the more you die, the more you’re learning.
“You need to design the game such that failure is interesting and understandable.”
An interesting failure is one the player can read. They can say why it happened, they can see it was their own doing, and they come away with something to try next: I jumped too early, so next time I’ll wait. The failure has fixed a small part of the mental model they’re building, and the next attempt tests the better version.
The bad kind comes in two flavors. In the first, the player loses and has no idea why. In the second, they know exactly what to do, but their hands can’t do it in time, and by the fortieth try they’re learning nothing. That’s where many people put a game down for good.
Getting good this way has a name, failure-based learning: you improve by failing often, cheaply, and on purpose. It’s what Level 1 said play was for. Most of us were taught to avoid failure, and a game is one of the few places where it costs almost nothing.
One more tool belongs here: humor. If the same failure plays out the same way every time, by the tenth go it feels like being taunted. Vary it, and make some versions funny, and the player is at least getting a laugh. Fall Guys, a game show where wobbly jelly-bean contestants race through obstacle courses, is mostly people falling over, and the falling is the fun part. Super Meat Boy, a very hard platform game, ends each level with a replay of every attempt you made at once, dozens of little Meat Boys flying into saw blades together. Humor is cheap, and it fixes problems that clever engineering can’t.
Loops inside loops
Failing and succeeding don’t happen at only one size. They come in loops, and the general name is the game loop: a cycle of trying something, seeing what happens, and trying again. Once you’ve got the hang of one loop, it hands you on to a bigger one.
Stardew Valley, a farming game, stacks them neatly. The smallest loop takes seconds: walk to a plant, water it. The next takes a day in the game: plant, tend, harvest, sell. Above that is the season, where the money from many days buys a bigger barn or a new tool. Each loop has its own success and failure, and you can’t work on the big one until the small ones are running.
Every loop a player closes adds a layer to their picture of the game, and the layers stack in the same order as the loops. When you design, you’re choosing what those layers are and the order they arrive in.
Ramp it up
The order of those layers matters, because nobody can hold a whole game in their head at once. Hand a beginner every rule in the first minute and they drown. The designer’s job is to start the player with a small, simple model and add to it a piece at a time. That’s the complexity ramp.
Portal, a puzzle game about a gun that makes linked doorways, does this beautifully. Each early test room teaches exactly one new idea, and the next room assumes you’ve got it. Plants vs. Zombies, where you plant a garden to hold off a lawn full of zombies, gives you one new plant per level, so you’ve always just learned the newest piece when it starts to matter.
The difficulty curve is the shape of how hard a game is at each point along the way. Good curves rise as the player gets better. What you’re watching for is a cliff: a spot where the slope suddenly becomes a wall and players pile up at the bottom. Cliffs are hard to spot from a desk, so teams find them by measuring where testers get stuck, with metrics (Level 18).
Who holds the dial
Players arrive with different skills, so no single difficulty suits everyone. There are two broad ways to fit the challenge to the person.
With dynamic difficulty adjustment, the game watches how you’re doing and quietly eases off when you’re struggling, or pushes harder when you’re coasting. Mario Kart, the go-kart racing game, gives players near the back better items than the ones in front, so races stay close. In Crash Bandicoot, a platform game about a cartoon bandicoot, dying again and again at the same spot quietly gets you a little extra protection on the next try.
With player-paced difficulty, the player holds the dial. The game offers a safe route and a tempting harder one with a bigger payoff, and players sort themselves. Celeste hides strawberries along its hardest paths. They do nothing except prove you got them, and plenty of players go after every one while others never bother.
Player pacing has a lovely bonus: it keeps agency with the player. The person who chose the hard road owns what happens on it, and can step back onto the easy one when they’ve had enough.
Try it · Pocket Game, step 5Loops and losing
Play your Pocket Game and time three things. What does a new player have to get the hang of in the first ten seconds? What do they learn in the first minute? What takes five minutes to pull off? Write one line for each. If one is missing, add something to fill it. Then look at how a round is lost. Change the rules or the pieces so that losing either shows the loser what to do differently, or makes both players laugh. Knocking over a tower counts.
Carry forward: Your players will spend most of their time failing, so design the failures with as much care as the wins.
Check
Easier, harder, or leave it?
You’re the designer, watching people play. For each one, decide: make it easier, make it harder, or leave it alone.