Mario levels teach by changing one tile

Super Mario Maker, Nintendo, 2015, did not invent level design. It made the Nintendo lessons playable from the other side of the screen. Seesaw Bridge in Super Mario Bros. does not open with a lecture about danger. It puts a bridge under you, then a second one, then the thing that can knock you off. The danger arrives as a change to a shape you already understand.

That is the whole course. Super Mario Bros. 3 can spend a level on a joke. Super Mario World can bring an old idea back with one new twist. A finale works because it spends ideas you have already paid for, not because it invents ten more. People who open an editor and place every enemy they like are skipping the class. One mutation, then another, is the assignment.

An editor is the cheapest design school there is. LittleBigPlanet, Trials, mod tools, the level editors bundled with older games, and Super Mario Maker all hand you the act of designing without asking you to learn an engine first.

Super Mario Maker has an unusual advantage among them. Its parts are not neutral. Goombas, pipes, coins and platforms arrive attached to decades of stages that used them well, so the editor ships with a reference library.

What it does not ship with is a starting point. Familiar parts do not tell you where the first block goes. The way in is to take real Mario levels apart and watch how they introduce an idea, press on it, surprise you with it, and pay you for surviving it.

A good Mario level lets you meet its mechanic before it can hurt you. Seesaw Bridge, in New Super Mario Bros. U, opens with a Wiggler that barely counts as a threat — a free win, handed over before the real subject of the level shows up.

The seesaw bridges follow, tipping up and down where tipping costs nothing. The platform never dips far enough to reach the poisonous water below, so you can learn the motion without being punished for the lesson.

Only then does the level ask for something. A power-up, then a wider jump from one seesaw to the next, with coins arcing through the air along the line your jump should take. Nobody explained timing to you. The arrangement did it.

An enemy does not have to be a hazard. The Koopa Troopa near the start of Seesaw Bridge is placed as a prop: kick the shell to the right, as most experienced players will, and the layout sends it clean through a row of Wigglers.

Worth holding on to, because it widens what enemies are for. Some of them are obstacles. Others are toys, or prompts, or the punchline the level has been quietly arranging.

The same stage keeps a barely hidden pipe that drops into a hard optional coin challenge. That challenge leans on ideas the player has not properly met yet, which is what makes it veteran territory rather than part of the main path.

Escalation in Mario is a series of small changes to one thing, not a stream of new things. After the safe introduction, Seesaw Bridge widens the gaps and lets the platform dip into the poison. Enemies arrive between platforms to intercept you, so jumping and landing now come with timing attached.

Then Wigglers appear on the bridges themselves and the same challenge has to be cleared by bouncing off them. Later the bridges pivot from a single point and swing the whole way around, their cycles deliberately out of sync so waiting for a clean alignment is not an option.

At no point does the level drop its idea and reach for a different one. It keeps the idea and complicates it, one turn at a time.

Texture comes from choices the level does not force. Mario stages dangle something optional and expensive: a block with a 1UP in it, reachable only by jumping at the worst possible moment. Walk past it and stay alive, or gamble your progress on the reward.

Red coin rings do a different job. Mario plays best fast, but staying alive rewards caution, and a timed coin sequence resolves that tension by shoving you into speed-run mode for a few seconds. A careful level gets one short burst of momentum out of it.

Seesaw Bridge also stretches the space between safe ground. Early on, somewhere to stand is never far away. Later you are asked to keep moving, to react to hazards before you can see where you will land, and to hold it together for longer before the next rest.

A level can also be built to land a gag. Stage 4-6 of Giant Land, in Super Mario Bros. 3, is the clearest example of it.

Magic doors let you flip the stage between giant enemies and normal-sized ones, but the real moment is the Starman. You grab it, you start charging forward the way anyone would, and invisible blocks materialise overhead at precisely the wrong instant.

The payload is not difficulty. It is a joke shared between designer and player: the level knew exactly what an invincible Mario wants to do, and it wrote the interruption in advance.

Everything around that moment is ordinary by comparison, though it still demonstrates how little an enemy needs to do. Koopas are no trouble to beat, yet their predictable patrols put a clock on jumps that would otherwise be free.

Reusing an idea works when the second visit is harder than the first. Chocolate Island 3, in Super Mario World, brings back a rotating platform the game introduced earlier and gives it a single arm, which makes the movement considerably fussier.

The opening is safe, as usual: one platform between two pieces of solid ground, no enemies, room to work out what the arm does. Then it stacks — a Koopa on the pivot, three platforms between safe zones, uneven gaps, rotations that refuse to line up.

The optional choices keep coming too. Deal with a Koopa to reach a block or leave it alone. A pipe leads somewhere genuinely worth going, offering a playful mini-game rather than just coins, and carries a cost the level never mentions: that route skips the checkpoint.

A finale is where a level stops adding and starts combining. Chocolate Island 3 mutates its idea a few more times first. Fuzzies take over the central platforms, so the safe middle is no longer safe. Some enemies force you to duck as the arm swings low, and interceptors turn up between jumps.

Then it plays several of those pressures together — a run of rotating platforms, some deliberately out of phase, enemies riding the pivots, collectibles for anyone greedy enough, and a flying enemy positioned to break your rhythm.

The last beat is one more echo of the mechanic, and bookending like that is what makes a stage feel finished rather than merely over. The ending is a closing statement about the idea you have spent the level learning.

The first game is plainer than the ones that followed it, largely because it was busy inventing the vocabulary they would inherit. Stage 8-2 still has things worth taking.

It puts Lakitu straight back in front of you, on a broken staircase where the footing is bad, then hands skilled players the chance to kill the thing that used to torment them. Power over a familiar threat is its own reward.

A spring that fires out a 1UP starts a short chase. You sprint forward with one eye on the mushroom above and one on the hazards below. The stage gets another quick surge of speed out of a single object.

It also marks a boundary. One late section amounts to a single enormous jump — hard, but not interesting. Stretching a gap only buys so much before the challenge tips over into unfair and dull at the same time.

Put those stages side by side and the rule is consistent: one idea, occasionally two, iterated until it is difficult. Very few Mario levels are built on more than that.

The iteration has a vocabulary of its own. Widen the gaps. Tighten the timing. Push the safe zones further apart. Drop an enemy in as an interceptor, or place one directly on the path. Change the core mechanic slightly. Take away a safety net the player has grown used to.

The trimmings are just as repeatable. Coins steer the player through a route. Optional rewards make risk tempting. Timed events buy a burst of speed. A moment of pure fun pays off a hard section, or clears the palate before the next one.

Begin a first Super Mario Maker stage by picking one mechanic out of the toolbox and asking how many ways you can iterate on it. The editor will not let you invent an entirely new gimmick, but it has more than enough parts for variations on an existing one.

Platforms on tracks make a good candidate. Show them off in a safe stretch first, with a net underneath so a mistake costs nothing, then hand over a mushroom and hint at what is coming next.

The first real iteration can be a jump between two moving platforms with a Koopa Paratroopa cutting across the gap. The section after that can put three platforms in play and offer a fork: the easy line, or a riskier one that threads through spikes.

Let the risky line become a small optional gauntlet — bullet bills, falling platforms, spikes, and a 1UP chase borrowed from the original Super Mario Bros. Move fast and the extra life is yours. Hesitate and the mushroom drops somewhere you cannot follow.

Some of the best moments in a level are not authored, and keeping them is a legitimate design decision. A Goomba that bounces off a music block by pure chance can be charming enough to leave exactly where it is.

The test is whether the accident belongs. A rare, low-probability interaction adds personality as long as it neither breaks the stage nor distracts from the idea the stage is about.

Later sections can keep adding: fire hazards, a safe vantage point to watch them from, an escape route, a question block placed to tempt the player somewhere dangerous. Running a track downward before the hazard buys extra seconds to read the pattern before answering it.

Bring an early element back at the end and the stage reads as composed. A late saw blade can rhyme with the opening, and a note block, familiar by then, can be the answer to clearing it. Coins hang out of reach as a tease, while the real prize is kicking a shell down a flight of steps to wipe out enemies and bank extra lives.

Make the final choice the same shape as every choice before it: ride the easy platform to the flag, or take the saw-blade route and be paid for the nerve.

Small as that structure is, it covers the ground. A main path, optional risks, an idea that recurs, escalation, a joke, rewards, and a last decision that echoes all of it.

Expect a first level built this way to have flaws. It may sit slightly too easy. One jump may want a tighter window. A chase may turn out to be almost impossible to fail. An enemy may crawl where you assumed it would rush.

None of that matters much, because the point was never a flawless stage. The point is that reading levels turned into building them. You only find out which parts land by loading the thing up and playing it yourself, and your own verdict is the one that counts.

So take a level apart. Name its core idea. Track how it introduces that idea safely, mutates it, adds risk, makes room for fun, and closes on an echo. Then build something short that arranges the same principles a different way.

In the Flockbay app, ask for three short platforming rooms that share a layout. The second may change one thing. The third may change that thing again. Play them in order. If room three needs a sentence of explanation, you added a mechanic instead of a mutation.

Use a 2D platformer template and the AI platformer maker. Keep the jump you already trust.