Teach the system after it matters

A complex game has a gap at the front, and that gap is where people quit. A grand strategy, a real-time strategy, a city builder: each one opens with a long stretch of instruction that lands before the player knows why any of it is worth the effort. The obvious fix is a shorter tutorial. The better fix is a later one.

Doing a thing once is not the same as learning it. A campaign that adds one system when the previous system has already become a problem, and an interface that appears a piece at a time, teaches more than a patient opponent who lets you practice everything on day one. Frame the lesson as a small problem whose answer is the new rule. If the player can ignore the rule and still succeed, you have not taught it. You have demonstrated it.

A complex game has to be learned before it can be enjoyed, and that gap is where players quit. Grand strategy, RTS, city builders and their neighbors open with a long stretch of instruction that lands before anyone knows why the game deserves the effort.

The frustrating part is that teaching has improved almost everywhere else. Lessons get folded into the world and the story, paced with care, kept quiet, sometimes hidden well enough that nobody notices them.

Deep games have not caught up. Their opening hours can be genuinely miserable, and at the extreme the tutorial becomes the reason a player never reaches the game underneath it.

So the useful question is whether the tricks that work elsewhere transfer. Can a complex game be taught the way a puzzle game is taught?

More on this: AI RTS maker for a genre with a hundred moving parts

The biggest shift in tutorial design is the recognition that teaching need not be a block of time before the game starts. It can be distributed.

Scatter the lessons through the experience and each one lands on a player with more reason to care than they had at the title screen. Willingness to learn rises with investment, and a wall of information at minute one asks for attention nobody has agreed to give.

Spreading lessons out also gets the player into the real game faster. Keep each piece small enough and the tutorial stops feeling like a tutorial at all. The opening chambers of Portal introduce one idea at a time, and the player experiences it as a run of puzzles rather than a lesson plan.

The bigger win is timing. A lesson taught at the moment it is needed sticks. Crafting makes sense when the player is standing in front of a crafting table, not an hour earlier when the words have nothing to attach to.

Complex games frontload because every system looks load-bearing from turn one. Economy, diplomacy, units, cities, resources and enemies are wired into each other, so none of them seems able to wait.

Some of these games already contain the ramp they need. Civilization is built on what Bruce Shelley called the inverted pyramid of decision making. Turn one may hold a single choice: where to settle. Then two. Eventually dozens, spread across units, cities, allies, enemies and resources.

That growth, from one settler on a fogged map to an empire, means Civilization is unusually well positioned to hand out its systems in installments rather than all at once.

Frostpunk shows the method. The opening job is small: collect resources by hand, and the game teaches that. The generator comes online, and the game teaches that. Each mechanic gets its own short lesson as it becomes relevant, without interrupting the campaign already underway.

For most newcomers the interface is the intimidating part, so it can be introduced in stages rather than all at once. Buttons, bars, panels and windows crowd the screen long before any of them mean anything.

The question worth asking during design is how much of that screen genuinely has to exist on the first turn.

Mini Metro opens with almost nothing on screen. Connect the stations. That is the whole job. The line menu, the clock and the passenger counter show up later, each one when it starts to matter.

Animal Crossing: New Horizons treats the tool wheel the same way. It is useful, but it does not arrive next to the basic inventory. The player buys it when they want it, which puts the pace of UI complexity in their hands.

A few complex games work this way already, such as Total War: Troy, which expands its interface during play. There is room to push much harder and keep whole systems off screen until a player has a reason to open them.

Games built for repeat play can spend their teaching across whole campaigns rather than across turns. A strategy game is replayed, not completed once like a story adventure, which opens a longer runway.

Fighting games make the comparison easy. Nobody needs frame data, matchup theory and specialist vocabulary before a first match. They need enough to stop mashing buttons and start choosing them.

Mortal Kombat 11 splits its tutorial into brackets: basics, advanced, strategy. They are not meant to be swallowed in one sitting. After each segment the game pushes the player back into matches to use what they learned and return when they want the next tier.

Strategy games could borrow that shape directly. Let the first campaign teach the core and nothing else. Let the second add systems. Keep going until the player is holding the full game.

Civilization V demonstrated this, probably by accident. It cut or simplified systems earlier entries had, espionage and religion among them. Longtime fans were annoyed, but it made a far gentler place to start for a first Civilization.

The expansions put those systems back. By then players had the fundamentals, so absorbing the returning pieces cost much less effort than learning them cold would have.

That points at a different definition of easy mode. Rather than the same game with a dumber opponent, the easy version could be the game with fewer systems switched on. The next run turns more on. Later runs keep building until the player arrives at the complete design.

None of this is cheap. A game would have to be planned around it from the start. Done well, it hands players complexity exactly as their investment grows to meet it.

Where instruction genuinely has to come first, it can at least be made active rather than passive.

Action games teach by making the player perform the move. The door stays shut until enemies are punched, the gap is jumped, the new ability is used. That is kinesthetic learning, and doing retains better than reading.

Complex games reach for the same idea and often get a hollow version of it. Arrows point at the interface. Click here. Click there. Select this. Drag that. The hands are busy and the brain is not.

Following instructions is not learning. The fix is to turn each lesson into a small problem the player has to actually solve.

A tutorial teaches more when it asks for a solution instead of a keystroke. The mobile puzzle game Threes makes the point. Rather than instructing the player to swipe left twice and then up twice, it asks them to use the walls to shuffle the numbers and bring a 1 and a 2 together.

The task is still tiny. The difference is that it demands the same kind of thinking the full game will demand later.

Deeper games can work the same way. Planet Zoo walks the player through fixing the welfare of a single animal, then hands them the whole zoo and asks them to raise welfare across it with far less hand-holding.

Now the lesson has to be applied. Inspect the animals, find the problems, fix them. Offworld Trading Company does something similar, setting objectives and leaving the method open. Often deleting the arrow that says click here is enough to make a tutorial feel like the game.

Learning by doing runs into a specific obstacle in complex games, which is that the results take a long time to appear.

Miss a jump and you know instantly. Make a poor economic call in a strategy game and the consequence may stay invisible for hours. Sometimes an entire campaign has to finish before the shape of a decision becomes clear.

That is a strong argument for short, repeatable campaigns. A run that ends quickly can be run again, and the second attempt is where the lesson lands.

There are other levers. A quick campaign mode can be presented as a training ground. Advisors can flag a choice while it is still being made. When Offworld Trading Company warns a player they are selling aluminum too cheaply, that warning teaches how the market behaves.

Tutorials in this genre are usually good at how and poor at why. Players finish them knowing the controls and not the strategy.

Pointing at the button that constructs a building in Civilization is trivial. What is worth building, when, and where it should go once it exists is the hard part, and the part that decides whether someone plays well.

Shortening the feedback loop helps here too, since consequences arrive while the decision is fresh. Advisors can carry more of the load, recommending and cautioning, telling a player to stop producing workers and spend those turns on something better.

The target is not a player who can operate the interface. It is a player who has started making strategic decisions of their own.

Anything a player already understands is something the tutorial does not have to teach. Spikes hurt. Ice is slippery. Coins buy things. Keys open locks. Skulls mean danger. Plants vs. Zombies never explains that a zombie in a metal bucket takes more punishment than one in a plastic cone, because everyone knows what those materials are worth.

Grounded and historical settings give complex games the same discount. Civilization gains a great deal from the fact that players walk in with some sense of history, technology, geography and politics.

The interface may be the richest source. People spend their days operating software, and games can borrow that fluency. Reigns takes its left-right swipe from dating apps. Disco Elysium lays out dialogue like a social feed.

Planet Zoo benefits when its lists sort the way website tables sort. When map icons look like the location buttons players use elsewhere. When red and green mean bad and good, and the filter icon matches the one in a spreadsheet. The interface reads as familiar before anyone explains it, so the player can get on with the problem.

Familiarity cuts both ways, and it fails when a designer assumes knowledge the audience does not have.

Playtests for Total War: Troy turned up players who could not find the button that ends a turn. One spent 40 minutes stuck on the opening turn, unsure how to move forward. The button carried an hourglass, which reads as end turn to a strategy veteran and as loading to almost everyone else.

Before release the studio swapped the hourglass for an arrow.

Two things follow from that. Do not assume the audience arrived fluent in the conventions of the genre. And playtest the tutorial harder than anything else, because its failures live in details that anyone experienced has stopped being able to see. Watching someone play the actual build is the only reliable way to find them.

Show, don't tell applies to teaching as much as to storytelling. A wall of text is hard to absorb mid-play, and one image or short animation often delivers the same rule faster.

Into the Breach does this with weapon previews. Rather than writing that a weapon damages one tile and shoves the tiles beside it, the interface plays a small animation of exactly that. Watching beats reading.

Text is not going away in a game this deep. It can still be trimmed hard, kept consistent in terminology, stripped of jargon, and checked so the load-bearing sentence is not buried under flavor dialogue.

The point is not to strip out explanation. It is to make explanation cheap to process at the moment a player is already busy learning a system.

Players forget, and they need somewhere to go when they do. Tooltips, tooltips inside tooltips, an encyclopedia, a glossary, tutorial steps that can be replayed: all of it earns its place. Quitting to a search engine should never be the only route back to a concept.

Good reference material also buys the designer freedom. When the information is reachable on demand, the tutorial can ask more of the player and trust that nobody ends up stranded.

People also learn differently from one another. Offworld Trading Company offers both scripted tutorials and practice challenges. Total War frequently separates the tutorial for a complete newcomer from the one for a returning player, which covers only what has changed.

The tutorial usually gets built last and valued least, which is backwards. For a complex game it is one of the things success actually depends on.

Teaching newcomers is how a deep franchise grows. Without it, the audience narrows to the experts who already know how to play.

The tools exist. Break the teaching across a campaign, or across several. Let players learn by doing rather than reading. Borrow interfaces and metaphors they already understand. Give them fast feedback, advisors worth hearing, visual explanations, and reference material when they need it.

Things are moving in the right direction, with more thoughtful teaching and more legible interfaces than the genre used to have. But while a curious player can still zone out halfway through the opening lesson, work remains.

In the Flockbay app, ask for a game with three rules and a first level that uses only one of them. The second level should break if the player never learns the second rule. Play the first level and notice what you still do not know. That ignorance is the tutorial you have not written yet.

A city builder makes the temptation to explain everything up front very strong. A strategy game maker is the same trap with units.