What makes a browser game worth five minutes

A browser game competes with everything. Not with other games, but with the eleven tabs already open, the message that just arrived, and the fact that closing a tab costs nothing at all. On Steam somebody has paid, downloaded and committed before your title screen appears. On the web they have committed to nothing and they will prove it.

I have watched a few thousand launches on this site now, and the games that hold people share a small set of properties. None of them are about how the game looks.

The first four seconds decide it

Not the first minute. Four seconds is roughly how long a stranger will look at a blank canvas before deciding your game is broken, and it is a wildly underestimated number by everyone who has only ever loaded their game from localhost.

The games that survive it show something immediately. Not a loading bar. A thing. A title drawn in code, a character idling, a board already half-populated with a demo playing itself. The best version of this is a game that has already started: you arrive and something is moving, and the instruction to press a key is the second thing you notice, not the first.

The corollary is that a 6MB engine payload behind a white screen is a design decision, not a technical one. It is fine to be big. It is not fine to be blank.

It has to survive being described badly

Somebody is going to send your game to a friend with the words "this one is weird, you place tiles but they steal from a shared pool". That sentence is your actual marketing, and it will be written by someone who played for four minutes and got it slightly wrong.

Games that spread well have a rule that survives a bad description. Games that do not spread are usually the ones where the interesting part is an accumulation: the third hour is great, the mechanics interlock beautifully, and there is no sentence that conveys any of it.

This is not an argument for shallow games. It is an argument for putting the strange bit up front. If your game becomes interesting at minute twelve, move that to minute one and let the rest be the elaboration.

Failure has to be cheap

Almost every browser game that people bounce off has the same bug, and it is not a bug: the restart is slow. Three seconds of death animation, a results screen, a menu, a fade. In a game somebody paid for, that is pacing. In a tab, it is an exit ramp.

The games at the top of the leaderboard here mostly restart in under half a second, and several restart instantly on the same key that killed you. The player never leaves the loop long enough to remember that email.

This extends past restarts. Undo in a puzzle game, a rewind in a platformer, a forgiving checkpoint. Anything that lowers the cost of being bad at the game buys you another five minutes. Difficulty is not the enemy. Ceremony around failure is.

There has to be a stopping point

The strangest finding from running this site is that games with an ending get more replays than games without one, which is exactly backwards from what an engagement mindset would predict.

Endless games ask you to decide when to stop, and that decision is a small unpleasantness that people learn to avoid by not starting. A game with a visible finish (twelve levels, five minutes, one board, a score you can beat) gives you permission to begin. And having finished, people frequently go straight back in, because now they know how long it takes.

You do not need a story. A results screen that says you did this in four minutes and seventeen seconds is an ending. So is a level counter that stops.

What five minutes is not

It is not a genre and it is not a size limit. Some of the most-played games on this site are twenty-hour idle games and some are one screen. The five minutes is a threshold, not a target: it is the window in which a game either earns a longer session or does not get one.

What it definitely is not is a smaller version of a big game. A browser game trimmed down from a bigger design usually keeps the parts that needed the bigger design (the tutorial, the progression, the menu tree) and loses the parts that made it worth it. The games that work here were designed for the tab.