How Difficulty Progression Works in Arcade Games

An arcade game has one axis for difficulty to rise along: the current run. There are no levels to unlock and no gear to acquire, so everything the game does to stay interesting has to happen in the next ninety seconds. This guide is about how that is done well, and about the one line a fair game never crosses.

The line a fair game never crosses

There are two ways to make an arcade game harder. You can add threats, or you can take away the player's ability to respond to them. Only one of those is legitimate.

Adding threats means more enemies, faster movement, tighter gaps, denser layouts. The player's job gets bigger, but every individual problem remains solvable with the skills they already have.

Removing reaction time means shortening telegraphs, hiding tells, or spawning things on top of the player. The individual problem becomes unsolvable, and players correctly experience this as the game cheating.

On this site the line is drawn in three places. Obstacle spacing in the runners is expressed in seconds of travel rather than pixels, with a floor of 0.62 seconds, so speed never eats the reaction window. Boss telegraphs shorten as a fight goes on but stop at 0.55 seconds. And a wave's creatures walk in from the edge on staggered delays rather than appearing next to you.

That last constraint is worth dwelling on. A creature that materialises inside you is not difficulty, it is a dice roll, and a player cannot get better at dice.

Why the curve is eased rather than linear

Difficulty on this site is stored as a value from zero to one and every game interpolates its own numbers along it - speeds, spawn rates, gap widths, window sizes. What matters is the shape of the interpolation.

A linear ramp is the obvious choice and the wrong one. It makes the first thirty seconds boring, because the game is at ten per cent of its interesting range, and it makes the last thirty impossible, because it never stops climbing.

The curve used here is a smoothstep: gentle at the start, steepest through the middle, and flattening at the top. That produces three distinct phases that players can actually feel - a warm-up where you settle in, a climb where the game gets genuinely harder, and a plateau where the difficulty stops rising and the question becomes whether you can hold it.

The plateau is the important part. A game that never stops getting harder has no skill ceiling, only an endurance ceiling, and players stop caring about the difference between a good run and a lucky one.

The breather, and why it is not generosity

In the wave-based games here, every fourth wave is deliberately smaller, with treasure scattered on the floor.

That is not the game going easy on you. A difficulty curve that only rises reads as a wall, and players stop at walls - they do not experience a wall as a challenge to overcome but as a signal that the game is over. A staircase with landings is the same total challenge arranged so that it can be beaten.

The landings also serve a mechanical purpose. They are when you collect the treasure that is worth points, reposition to a better part of the arena, and let a stamina bar refill. A game with no landings has no moments where a player can prepare, which means every wave is entered in whatever state the last one left you in.

Recognising the breather is a skill in itself. When a wave-based game suddenly feels easy, the correct response is not to relax but to spend the wave doing the things you cannot do during a hard one.

Introducing one thing at a time

The other half of fair progression is what arrives and when.

In the runners, a new hazard type is always emitted alone the first time, with double the usual clear space on the far side, and in several games its name is printed on screen. You meet it in isolation, with time to look at what just happened to you.

In the wave games the rule is that a wave never opens with a creature you have not met, and new types are unlocked one wave at a time. You are never learning a behaviour while already under pressure from it.

There is a second composition rule in the battle games: never more than one ranged creature per two melee ones. A player forced to dodge projectiles while closing distance with several melee attackers has no route that is not a guess.

These sound like small courtesies. They are the difference between a game that feels hard and a game that feels arbitrary, and players can tell instantly which one they are playing even if they cannot say why.

Difficulty that responds to the player - and why it is not used here

Some games adjust difficulty based on how well you are doing. It sounds fair and it usually is not.

The problem is that it makes the score meaningless. If the game gets easier when you are struggling, then a high score partly measures how much help you were given, and two runs with the same number are not comparable.

It also produces a perverse incentive that players find surprisingly quickly: playing badly on purpose early to be given an easier late game.

So difficulty on this site advances with time or distance on a fixed curve, identically whether you are playing brilliantly or badly. The only thing that responds to your performance is one exception worth knowing: in Dragon Treasure the dragon unlocks new attacks when you bank rather than as time passes, so a cautious player meets the fourth attack later - which is a reward for caution rather than compensation for failure.

How difficulty is expressed in each category

The same curve, applied to different variables.

• Runners: speed rises, obstacle spacing falls - but never below the reaction floor in seconds.
• Platformers: gaps widen, platforms narrow, awkward platform types unlock at set heights.
• Collection games: quotas rise and clocks shorten, while the layouts stay readable.
• Battle games: wave size, creature tier and composition, with a breather every fourth.
• Reflex games: windows narrow and speeds rise, with the window size often shown in milliseconds.
• Adventure games: more traps and a further exit, rather than faster traps.

That last one is a deliberate choice. Making traps faster in an adventure game would push it towards being a reflex game; making the space bigger and busier keeps it a routing problem, which is what the category is for.

What to do as the difficulty rises

Three habits that hold across every category.

Simplify as things speed up. The instinct is to do more when there is more happening, and the correct response is usually to do less - take the safe line, protect the multiplier, stop chasing the marginal pickup.

Recalibrate rather than react faster. When a runner speeds up, the answer is not quicker reflexes; it is looking further right. The information you need is arriving at the same rate, just from further away.

And expect the plateau. Most runs here end during the climb rather than on the plateau, because players assume the game will keep getting harder and give up before they find out that it does not.

Frequently asked questions