Real Physics vs Simplified Game Physics

Every physics game lies, and the interesting question is which lies it tells and why. Some are forced by the hardware, some are chosen for playability, and a few are simply conventions that nobody has revisited.

This guide is an honest account of where the games on this site — and almost every other arcade physics game — depart from real physics. It matters because these games are genuinely useful for building intuition, and knowing where the model stops is what makes that intuition trustworthy.

Why games simplify at all

Three separate pressures push in the same direction, and they account for essentially every simplification in the genre.

Time

A physics step has roughly 8 milliseconds to complete if the game is to hold 60 frames per second with room for rendering. Everything that does not fit in that budget is cut, regardless of how accurate it would be.

Stability

Accurate models are often numerically fragile. A perfectly rigid contact, a genuinely stiff spring or a true inverse-square force at close range all destabilise a fixed-timestep simulation. Softening them is not laziness — it is what keeps the game running.

Playability

The biggest one. Real physics frequently produces bad games. Realistic restitution varies with impact speed, making bounces unpredictable. Realistic friction makes most characters feel sluggish. Realistic air resistance makes projectiles frustrating to aim.

The goal is not accuracy. The goal is a model that behaves the way players expect physics to behave, quickly enough to run sixty times a second.

The specific departures

These are the ones that apply to essentially every arcade physics game, including all thirty on this site.

Restitution is treated as constant

In reality the coefficient of restitution falls as impact speed rises — a ball dropped harder returns proportionally less. Games treat it as a fixed material property because a speed-dependent value makes bouncing unpredictable and aiming impossible.

Air resistance is linear or absent

True drag rises with the square of speed and depends on the object's shape, orientation and the fluid's density. Most games use linear damping, which is cheaper and more stable. Air Balloon Physics and Fluid Drift use quadratic drag because the feel matters there; nothing else does.

Contacts are points, not areas

A real box resting on a floor contacts it across a whole face. A simulated one usually contacts at a small number of points, which is why simulated stacks are less stable than real ones and need solver iterations to compensate.

Rotation is often disabled

Many player characters have rotation locked entirely, because a player character that tumbles when it lands is usually less fun than one that does not. That is a gameplay decision wearing physics clothing.

Collisions are discrete

Objects are tested for overlap after they move. A fast enough small object passes through a thin wall without ever overlapping it. Continuous collision detection solves this and costs considerably more.

Forces are capped

An inverse-square force becomes infinite at zero distance. Every implementation adds a softening term and usually a hard cap. Magnet Motion does both, which is why very close control there is merely difficult rather than impossible.

The units problem

Real physics uses metres, kilograms and seconds. A game uses pixels, arbitrary mass units and seconds. Bridging them is the first design decision in any physics game, and most games never bother.

The honest approach is to pick a scale and be consistent: decide how many pixels make a metre, and derive everything from there. The practical approach, used almost universally, is to work backwards from feel: choose how long a full-screen fall should take and solve for the acceleration that produces it.

For a 700-pixel play area and a one-second fall, that gives roughly 1,500 pixels per second squared, which is the baseline gravity here. Calling it '9.8' would be false precision; calling it 'normal' and documenting what it means is honest.

Notice that friction is the one row where the game value is directly comparable to reality, because the coefficient is dimensionless. That is why the numbers printed on the track in Friction Runner are genuinely in the right range.

What games do keep faithful

It would be misleading to leave the impression that nothing is real. The relationships are almost always correct, and that is the part that builds intuition.

• Mass independence of gravitational acceleration. Everything falls at the same rate, and games get this right because getting it wrong would look obviously broken.
• Conservation of momentum. Impulse-based resolution conserves it by construction — every impulse applied to one body is applied equally and oppositely to the other.
• The inverse-square law. Gravity wells genuinely follow 1/r², which is why orbital mechanics in a game behaves the way it does.
• Pendulum period independence from mass. Falls directly out of the equation, and games that integrate it get it for free.
• Torque as force times lever arm. Balance games depend on this being exactly right, and it is.
• Energy relationships. Spring energy going with x², kinetic energy with v², stopping distance with v² — all preserved.

What a physics game can and cannot teach

What it teaches well

• Direction of effects. Increase friction and stopping distance falls. Increase compression and stored energy rises much faster than linearly.
• Scaling. After enough Spring Jump, the fact that a full charge is worth four half charges stops being a formula and becomes an expectation.
• Vector thinking. Games that draw velocity vectors teach you to think about direction and magnitude separately, which is a genuine conceptual step.
• Conservation intuition. Understanding that a chain reaction has a finite reach because energy is lost at each link is a real physical insight.

What it does not teach

• Numbers. Nothing in an arcade game is in SI units.
• When models break down. Games hide their approximations rather than exposing them.
• Formal reasoning. Intuition is not derivation, and a physics course will ask for the latter.
• Edge cases. Relativistic effects, quantum behaviour and even ordinary fluid turbulence are entirely absent.

The honest summary: physics games build the intuition a textbook assumes you already have. Used before a course they make the material feel familiar; used instead of one they teach a confident set of approximately-right beliefs.

Why labelling matters

A game that claims to be a simulation and is not is doing something worse than simplifying — it is misleading.

That is why every game page on this site explains what its game models and where it stops. Air Balloon Physics says on the page and on screen that it is a simplified arcade model. Fluid Drift says that real fluid dynamics involves pressure fields no browser game solves in real time. Ricochet Lab says outright that its perfectly elastic walls are physically impossible and explains why they are there anyway.

Four things games get wrong that everyone believes

Some game physics conventions are so widespread that players have absorbed them as facts about the world. These four are worth correcting explicitly.

1. Air control

Almost every platformer lets you change direction in mid-air. Nothing about a real jump permits this: once you leave the ground there is nothing to push against. It is in games because jumps would be unforgiving without it, and it is such an old convention that removing it reads as a bug.

The games here include a small amount deliberately and keep it genuinely small, so that a jump is still a commitment.

2. Instant stopping

Releasing the movement key halts most game characters within a few frames. Real deceleration takes a distance proportional to the square of speed. The convention exists because responsive controls feel good, and it is exactly the convention physics games break on purpose.

3. Terminal velocity as a hard cap

Many games simply clamp falling speed at a maximum. Real terminal velocity is an equilibrium — drag grows until it balances gravity, and the approach is gradual. A clamp accelerates at full rate and then stops accelerating abruptly, which is visibly different.

4. Heavier things fall faster

Some games genuinely implement this, usually by accident, and it reinforces one of the most persistent misconceptions in physics. Gravity produces the same acceleration on every mass. Only drag distinguishes them, and drag depends on shape and size rather than weight.

Frequently asked questions

References