The claim: we are probably living in a computer simulation
Simulation theory says that the world we experience may be a computer simulation run by a more advanced civilization. In its popular form it is stated as a probability: if simulations are possible, simulated minds would outnumber real ones, so we are probably simulated.
The theory descends from a precise argument by the philosopher Nick Bostrom, ‘Are You Living in a Computer Simulation?’, published in the Philosophical Quarterly in 2003. The popular version drops the conditions that the argument stated.
What Bostrom actually argued
Bostrom’s paper states a trilemma. At least one of three propositions is true: almost all civilizations at our level die out before they become able to run detailed simulations of their ancestors; or almost none of those that survive are interested in running such simulations; or we are almost certainly living in a simulation. He does not claim the third is true. He claims that the three cannot all be rejected.
The argument is a piece of arithmetic about observers. Let fP be the fraction of civilizations that reach a stage able to run simulations, fI the fraction of those that choose to run ancestor simulations, and N the average number of simulations each such civilization runs. The fraction of all observers who are simulated is then fP·fI·N ÷ (fP·fI·N + 1).
That equation is not in dispute. What is disputed is every input. Nobody has measured how many civilizations there are, how many survive, whether a simulation of a conscious mind is possible, or what a powerful civilization would want to run. Bostrom’s paper also assumes that a simulated mind could be conscious, a claim about minds that cannot be settled by physics alone.
The fraction as a function of three guesses
Take the default odds: 10% of civilizations mature, 1% of those run simulations, and each runs 1,000 of them. The product fP·fI·N is exactly 1, so the simulated minds equal the real ones and the fraction is one half. Multiply the number of simulations by ten and 91% of minds are simulated; divide it by ten and the figure is 9%.
The break-even number of simulations, N = 1 ÷ (fP·fI), is the value at which half of all minds are simulated. With the default odds it is 1,000. If the product of the first two odds is one in a million, the break-even N is a million.
The calculation shows what the argument is made of. It turns a gap in knowledge into a number, and the number follows whichever guess is made first. Nothing in the equation measures the world.
What fraction of minds are simulated?
Set the three odds in Bostrom’s equation. They are assumptions, not measurements: nobody knows any of them.
The share of all minds that live inside a simulation under these odds.
The number of simulations at which simulated and real minds are equally many.
Working tape
- fP × fI
0.1 × 0.01 = 0.001 - fP × fI × N
0.001 × 1,000 = 1 - fP × fI × N + 1
1 + 1 = 2 - Fraction of minds simulated
1 ÷ 2 = 0.5 - Simulated minds as a percentage
0.5 × 100 = 50 - Simulations per civilization for an even split
1 ÷ 0.001 = 1,000
An argument about odds, with no experiment attached
No experiment has detected a simulation. The physicists Silas Beane, Zohreh Davoudi and Martin Savage proposed in 2012 that a simulation on a spatial grid would leave a signature in the highest-energy cosmic rays, a preferred direction in which they travel, and they showed what grid spacings current data allow. They did not find such a signature, and their proposal applies only to a particular kind of simulation on a cubic lattice.
The theory remains a statement about probabilities, not a finding. That does not make it foolish: it belongs with other thought experiments that use probability to ask what we should believe about the world. It also means that the page you are reading cannot confirm or refute it, and anyone who says the question is closed in either direction is overstating what is known.
Why people believe it
Video games and virtual reality make a convincing world seem within reach, and the sense that a few lines of code can produce a world makes the idea feel familiar. Physics also offers hints that appeal: quantised energy levels, a finite speed of light, and a universe that follows exact equations feel like the output of rules. Public figures have said the odds are high, and the argument sounds like a calculation. The step from a trilemma, in which one branch is a simulation, to a claim that simulation is likely, requires accepting the first two propositions as improbable, which is the part that is rarely examined.
Where to test it
This is the one page on the site where the measurement is an argument rather than an observation. For claims about the physical world with an experiment attached, start with the Foucault pendulum or the Eratosthenes experiment, where the prediction is a number you can compare with a reading.
