2a. Central Limit Theorem Visualizer

Sample from any distribution, compute the mean of each sample, and watch the distribution of means converge to a normal (bell) curve regardless of the source distribution.

Controls

Total Trials0
Mean of Means--
Std Dev--

Mechanism

Regardless of the underlying distribution -- uniform, exponential, bimodal, or anything else -- the average of many independent random variables approaches a normal distribution. This is the Central Limit Theorem. As sample size N increases, the distribution of sample means becomes increasingly bell-shaped and tighter around the true mean. The standard deviation of means shrinks proportionally to 1/sqrt(N).

Real-World Examples

  • Height distributions -- the sum of many genetic and environmental factors produces a bell curve
  • Measurement errors -- aggregate many small random errors and the total error is normally distributed
  • Stock portfolio returns -- diversified portfolio returns approximate normality (in normal times)
  • Polling averages -- averaging many polls produces a more reliable, normally-distributed estimate

Key Insight

The bell curve is not an assumption about how the world works -- it is a mathematical consequence of aggregation. Whenever you see a normally-distributed outcome, ask: "What independent factors are being summed here?" The CLT explains why so many phenomena in nature and society follow this pattern.

2b. Conway's Game of Life

Cells on a grid live, die, or are born based on four simple rules. From these emerge gliders, oscillators, spaceships, and even structures capable of universal computation.

Controls

Presets

Click on the grid to draw/erase cells.

Generation0
Population0
Births0
Deaths0

Mechanism

Just four simple rules: (1) Any live cell with fewer than 2 live neighbors dies (underpopulation). (2) Any live cell with 2 or 3 live neighbors survives. (3) Any live cell with more than 3 live neighbors dies (overcrowding). (4) Any dead cell with exactly 3 live neighbors becomes alive (reproduction). From these emerge gliders (moving patterns), oscillators (repeating patterns), spaceships, and even universal computation. Simple rules produce complex emergent behavior.

Real-World Examples

  • Traffic patterns -- cars follow simple rules but produce complex jams and waves
  • Urban growth -- cities expand through local decisions, creating emergent structure
  • Chemical reaction-diffusion patterns -- Turing patterns in biology
  • Flocking behavior -- birds follow 3 simple rules (separation, alignment, cohesion) to create murmurations

Key Insight

The Game of Life demonstrates that complexity does not require complex rules. The whole performs functions that far exceed the capacities of its parts. This raises deep questions: if consciousness and intelligence can emerge from simple neuronal rules, could complex social phenomena similarly emerge from simple individual behaviors?

2c. 1D Cellular Automata (Wolfram)

A single row of cells evolves according to a rule that maps each 3-cell neighborhood to a new state. All 256 possible rules are encoded in a single byte, yet they produce behaviors ranging from uniform to chaotic to complex.

Controls

Rule Lookup Table

Rule110
ClassComplex
Width201

Mechanism

Each cell looks at itself and its two neighbors (3 cells = 8 possible patterns). The rule number, converted to binary, tells each pattern whether the center cell should be alive or dead in the next generation. A single byte (0-255) completely determines the behavior. Wolfram classified outcomes into four classes: Class 1 (uniform/empty), Class 2 (periodic/repetitive), Class 3 (chaotic/random), and Class 4 (complex). Rule 110 is provably Turing complete -- it can compute anything a computer can.

Real-World Examples

  • Pattern formation on Conus textile sea snail shells -- strikingly similar to Rule 30 output
  • Crystal growth patterns -- simple atomic-level rules create complex macroscopic structure
  • Traffic flow models -- Rule 184 models basic traffic dynamics (cars as 1s, empty road as 0s)
  • Random number generation -- Rule 30 is used by Wolfram Mathematica for pseudorandom numbers

Key Insight

A single number between 0 and 255 can specify behavior ranging from trivially empty to capable of universal computation. This suggests that the boundary between simplicity and complexity is razor-thin, and that immense computational power can hide in extremely simple systems.

2d. Preference Aggregation (Condorcet Paradox)

Different voting methods applied to the same set of voter preferences can produce different winners. Arrow's Impossibility Theorem proves no ranked voting system can satisfy all fairness criteria simultaneously.

Controls

Voters0
Candidates0
Condorcet Winner?--
Methods Agree?--

Mechanism

With three or more candidates, individual rational preferences can produce irrational group outcomes. The Condorcet Paradox: A beats B, B beats C, but C beats A -- a cycle with no clear winner. Different aggregation methods break this cycle differently: Plurality picks the most first-place votes, Borda Count scores all rankings, Condorcet checks head-to-head matchups, and Runoff eliminates the weakest candidates. Arrow's Impossibility Theorem proves that no ranked-choice voting system can simultaneously satisfy all fairness criteria (non-dictatorship, Pareto efficiency, independence of irrelevant alternatives).

Real-World Examples

  • Presidential elections -- the 2000 US election (Bush/Gore/Nader) is a classic spoiler-effect case
  • Committee decisions -- department hiring committees often cycle through candidates
  • Product feature prioritization -- stakeholders rank features differently, aggregation method matters
  • AI alignment -- aggregating human preferences into an objective function faces the same impossibility

Key Insight

There is no perfect voting system. The method of aggregation is not a neutral procedure -- it actively shapes the outcome. This means the choice of how to choose is itself a consequential decision. In any group decision with 3+ options, someone must decide the rules, and the rules partly determine the winner.