Why does the path look so messy, but the average ⟨r²⟩ plot is a straight line?

An individual random walk path is unpredictable and 'jagged' because each step is random. However, the mean squared displacement ⟨r²⟩ is a statistical average over many possible paths. The linear trend emerges from the law of large numbers, showing that while individual outcomes are random, their average behavior is predictable and follows the diffusion law ⟨r²⟩ ∝ N.

Is this how real particles like pollen grains in water actually move?

Yes, this is a simplified model of Brownian motion. Real particles are constantly bombarded by water molecules, resulting in a random, jittery path. The random walk model captures the essential statistical feature—the linear growth of mean squared displacement with time—which Albert Einstein used in 1905 to prove the atomic theory of matter.

What does the 'running mean' of ⟨r²⟩ show?

The running mean calculates the average of r² up to each step number N for the single walker you are watching. It starts very noisy for small N but converges toward the theoretical straight line (N L²) as N increases. This demonstrates how statistical predictability improves with more data, a key concept in probability.

What are the main limitations of this simple random walk model?

This model assumes fixed step lengths, instantaneous steps, and no interactions between walkers or with an environment. Real-world diffusion can involve variable step sizes, waiting times between steps, and external forces (leading to biased walks). It also models 'normal' diffusion; many complex systems like transport in cells exhibit 'anomalous' diffusion where ⟨r²⟩ ∝ N^α with α ≠ 1.

Random Walk

1D or 2D steps; trail and running mean ⟨r²⟩ vs diffusion intuition.

About this model

A random walk is a fundamental stochastic process that models the path of an object taking successive steps in random directions. This simulator visualizes both one-dimensional and two-dimensional random walks, where each step is of fixed length but its direction is chosen randomly (e.g., left/right in 1D or along a grid in 2D). The core mathematical principle is that the mean squared displacement, denoted ⟨r²⟩, grows linearly with the number of steps, N. Specifically, in an unbiased walk with step length L, ⟨r²⟩ = N L² in 1D, and ⟨r²⟩ = N L² in 2D for a square lattice. This linear relationship ⟨r²⟩ ∝ N is the signature of normal, or Fickian, diffusion. The simulator plots the running average of r² versus N, allowing users to observe how this linear trend emerges from the noise of individual random paths. Key simplifications include the fixed step length, the absence of external forces or biases, and the assumption that each step is independent of all previous steps (a Markov process). By interacting with the simulation, students learn to connect the microscopic random motion of individual steps to the emergent macroscopic behavior of diffusion, building intuition for phenomena like Brownian motion, heat conduction, and population dispersal.

Who it's for: High school and undergraduate students in physics, mathematics, or biology courses introducing probability, statistics, or diffusion processes.

Key terms

Random Walk
Mean Squared Displacement
Diffusion
Stochastic Process
Brownian Motion
Markov Process
Normal Distribution
Central Limit Theorem

How it works

Contrasts with the smooth Brownian path in thermodynamics: here each step is a clean coin flip or uniform angle, ideal for proving scaling laws.

Frequently asked questions

Why does the path look so messy, but the average ⟨r²⟩ plot is a straight line?: An individual random walk path is unpredictable and 'jagged' because each step is random. However, the mean squared displacement ⟨r²⟩ is a statistical average over many possible paths. The linear trend emerges from the law of large numbers, showing that while individual outcomes are random, their average behavior is predictable and follows the diffusion law ⟨r²⟩ ∝ N.
Is this how real particles like pollen grains in water actually move?: Yes, this is a simplified model of Brownian motion. Real particles are constantly bombarded by water molecules, resulting in a random, jittery path. The random walk model captures the essential statistical feature—the linear growth of mean squared displacement with time—which Albert Einstein used in 1905 to prove the atomic theory of matter.
What does the 'running mean' of ⟨r²⟩ show?: The running mean calculates the average of r² up to each step number N for the single walker you are watching. It starts very noisy for small N but converges toward the theoretical straight line (N L²) as N increases. This demonstrates how statistical predictability improves with more data, a key concept in probability.
What are the main limitations of this simple random walk model?: This model assumes fixed step lengths, instantaneous steps, and no interactions between walkers or with an environment. Real-world diffusion can involve variable step sizes, waiting times between steps, and external forces (leading to biased walks). It also models 'normal' diffusion; many complex systems like transport in cells exhibit 'anomalous' diffusion where ⟨r²⟩ ∝ N^α with α ≠ 1.

Similar topics nearby (including related fields) — or browse all 85 in Math Visualization.

View category →

NewSchool

Brownian Motion

Launch Simulator

Heavy particle with random kicks and friction; trail and running ⟨r²⟩ vs time.

NewSchool

Random Walk (2D / 3D Lattice)

Launch Simulator

Nearest-neighbor SRW on Z² or Z³: ensemble mean ⟨r²⟩ vs time with y = t reference, histogram of r² across walkers, and Monte Carlo first-return times (recurrent vs transient).

NewSchool

Markov Chain (Weather)

Launch Simulator

Sun/Rain two-state chain: P matrix, stationary π, empirical vs theory.

NewUniversity / research

Kalman Filter (1-D)

Launch Simulator

Recursive optimal estimation: noisy measurements, hidden truth, predict + update with Q and R; random-walk or constant-velocity model with ±2σ band and innovations.

NewUniversity / research

LMS / NLMS Adaptive Noise Cancellation

Launch Simulator

Primary p = s + v with v a fixed unknown FIR of Gaussian reference x[n]. Watch an L-tap FIR adapt by LMS or NLMS so error e = p − wᵀx → s; running MSE and ‖w − h‖.

NewSchool

Least Squares Fit

Launch Simulator

Noisy linear data; fitted slope and intercept with residuals.

Random Walk

About this model

How it works

Frequently asked questions

Related simulators

Brownian Motion

Random Walk (2D / 3D Lattice)

Markov Chain (Weather)

Kalman Filter (1-D)

LMS / NLMS Adaptive Noise Cancellation

Least Squares Fit

Random Walk

About this model

How it works

Frequently asked questions