The Drunkard's Walk Home

RMS distance after n steps = √n in any dimension. After 100 steps: 10m. After 10,000 steps: 100m. Bonus: in 1D and 2D the walk returns to the origin with probability 1 (recurrent). In 3D it escapes with probability ~66% (transient). WOW Moment: - After 1 step: 1m from bar. - After 100 steps: 10m from bar. - After 10,000 steps: 100m from bar. - After 1,000,000 steps: 1,000m = 1 km from bar. - To be 10x farther, you need 100x more steps. To be 100x farther, you need 10,000x more steps. - This is why oxygen molecules in a room take minutes to diffuse across it even though they travel at 500 m/s. Their random collisions cancel most progress. - In 1D, will the drunkard EVER return to the bar? YES — with probability 1. In 2D? Also YES — probability 1. In 3D? NO — probability of return is only 34%. A 3D random walk escapes to infinity with 66% probability. 'A drunk man will find his way home, but a drunk bird may get lost forever.' — Shizuo Kakutani

The random walk (or Drunkard's Walk) is a fundamental mathematical model for diffusion, brownian motion, and statistical mechanics. It reveals how random, uncorrelated movements lead to a slow, square-root expansion over time. The Mathematical Proof for RMS Distance: Let the steps be represented by independent and identically distributed (i.i.d.) random variables X1, X2, ..., Xn. In 1D: - Each step Xi is either +1 or -1, with P(Xi = +1) = 0.5 and P(Xi = -1) = 0.5. - The expected value of a single step is E[Xi] = 0. - The variance of a single step is E[Xi^2] = 0.5*(1)^2 + 0.5*(-1)^2 = 1. The position after n steps is S_n = X1 + X2 + ... + Xn. 1. The expected position is E[S_n] = E[X1] + ... + E[Xn] = 0. 2. The expected squared distance is: E[S_n^2] = E[(X1 + X2 + ... + Xn)^2] E[S_n^2] = E[Σ Xi^2 + 2 * Σ_{i < j} Xi * Xj] By linearity of expectation: E[S_n^2] = Σ E[Xi^2] + 2 * Σ_{i < j} E[Xi * Xj] 3. Since the steps are independent, E[Xi * Xj] = E[Xi] * E[Xj] = 0 * 0 = 0. 4. Therefore: E[S_n^2] = Σ E[Xi^2] = n * 1 = n. 5. The Root Mean Square (RMS) distance is: RMS = √E[S_n^2] = √n. This result generalizes to any dimension d. If we take steps of length 1 along the coordinate axes, the sum of the variances along each dimension still adds up to n, giving an RMS distance of √n. To explain the WOW part: - The √n scaling explains why diffusion is extremely slow over long distances. If a molecule takes 10^12 collisions (steps) per second, it will travel 1 million steps (1 mm) in 1 second. To travel 2 mm (double the distance), it must travel 4 million steps, which takes 4 seconds. The time required scales quadratically with distance. - The recurrence of random walks (Pólya's Theorem) is another mind-blowing result. Shizuo Kakutani summarized it beautifully: "A drunk man will find his way home, but a drunk bird may get lost forever." - In 1D and 2D, the random walk is recurrent. The probability that the walk returns to the starting point is exactly 1. - In 3D (and higher), the walk is transient. The extra degrees of freedom allow the walk to escape. The probability of returning to the origin in 3D is only about 34.05%, meaning there is a 65.95% chance the walker never returns.