2.2
Brownian motion
Continuous everywhere, differentiable nowhere, and the same roughness at every magnification. Independent Gaussian increments, martingale, Markov; simulated as a cumulative sum of √dt·Z.
Shrink the time step and most functions flatten; Brownian motion refuses to.
What it is
Brownian motion is a random walk with variance for a walk with time step . It's known as a Wiener process and often represented using . There are six key properties of the walk:
- It starts at zero, .
- Its increments are independent of each other — what does over the interval is independent of anything it did before . This also makes it a Markov process (the future is only dependent on the present, not the past — it retains no memory).
- It is a martingale (the best prediction of the future is the present — for all ).
- Step increments are Gaussian with variance equal to elapsed time .
- It is continuous everywhere with no jumps or discontinuities.
- It is not differentiable — the slope grows towards infinity, with as .
Variances subtract, standard deviations don't
is normally distributed with variance and standard deviation . Variances subtract, standard deviations don't.
Simulating it
Build the path with time steps , then — Gaussian because that is the distribution the continuous limit produces from any finite-variance step, according to the central limit theorem.
So the path becomes , which is the cumulative sum of the standard deviation of multiplied by a Gaussian variable , or — .
Excess kurtosis of the increment over steps, 40 000 paths of 2 000 steps at
, seed 20260903 — from the committed
experiments/2_2_brownian_motion.json. Zero is Gaussian. The coin column is
the central limit theorem arriving.
| lag (steps) | coin walk | Gaussian walk |
|---|---|---|
| 1 | −2.000 | −0.0005 |
| 10 | −0.198 | −0.0005 |
| 100 | −0.023 | −0.0045 |
| 250 | −0.001 | −0.0018 |
Continuous everywhere, differentiable nowhere
The typical size of the slope is therefore , so as the slope infinity — a smooth function's slope tends to its derivative.
There are no jumps, no discontinuities, since as — the data shows that with sampling: the largest move was 0.0056 at .
100 paths on a fine grid of step over , measured over every
non-overlapping window of width — the experiment's h, and the row
labels below — seed 20260903, from the committed
experiments/2_2_brownian_motion.json. Read down: the move shrinks, the slope
grows.
| mean | mean | largest | |
|---|---|---|---|
| 0.2478 | 2.478 | 1.0521 | |
| 0.07923 | 7.923 | 0.3492 | |
| 0.02516 | 25.16 | 0.1414 | |
| 0.007977 | 79.77 | 0.04842 | |
| 0.002523 | 252.3 | 0.01589 | |
| 0.0007979 | 797.9 | 0.005551 |
is a -sized random increment.
The same at every magnification
Zoom keeps the first 1/z of the path and stretches space by √z — the scaling Brownian motion is invariant under. The zoomed picture is as rough as the whole one; that roughness is what stops the path having a derivative.
Zooming into the first 1 % of a Brownian path only looks "small" because its variance is ; multiplying the vertical axis by ( for a 1 % zoom) restores the variance to , showing that the zoomed segment has the same roughness as the full path. Because every zoomed-in piece, once scaled by , statistically reproduces the whole path, Brownian motion never becomes smooth at any magnification.
The best forecast is where you are
because this is a martingale — the expected value of , given the known value of , is the known value of . The spread is 1 because is a Markov process — no memory: the variance at is for the time step , therefore the spread is .
Into 2.3
is continuous but infinitely rough, so the usual idea "differentiate first, then integrate" is impossible. With no derivative to supply a term like , calculus on must instead be built from its increments — the -sized kicks whose variance is — and from how those increments accumulate. Section 2.3 begins by replacing classical derivatives with this increment-based geometry, and begins to build the rules of stochastic calculus from that starting point.
Check
W is a Brownian motion. What is the distribution of W(0.5) − W(0.25)?
At a fixed time t, you measure the secant slope (W(t+dt) − W(t))/dt for smaller and smaller dt. As dt → 0 the typical size of the slope:
A Brownian path has reached W(1) = 1.5. Seen from now, W(2) has expected value and standard deviation:
You zoom into the first 1 % of a Brownian path and stretch time by 100 to fill the screen. For the picture to look statistically like the whole path, the vertical axis must be stretched by:
Build a walk from ±1 coin flips with √dt scaling, and look at its increments over 250 steps. Their distribution is: