Mandelbrot field guide
What are you actually looking at?
The Mandelbrot set is the collection of complex numbers c for which repeated application of z ← z² + c, starting from z = 0, remains bounded. The black interior represents points that did not escape; exterior color records how their orbits departed.
One rule, one questionChoose c. Start at zero. Square, add c, repeat.Does the orbit stay bounded forever—or eventually fly away?
01Complex number
A number a + bi represented as one point on a two-dimensional plane.
02Iteration
One repetition of z ← z² + c, using the last result as the next input.
03Orbit
The sequence of z values generated for one chosen point c.
04Escape time
How many iterations pass before an orbit grows beyond the escape radius.
05Boundary
The infinitely intricate frontier between escaping and bounded points.
06Self-similarity
Small regions echo the whole set without being perfectly identical copies.
How this renderer works
From pixel to coastline.
- 1Map the pixel
Convert its screen position into a complex number c.
- 2Run the orbit
Repeat z ← z² + c up to the current iteration budget.
- 3Test escape
If |z| exceeds the radius, its orbit is guaranteed to diverge.
- 4Color the evidence
Escape speed becomes a smooth color; unescaped points remain black.
Why deep zoom eventually stops
Infinity meets floating-point arithmetic.
The mathematical boundary has detail at every scale, but ordinary JavaScript numbers carry about 15–16 significant decimal digits. Near 10¹⁴× magnification, neighboring pixels can collapse onto the same number and the image begins to lose trustworthy detail.
Going farther requires arbitrary-precision arithmetic or perturbation methods. This laboratory shows its remaining precision live so the limit is visible rather than hidden.
1905Fatou begins studying iterative rational functions.
1918Julia publishes foundational work on iteration in the complex plane.
1978–80Benoît Mandelbrot renders and popularizes the connectedness set now bearing his name.
TodayDeep-zoom methods reveal scales far beyond direct floating-point calculation.