Science & Technology
How Fractals Are Used in the Real World
From the antenna folded inside your phone to the way a radiologist reads a tumour, fractal geometry quietly powers medicine, wireless engineering, finance, computer graphics and the measurement of the Earth itself.
Fractals are usually introduced as a curiosity — a screensaver, a maths-class snowflake, a pretty thing to zoom into. But the reason Benoit Mandelbrot fought for decades to be taken seriously is that the rough, self-similar geometry he named turned out to describe the real world far better than the smooth lines and perfect circles of Euclid. Coastlines, lungs, markets, the static on a phone line: none of them are smooth. Once engineers and scientists had the language of fractals, they could not only describe these irregular systems but build with them and diagnose through them.
This is the working floor of fractal geometry. Below are the five domains where fractals have moved from elegant theory into devices, clinics, trading desks, film studios and atlases — each grounded in real inventors, real instruments and real published research.
The big idea in one line: a single rule repeated at many scales lets you fold a long antenna into a tiny phone, pack a tennis court of lung tissue into a chest, measure how rough a market or a tumour is with one number (the fractal dimension), and generate an entire mountain range from a few lines of code. The applications all exploit the same property: self-similarity across scale.
How are fractals used in technology and engineering?
The most ubiquitous fractal application is probably within arm's reach right now. In 1987 the radio astronomer Nathan Cohen attended a lecture by Benoit Mandelbrot and wondered what would happen if an antenna were bent into a fractal shape. In 1988 he built the first deliberate fractal-element antenna out of aluminium foil and a pagoda-shaped Koch curve, taped to his apartment railing — until his landlord cut it down. Cohen went on to found Fractal Antenna Systems and hold close to a hundred patents on the idea, as the IEEE History Center records.
The physics is elegant. A conventional antenna resonates best when its length matches the wavelength it receives, which is why old radios needed long aerials. A fractal antenna instead folds a long conductor into a self-similar pattern — a Koch curve, a Sierpiński gasket, a Minkowski loop — cramming a great deal of electrical length into a tiny footprint. Crucially, because the same shape repeats at several scales, the antenna resonates at multiple frequencies at once. That single property is why one small element inside your phone can handle Wi-Fi, Bluetooth, GPS and several cellular bands simultaneously, as Wikipedia's fractal-antenna entry summarises. The University of Waterloo's mathematics faculty lists antennas among its top five applications of fractals, noting that Hilbert-curve geometries enable especially low-profile, high-performance designs.
Fractal geometry also turns up in heat exchangers (branching channels maximise surface contact), in fractal camouflage patterns such as the US military's MARPAT, and in the layout of efficient electronic circuits — anywhere engineers need to pack a large amount of surface, length or branching into a small, bounded space.
How are fractals used in medicine?
Biology is built fractally, so medicine reads the body fractally too. Your lungs, blood vessels and neurons are all branching, self-similar networks, and their fractal dimension — a single number capturing how intricately they fill space — changes measurably when something goes wrong. The University of Waterloo notes that healthy blood-vessel cells grow in orderly fractal patterns, which makes the disorderly growth of cancerous tissue easier to flag through fractal analysis of medical images.
The most striking medical use is in the heartbeat itself. The cardiologist Ary Goldberger and colleagues showed that a healthy heart is not a metronome: a normal interbeat interval fluctuates in a complex, fractal way, with long-range correlations stretching over thousands of beats. Their landmark paper in the Proceedings of the National Academy of Sciences (Goldberger et al., 2002) reframed health itself: it is not stability but complexity that marks a robust system. In patients at high risk of sudden cardiac death — for example those in congestive heart failure — that fractal complexity breaks down. The same loss of fractal ‘richness’ accompanies aging, which is why fractal analysis of heart-rate variability, retinal vessels and brain MRIs has become a quiet diagnostic tool. Geometry becomes a vital sign.
How are fractals used in finance?
Mandelbrot's very first brush with real-world data was financial, not biological. In 1963 he analysed more than a century of daily cotton prices and found that they did not follow the tidy bell-curve distribution that orthodox economics assumed. Instead they followed a power-law (Pareto-Lévy) distribution with so-called “fat tails”: violent, large moves happen far more often than a Gaussian model predicts. He laid out the full argument for a general audience in The (Mis)Behavior of Markets (the book; the cotton work itself predates it).
His three claims were that price changes are not independent from one period to the next, that a chart of a single day looks statistically like a chart of a month or a decade (self-similarity across time), and that markets are therefore far rougher and riskier than standard models such as Value-at-Risk imply. Volatility, he argued, is not a fixed number but clusters — large changes follow large changes. This fractal-market view has not replaced mainstream finance, but it deeply informs modern risk management, the modelling of crashes, and trading tools such as the fractal indicator used in technical analysis. The table below summarises how each field puts the same underlying property to work.
| Domain | What is exploited | Real example |
|---|---|---|
| Wireless engineering | Multi-scale resonance & space-filling | Multiband fractal antenna in a phone (Nathan Cohen, 1988) |
| Medicine | Fractal dimension as a health signal | Heart-rate variability; tumour & retinal-vessel analysis (Goldberger et al., 2002) |
| Finance | Power-law “fat tails” & scale invariance | Mandelbrot's cotton-price study (1963); fractal-market risk models |
| Computer graphics | Recursive detail from simple rules | CGI mountains in Star Trek II (Loren Carpenter, 1982) |
| Geography | Fractional dimension of rough curves | Coastline paradox (Mandelbrot, Science, 1967) |
How are fractals used in computer graphics?
If you have ever been moved by a digital mountain range in a film or a video game, you have almost certainly seen a fractal. In 1980 a young programmer named Loren Carpenter — then at Boeing — wrote software that generated realistic mountains by recursively subdividing triangles and randomly displacing the midpoints, a technique that produces ever-finer, self-similar detail. He showed the result, a two-minute flight over fractal peaks called Vol Libre, at the SIGGRAPH conference in July 1980. It earned a standing ovation and an immediate job offer from Lucasfilm's nascent computer division.
Two years later that same fractal-terrain technique produced cinema's first fully computer-generated sequence: the Genesis Effect in Star Trek II: The Wrath of Khan, in which a dead planet's fractal mountains, atmosphere and shock wave bloom into life. Carpenter went on to co-found Pixar and won two Academy Awards for the underlying technology. The same family of methods — fractal and related procedural noise — now generates terrain, clouds, coastlines and foliage across the games and effects industries, because a few lines of recursive code can stand in for billions of hand-placed details. Fractal image compression applies the inverse idea: it encodes a picture as the self-similar transformations that regenerate it, packing complex imagery into very little data.
What is the coastline paradox, and why does it matter?
The application that started it all is also the most philosophically unsettling. In 1967, before he had even coined the word “fractal,” Mandelbrot published How Long Is the Coast of Britain? in the journal Science. Building on an obscure earlier observation by the meteorologist Lewis Fry Richardson, he showed that a coastline has no single well-defined length. Measure Britain's coast with a 100-km ruler and you get roughly 2,800 km; switch to a 50-km ruler, which captures more inlets and headlands, and the figure jumps to about 3,400 km. As the ruler shrinks, the measured length rises without limit.
This is the coastline paradox, and the resolution is fractal: a coastline is too rough to be a one-dimensional line but does not fill a two-dimensional plane, so its fractal dimension sits between the two — about 1.25 for Britain's west coast, near 1.02 for the smoother coast of South Africa, and as high as ~1.5 for Norway's fjord-cut shore. The practical upshot reaches into geography, GIS and mapping: any reported coastline or border length is meaningless without stating the measurement scale, and software that models terrain, erosion or river networks must build in this scale-dependence rather than assume a fixed answer.
Across all five domains the lesson is the same one Mandelbrot spent his life arguing. The world is not smooth, and pretending otherwise — with straight antennas, metronomic hearts, bell-curve markets, hand-drawn mountains or rulers of a single length — gets you the wrong answer. Fractal geometry is simply the mathematics that takes roughness seriously, and that is why it keeps showing up wherever the real world refuses to behave.
Frequently asked
What are fractals used for in the real world?
Fractals are used wherever a system is rough, branching or self-similar across scales. The headline applications are wireless engineering (compact multiband antennas folded into phones and other devices), medicine (using fractal dimension to analyse tumours, retinal blood vessels and heart-rate variability), finance (power-law "fat-tail" models of volatile markets), computer graphics (procedurally generated mountains, clouds and terrain, plus fractal image compression), and geography (measuring irregular coastlines and modelling river networks). In each case the same underlying property — one rule repeated at many scales — lets engineers pack length or surface into a small space, or lets scientists summarise complex roughness with a single number, the fractal dimension.
How do fractal antennas work?
A fractal antenna folds a long electrical conductor into a self-similar shape such as a Koch curve or Sierpiński gasket, packing a great deal of effective length into a tiny physical footprint. Because the same pattern repeats at several scales, the antenna resonates at multiple frequencies simultaneously rather than just one. That is why a single compact fractal element inside a smartphone can serve Wi-Fi, Bluetooth, GPS and several cellular bands at once. The radio astronomer Nathan Cohen built the first deliberate fractal-element antenna in 1988 after a Mandelbrot lecture inspired him, and later founded Fractal Antenna Systems, which holds dozens of patents on the technology now found in most wireless devices.
How are fractals used in medicine?
Many tissues and signals in the body are fractal — lungs, blood vessels, neurons and even the rhythm of the heartbeat — so doctors can measure their fractal dimension and watch how it changes with disease. Healthy blood-vessel networks grow in orderly fractal patterns, which makes the disorderly growth of cancer easier to detect through fractal image analysis. Most strikingly, cardiologist Ary Goldberger and colleagues showed that a healthy heartbeat is not a steady metronome but a complex, fractal signal; that complexity breaks down in patients at high risk of sudden cardiac death and declines with aging. Fractal analysis of heart-rate variability, retinal vessels and brain scans has therefore become a quiet diagnostic tool that turns geometry into a vital sign.
Did Mandelbrot really apply fractals to the stock market?
Yes — finance was one of Mandelbrot's earliest real-world subjects. In 1963 he studied more than a century of daily cotton prices and found they did not follow the neat bell-curve distribution that orthodox economics assumed. Instead they followed a power-law distribution with "fat tails," meaning extreme moves happen far more often than standard models predict. He argued that price changes are not independent, that a chart looks statistically the same across hours, months or decades, and that volatility clusters rather than staying fixed. He set out the full fractal view of markets in his book The (Mis)Behavior of Markets. This work did not replace mainstream finance but it reshaped how crashes, risk and volatility are modelled today.
How are fractals used in movies and computer graphics?
Fractals let computers generate enormous, realistic detail from a few simple rules. In 1980 Loren Carpenter wrote software that built convincing mountains by recursively subdividing and displacing triangles, and demonstrated it in a short film, Vol Libre, at the SIGGRAPH conference. The technique produced cinema's first fully computer-generated sequence two years later: the Genesis Effect in Star Trek II: The Wrath of Khan. Carpenter went on to co-found Pixar and won two Academy Awards for the underlying graphics technology. The same fractal and procedural-noise methods now generate terrain, clouds, coastlines and vegetation across films and video games, and fractal image compression uses the inverse idea to store complex pictures in very little data.
What is the coastline paradox?
The coastline paradox is the discovery that a coastline has no single, well-defined length. Benoit Mandelbrot demonstrated it in a 1967 paper in the journal Science titled "How Long Is the Coast of Britain?", building on an earlier observation by Lewis Fry Richardson. If you measure Britain's coast with a 100-km ruler you get about 2,800 km, but with a 50-km ruler — which captures more inlets and headlands — the figure rises to roughly 3,400 km, and it keeps growing as the ruler shrinks. The explanation is fractal: a coastline is rougher than a simple line but does not fill a plane, so its fractal dimension lies between 1 and 2 (about 1.25 for Britain). The practical lesson is that any stated coastline or border length is meaningless without specifying the measurement scale.
Why is fractal geometry so useful across such different fields?
Because all of these fields deal with the same problem: real systems are rough, irregular and self-similar, and classical Euclidean geometry — built from smooth lines, circles and planes — describes them poorly. Fractal geometry takes roughness seriously. The single concept of self-similarity across scale lets an engineer pack a long antenna into a small phone, lets a doctor summarise the complexity of a heartbeat or a tumour with one fractal-dimension number, lets a financier model markets that move in violent jumps, and lets an artist generate a whole mountain range from a recursive rule. Mandelbrot's lasting argument was that this shared roughness is not noise to be smoothed away but structure to be measured and exploited, which is why the same mathematics reappears in medicine, engineering, finance, graphics and geography.