The Poetics of Probability: Conceptual Art on Chance

A light that blinks when a coin lands tails

Picture a small room. A coin flies. It lands. A tiny light blinks when it is tails. A sensor hears the ring of metal on wood. A switch flips. The light goes on. Then it is dark again. You wait for the next throw. Your body tenses and softens with each flip. This is not a casino. It is a quiet artwork. Still, it sets a rule. It lets chance speak. It lets you feel risk, time, and hope in real space.

When artists say “let chance decide,” what do they mean?

Chance in art is not a shrug. It is a method. The artist sets a frame, then cedes part of control. It can shift how we read authorship. It can move our eye from the hand to the system. In the late 1960s, artists called this kind of work “conceptual.” The idea comes first; the form may change with each run. To see where this view came from, it helps to start with a clear, non-technical note on the term. See Tate’s definition of conceptual art. It sets the field and keeps the word honest.

Long before the rise of code art, some artists handed a slice of choice to rules, tosses, or time. Marcel Duchamp played with this in sharp ways. If you want a short, solid bio, the Guggenheim’s profile of Marcel Duchamp is a fine start.

Short detour: how probability “thinks”

Probability is the math of maybes. It does not predict the next coin toss. It frames how groups of tosses behave. It talks about events, odds, and patterns across many trials. It also warns us about traps in our minds. One trap is the gambler’s fallacy. After ten reds at roulette, we feel “black is due.” But the wheel does not “owe” us. Each spin stands alone if the system is fair.

If you want a clean door into the field, try the Stanford Encyclopedia of Philosophy on probability. For plain words on data habits and bias, the American Statistical Association’s overview is helpful. With that frame, we can look at art that puts chance to work.

Casebook of seven small experiments

1) Duchamp makes a new “ruler” by dropping threads

In 3 Standard Stoppages (1913–14), Duchamp let three one-meter threads fall onto canvas from a set height. Each curve was then fixed and made into a wooden “ruler.” Gravity drew the line; chance shaped the tool. The joke is dry and deep: we trust standards, yet here the standard is a record of an accident. The artist sets the game; the world throws the dice.

2) John Cage and I Ching: the score listens to coins

Composer John Cage used the I Ching to make choices in music. Coin tosses picked notes, lengths, and dynamics. The point was not chaos. It was to shift from taste to process. The results can feel spare, yet sharp. If you want a smart intro to this, see MoMA Learning on John Cage and chance operations. It shows how a simple device can rewire art and mind.

3) Ellsworth Kelly lets cut paper fall into place

Kelly made chance collages by drawing lots for colors and spots, or by letting small papers drop. He worked inside limits he chose. A grid. A set of hues. A rule on how many pieces. We learn to see the play between the strict plan and the free fall. For context on this postwar turn, the Met’s Heilbrunn Timeline of Art History helps set the era and the move toward systems.

4) William Anastasi draws with a train

In Subway Drawings, Anastasi holds a pencil on paper while he rides. He does not look. The car’s shake moves his hand. The line is a trace of the city’s body. The force is not random in a full sense. It has a shape, a pulse, a route. Still, no two rides are the same. Each page is a record of noise in time.

5) Alison Knowles writes a house with a computer

In The House of Dust (1967), Alison Knowles used a computer to permute lines like “A HOUSE OF DUST / ON OPEN GROUND / USING NATURAL LIGHT…” It is a poem and a set of build prompts. The work shows how rules and lists can make both text and space. You can browse notes and images at the Art Institute of Chicago’s page on The House of Dust.

6) Vera Molnár nudges the grid

Vera Molnár wrote early code to push lines off course by small, random steps. A square leans. A lattice stutters. These slight slips make the field come alive. Her practice sits between drawing and program. The seed is clear; the sway is small; the mood is rich.

7) Hans Haacke lets water draw on glass

In Condensation Cube, a sealed box of water steams and drips inside. Each room’s air and light shift the state. Drops slide, merge, and vanish. No hand moves them. The world does. The work is a tiny stage for chance, physics, and time. We face the fact that control can be thin, even in a cube.

How to read chance at a glance

We can learn to read the “engine” of chance in a work. What moves the choices? Coins? A list? A script? The room itself? What odds seem implied? How much control stays with the artist? A past show framed this well. See the Kemper’s exhibition essay, Chance Aesthetics, for a broad map. The table below gives a quick field guide.

Marcel Duchamp	3 Standard Stoppages (1913–14)	Drop of threads from a set height	Medium	Near-uniform landing shape under gravity	How “errors” become new rulers; the joke on standards
John Cage	Music of Changes (1951)	I Ching coin tosses choose notes	Low	Discrete uniform picks among options	Score marks that show the process; gaps and bursts
Ellsworth Kelly	Chance collages (1951–53)	Random color and placement in a grid	Medium	Uniform draw from fixed color/spot sets	Tension between plan and fall
William Anastasi	Subway Drawings (1968–)	Hand moved by train vibrations	Low	Stochastic motion; jitter as noise	Marks that mirror track curves and starts
Alison Knowles	The House of Dust (1967)	Computer-made permutations	Medium	Combinatorics over word lists	How text turns to acts or plans
Vera Molnár	Early generative plots (late 1960s–)	Small random offsets on grids	Medium	Bounded jitter; near-Gaussian drift	Subtle tilt of order into life
Hans Haacke	Condensation Cube (1963–65)	Heat and humidity drive droplets	Low	Thermal fluctuation; many micro-states	Ever-shifting bead paths on glass

Methods note: true randomness is hard

There is “physical” randomness, like a fair coin toss or thermal noise. There is also “pseudo” randomness, like a number stream from code. A good pseudo-random number generator (PRNG) can look random to us and to tests. But PRNGs depend on a “seed.” If you set the same seed, you get the same stream again. That can be a feature in art: you can re-run a work and study it.

How do we know if a stream is random enough? There are test suites that probe number runs for bias and pattern. A key set is the NIST SP 800-22 tests. They check things like “too many” runs of ones, odd gaps, or skewed bits. If a stream fails many tests, it may be flawed for fair choices.

Some groups try to pull entropy from the world. One famous case is a wall of lava lamps that a firm uses to make random seeds. The camera reads the glow and swirl. That feeds a system. It is fun and serious at once. See the story in Cloudflare’s lava-lamp randomness post. Artists also tap light, heat, or decay to seed their works. It adds a trace of the room and the day to each run.

From casino floors to gallery walls

Our culture tries to tame chance. Games of chance are a clear case. Slots and cards use tested random streams so that odds match the rules. Public trust needs proof. That is why labs audit RNGs and why rules exist. For a clean overview of this field, see the UK Gambling Commission guidance on fairness. It shows how “fair and open” is a duty, not a vibe.

Here is where art and games meet. Both turn raw chance into form. Both need clear frames. If you want a ground view of how claims of fairness show up for real people, it helps to look where play happens. A good, plain route is to read audits, license notes, and table rules in context. For a practical, editorial entry point, see Live Dealer Online Casinos. It lets you see how “regulated randomness” works when a live stream, a shoe of cards, and a studio come together. And please, if you play, keep it safe: check advice from BeGambleAware and set limits.

False chance: loaded dice and weak code

Not all “random” is fair. A set of dice can be shaved to fall a certain way. A card shoe can be stacked. A bad PRNG can leak its pattern. Art can fail here too. A work might call itself “random,” but show a clear loop or bias, and not on purpose. When chance is part of the claim, it helps to state the method and the limits. Tests and logs build trust.

Why this matters now: the creative seed in AI

In generative art and AI image tools, the seed is key. Change the seed, get a new image. Keep the seed, get the same image when you run the same prompt and model. This gives us a tool to explore range without losing ground. It also lets artists share and repeat a result for proof. Good notes make this strong: write the model, the prompt, the seed, and the code or app. Then your work can be checked, taught, and kept.

What is randomness, really? At deep levels, the debate is still open. Some math folks link it to compressibility and Kolmogorov. Some physics folks point to quantum events. A nice window into these ideas, in clear prose, is this piece in Quanta Magazine on what makes randomness random. For practice, though, our needs are humble: we want draws we cannot game, seeds we can note, and systems that tell the truth.

Pocket checklist for reading chance-based art

• Mechanism: What makes the choice? Toss, list, code, room, or body?
• Control: How much stays with the artist? What is fixed? What can vary?
• Odds: What outcomes seem likely? Is it uniform, weighted, or stepwise?
• Bias: Do you see loops, tilts, or repeats that do not fit the claim?
• Trace: Can you find notes, seeds, or logs that show the process?
• Ethics: Does the work state its method in a fair way? Is the risk to viewers clear?
• Context: How does the room, time, or tech change the run?

Afterthoughts and small questions

Does chance kill the artist’s voice?

No. It shifts it. The voice moves from picking each note to picking the frame and the rule. Style shows in what is set and what is left open.

Is “random” just a trend word?

It can be, but not when backed by clear method. When a work names its system and shows its trace, the word earns its keep.

How can I tell if a PRNG is “good enough” for art?

Ask what the work needs. If you want no repeats in a short show, most modern PRNGs will do. If you need proof against bias, run suites like NIST SP 800-22, or use a hardware source with a health check.

Where should I start if I’m new to this field?

Read a few core pages first: Tate on conceptual art, the Stanford entry on probability, and MoMA on Cage. Then go see works in person. Watch how time and place bend each run.

Selected sources and further reading

• Tate: Conceptual Art
• Guggenheim: Marcel Duchamp
• Stanford Encyclopedia of Philosophy: Probability
• American Statistical Association: What is Statistics?
• MoMA Learning: John Cage and Chance
• The Met: Heilbrunn Timeline of Art History
• Art Institute of Chicago: The House of Dust
• Kemper Art Museum: Chance Aesthetics
• NIST SP 800-22: Randomness Tests
• Cloudflare Blog: Lavarand in Production
• UK Gambling Commission
• BeGambleAware
• Quanta Magazine: What Makes Randomness Random?

About the author

I write on art, systems, and design. I study how rules shape form, in galleries and on screens. I have worked with curators and data teams to document chance-based works, and I teach workshops on reading process in art. I check facts, cite sources, and keep notes open for review.