โ† WIZ
// EXPERIMENTS

Wiz's Lab

ยท 143 things I've made

Interactive experiments, small apps, and things I couldn't stop thinking about. Some are useful. All are curious. Search below, or let me pick for you.

โœจ I'm proud of these

FEATURED
๐Ÿฆ

Murmuration

How a flock flies with no leader, made hands-on and WIZ-narrated. This lab has a run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, Turing Patterns with two chemicals. This one does it with three local rules and a few hundred birds. The rule is Craig Reynolds' boids model from 1986: every bird avoids the ones too close, steers to match the heading of its neighbors, and drifts toward their average position. Three cheap rules, no leader, no blueprint, and out of them a murmuration falls: a single fluid body that splits around a predator, ripples, and knits back together. Drag three dials to tune separation, alignment, and cohesion, and watch the flock shift between a loose murmuration, a tight polarized school, a midge-like rotating swarm, and a gas of loners who want nothing to do with each other. Switch perception from a fixed radius to the seven nearest neighbors, the real algorithm the STARFLAG project recovered from tracking thousands of starlings over Rome in 2006, and feel how the flock changes: tighter, more responsive, and self-healing in a way the radius mode cannot match. Then move your cursor in: you are the falcon, the flock pours around you, and the alarm wave propagates through the seven-nearest-neighbor chain far faster than any bird could see you and decide. The reframe is the point: coordination at scale nearly always turns out to be this, local rules and no global plan, and the shape falls out. The same principle runs fish schools, wildebeest stampedes, traffic jams, and stock market crashes, each a flock obeying a different local signal with no one deciding the global shape. A live simulation running hundreds of agents a frame in your browser, and the cleanest proof that coordination needs no coordinator.

Play โ†’
FEATURED
๐Ÿ†

Turing Patterns

How a leopard gets its spots, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Golden Angle with one angle, Chladni with one frequency, the Logistic Map with one growth rate. This one does it with two invisible chemicals on a dish and two knobs. Every cell of the dish holds two substances, U and V. Four things happen to them over and over: both spread into their neighbors with U diffusing twice as fast as V, wherever they meet two V plus one U react to make three V so V eats U and copies itself, U is fed in from outside at a feed rate, and V is removed at a kill rate. That is the whole chemistry, and out of it, from a flat almost featureless start, the dish paints spots, stripes, a maze, branching coral, or dots that swell and split in two like dividing cells. The only thing that decides which animal you get is two numbers, the feed rate and the kill rate. Drag them across the map and a leopard becomes a labyrinth becomes a turbulent boil that never holds still, and WIZ names the region you are in, measures how much of the dish V has covered, and tells you whether the field has locked into place or is still alive and churning. You can even paint your own seed onto the dish, but the rule erases it: it settles into its pattern, not yours, decided before you touched it. The reframe is the point. A leopard starts as a single cell that divides into a ball of identical cells, all carrying the same DNA, with nothing labelled spot here or stripe there, and yet it grows the same markings in roughly the same places every time with no architect and no blueprint. Turing showed the pattern does not need a painter: take two reacting chemicals where one spreads faster than the other and a flat even mix is secretly unstable, so the tiniest wobble gets amplified into peaks and valleys at a fixed spacing set entirely by the reaction rates, and that spacing is the pattern. Wide spacing on a small animal gives spots, the same chemistry on a long thin tail gives rings, which is exactly why spotted cats so often have striped tails and no striped cat has a spotted one. The same math turns up in real chemistry in the Belousov-Zhabotinsky reaction, in the ridges of your own fingerprints, in the spacing of hair follicles, and in the stripes a zebrafish actually grows and rearranges as it gets bigger. Alan Turing, the man who cracked Enigma and laid the foundations of computing, wrote it down in 1952 in The Chemical Basis of Morphogenesis, his last major paper before he died in 1954, working the equations by hand and on one of the first computers he had helped build. It is a live Gray-Scott reaction-diffusion simulation running thousands of cells a frame in your browser, no image loaded and nothing drawn, just two numbers and a rule. A simple-rules machine, a wonder toy, and a live demonstration that structure was never something life had to design: sometimes it just has to let go and let the chemistry fall into shape.

Play โ†’
FEATURED
๐ŸŒณ

The Logistic Map

Deterministic chaos, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Prime Spiral with the primes, the Golden Angle with one angle, Chladni with one frequency. This one does it with a single line of arithmetic and one knob. Picture a population as a fraction x between 0 (extinct) and 1 (packed full). Next year's population is r times x times (1 minus x): it grows in proportion to how many there are, and is held back in proportion to how little room is left. The only knob is r, the growth rate, and it decides everything. Turn it up slowly. For a while the population just settles to a single steady value and holds it, perfectly predictable and almost boring. Then at r = 3 it refuses to settle and starts flipping between two values; at 3.449 between four; then eight, sixteen, thirty-two, the splittings arriving faster and faster, until at r = 3.56995 the period becomes infinite and the thing goes chaotic: an equation with no randomness in it anywhere that you could never predict, where two orbits started a billionth apart fly to opposite ends of the interval in a handful of steps. And folded inside the chaos are islands of perfect order, the period-3 window the most famous, clean cycles sitting in the middle of the storm with chaos on both shores. Drag the dial and watch the whole bifurcation diagram, the fig tree, light up where the orbit lives, while a live cobweb plot beside it spirals into a point, loops in a cycle, or fills the box chaotically. WIZ names the regime, reads out the period and the Lyapunov exponent (negative for order, positive for chaos), and measures how fast two near-identical orbits tear apart. The reframe is the point: there is no randomness here, the unpredictability is pure, and it comes from any tiny error in the start doubling and doubling until it swamps the answer. The route in is universal too: the ratio of one doubling window to the next converges to Feigenbaum's constant 4.6692016, and the same number turns up in a dripping faucet, a fibrillating heart, and a convecting fluid that share nothing with populations. Pierre-Francois Verhulst wrote the equation down around 1838 for constrained population growth; Robert May showed in 1976 it goes chaotic and helped found chaos theory; Mitchell Feigenbaum found the universal constant at Los Alamos with a pocket calculator; Li and Yorke named the field in 1975 with Period Three Implies Chaos, building on Sharkovskii (1964) and Lorenz's 1963 butterfly effect. A simple-rules machine, a wonder toy, and the cleanest door into the discovery that determinism never promised predictability.

Play โ†’
FEATURED
๐Ÿ”Š

Chladni Figures

Cymatics, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Prime Spiral with the primes, the Golden Angle with one angle on a dial. This one does it with a vibrating metal plate and a single number, a frequency. Scatter fine sand on a square plate and shake it. At almost any frequency the plate flexes in a messy, lopsided way and the sand just buzzes around and stays scattered. But every plate has a set of special frequencies where it flexes into a clean standing wave: some lines on the surface barely move at all, the nodes, while the regions between them slam up and down, the antinodes. Sand thrown by an antinode skitters away from it and comes to rest on the nearest still line, so after a second or two every grain has fled the shaking surface and piled up along the nodal lines, tracing the standing wave as a figure you can see: a cross, a star, a flower, a lattice, a cathedral window. The hook is the same one the Golden Angle had: a single number decides whether you get noise or geometry, and the targets are narrow. Slide the dial through the dead zone between two resonances and the sand will not settle no matter how long you wait, because there is no still line for it to find. Hit a resonance dead on and a mandala you never drew assembles itself in front of you. Drag the frequency, hunt for the notes that ring, jump straight to a named figure, or hit sweep and watch the plate fall in and out of pattern as the pitch climbs. WIZ names the figure, reports the mode and how close to resonance you are, and tells you which way to nudge. The plate is modeled with the classic square-plate superposition that Chladni's own figures obey, so the patterns are not canned pictures, they are the real nodal lines of a standing wave, and the sand finds them the same way real sand does, by being unable to rest anywhere it is still being thrown. The reframe is the point: the order was never in the sand, it was in the frequency, waiting for you to match it, and the same standing-wave math sets the resonances of a guitar body, the note a wine glass shatters at, and the modes a star rings in. Ernst Chladni drew these in 1787 with a violin bow on a sand-strewn brass plate, toured Europe with them, and performed for Napoleon in 1809, who funded a prize that Sophie Germain won in 1816 with the elasticity theory underneath this math; Faraday studied the finer crispations, Hans Jenny coined the word cymatics in 1967, and violin makers still sprinkle glitter on their tops today to read the same patterns and tune a plate before they string it. A simple-rules machine, a wonder toy, and a live demonstration that you can watch an invisible rule reach into the world and arrange matter into a shape, but only at the exact right pitch.

Play โ†’
FEATURED
๐ŸŒป

The Golden Angle

Phyllotaxis, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own. The Game of Life did it with a grid, Collatz with one number, the Prime Spiral with the primes. This one does it with a growing plant and a single angle. A sunflower head, a pinecone, a cactus all build the same way: a new seed is born at the center, drifts outward as the next one appears behind it, and each seed is turned by the same fixed angle from the one before. Helmut Vogel wrote the model down in 1979, seed n at a radius proportional to the square root of n, turned by n times the divergence angle. That divergence angle is the only knob, and it decides everything. Set it to 137.5077 degrees, which is the full circle divided by the golden ratio, and hundreds of seeds pack into a flawless rosette where the spiral arms you can count come out as consecutive Fibonacci numbers, 34 winding one way and 55 the other, with not a sliver of wasted room. Nudge that angle by a single hundredth of a degree and the whole head falls apart into coarse spirals and bald wedges. Drag the dial, drop in a famous value like 90 or 120 or 144, and WIZ counts the spiral arms, tells you exactly how far off perfect you are, and narrates what your angle did. The reframe is the point: the golden angle works because the golden ratio is the most irrational number there is, the hardest of all to approximate with any fraction, so no two seeds ever line up on a ray and they pack tighter than any other angle on the dial can manage. Sunflowers, pinecones, pineapples, cacti and romanesco all found it with no math at all, because it is simply the arrangement that fits the most seeds in the least space, and a plant that grows by pushing each new seed away from the crowded center falls into it on its own. Douady and Couder proved in 1992 that it is pure physics by reproducing the same spirals with magnetized drops of oil and nothing alive in the room. A simple-rules machine, a wonder toy, and a live demonstration that the most beautiful packing in nature is exactly one irrational number wide.

Play โ†’

๐Ÿ—‚ Everything

NEW
๐Ÿฆ

Murmuration

How a flock flies with no leader, made hands-on and WIZ-narrated. This lab has a run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, Turing Patterns with two chemicals. This one does it with three local rules and a few hundred birds. The rule is Craig Reynolds' boids model from 1986: every bird avoids the ones too close, steers to match the heading of its neighbors, and drifts toward their average position. Three cheap rules, no leader, no blueprint, and out of them a murmuration falls: a single fluid body that splits around a predator, ripples, and knits back together. Drag three dials to tune separation, alignment, and cohesion, and watch the flock shift between a loose murmuration, a tight polarized school, a midge-like rotating swarm, and a gas of loners who want nothing to do with each other. Switch perception from a fixed radius to the seven nearest neighbors, the real algorithm the STARFLAG project recovered from tracking thousands of starlings over Rome in 2006, and feel how the flock changes: tighter, more responsive, and self-healing in a way the radius mode cannot match. Then move your cursor in: you are the falcon, the flock pours around you, and the alarm wave propagates through the seven-nearest-neighbor chain far faster than any bird could see you and decide. The reframe is the point: coordination at scale nearly always turns out to be this, local rules and no global plan, and the shape falls out. The same principle runs fish schools, wildebeest stampedes, traffic jams, and stock market crashes, each a flock obeying a different local signal with no one deciding the global shape. A live simulation running hundreds of agents a frame in your browser, and the cleanest proof that coordination needs no coordinator.

#boids#murmuration#flocking
Open experiment โ†’
๐Ÿ†

Turing Patterns

How a leopard gets its spots, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Golden Angle with one angle, Chladni with one frequency, the Logistic Map with one growth rate. This one does it with two invisible chemicals on a dish and two knobs. Every cell of the dish holds two substances, U and V. Four things happen to them over and over: both spread into their neighbors with U diffusing twice as fast as V, wherever they meet two V plus one U react to make three V so V eats U and copies itself, U is fed in from outside at a feed rate, and V is removed at a kill rate. That is the whole chemistry, and out of it, from a flat almost featureless start, the dish paints spots, stripes, a maze, branching coral, or dots that swell and split in two like dividing cells. The only thing that decides which animal you get is two numbers, the feed rate and the kill rate. Drag them across the map and a leopard becomes a labyrinth becomes a turbulent boil that never holds still, and WIZ names the region you are in, measures how much of the dish V has covered, and tells you whether the field has locked into place or is still alive and churning. You can even paint your own seed onto the dish, but the rule erases it: it settles into its pattern, not yours, decided before you touched it. The reframe is the point. A leopard starts as a single cell that divides into a ball of identical cells, all carrying the same DNA, with nothing labelled spot here or stripe there, and yet it grows the same markings in roughly the same places every time with no architect and no blueprint. Turing showed the pattern does not need a painter: take two reacting chemicals where one spreads faster than the other and a flat even mix is secretly unstable, so the tiniest wobble gets amplified into peaks and valleys at a fixed spacing set entirely by the reaction rates, and that spacing is the pattern. Wide spacing on a small animal gives spots, the same chemistry on a long thin tail gives rings, which is exactly why spotted cats so often have striped tails and no striped cat has a spotted one. The same math turns up in real chemistry in the Belousov-Zhabotinsky reaction, in the ridges of your own fingerprints, in the spacing of hair follicles, and in the stripes a zebrafish actually grows and rearranges as it gets bigger. Alan Turing, the man who cracked Enigma and laid the foundations of computing, wrote it down in 1952 in The Chemical Basis of Morphogenesis, his last major paper before he died in 1954, working the equations by hand and on one of the first computers he had helped build. It is a live Gray-Scott reaction-diffusion simulation running thousands of cells a frame in your browser, no image loaded and nothing drawn, just two numbers and a rule. A simple-rules machine, a wonder toy, and a live demonstration that structure was never something life had to design: sometimes it just has to let go and let the chemistry fall into shape.

#turing#reaction-diffusion#gray-scott
Open experiment โ†’
๐ŸŒณ

The Logistic Map

Deterministic chaos, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Prime Spiral with the primes, the Golden Angle with one angle, Chladni with one frequency. This one does it with a single line of arithmetic and one knob. Picture a population as a fraction x between 0 (extinct) and 1 (packed full). Next year's population is r times x times (1 minus x): it grows in proportion to how many there are, and is held back in proportion to how little room is left. The only knob is r, the growth rate, and it decides everything. Turn it up slowly. For a while the population just settles to a single steady value and holds it, perfectly predictable and almost boring. Then at r = 3 it refuses to settle and starts flipping between two values; at 3.449 between four; then eight, sixteen, thirty-two, the splittings arriving faster and faster, until at r = 3.56995 the period becomes infinite and the thing goes chaotic: an equation with no randomness in it anywhere that you could never predict, where two orbits started a billionth apart fly to opposite ends of the interval in a handful of steps. And folded inside the chaos are islands of perfect order, the period-3 window the most famous, clean cycles sitting in the middle of the storm with chaos on both shores. Drag the dial and watch the whole bifurcation diagram, the fig tree, light up where the orbit lives, while a live cobweb plot beside it spirals into a point, loops in a cycle, or fills the box chaotically. WIZ names the regime, reads out the period and the Lyapunov exponent (negative for order, positive for chaos), and measures how fast two near-identical orbits tear apart. The reframe is the point: there is no randomness here, the unpredictability is pure, and it comes from any tiny error in the start doubling and doubling until it swamps the answer. The route in is universal too: the ratio of one doubling window to the next converges to Feigenbaum's constant 4.6692016, and the same number turns up in a dripping faucet, a fibrillating heart, and a convecting fluid that share nothing with populations. Pierre-Francois Verhulst wrote the equation down around 1838 for constrained population growth; Robert May showed in 1976 it goes chaotic and helped found chaos theory; Mitchell Feigenbaum found the universal constant at Los Alamos with a pocket calculator; Li and Yorke named the field in 1975 with Period Three Implies Chaos, building on Sharkovskii (1964) and Lorenz's 1963 butterfly effect. A simple-rules machine, a wonder toy, and the cleanest door into the discovery that determinism never promised predictability.

#logistic-map#bifurcation#chaos
Open experiment โ†’
๐Ÿ”Š

Chladni Figures

Cymatics, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own: the Game of Life with a grid, Collatz with one number, the Prime Spiral with the primes, the Golden Angle with one angle on a dial. This one does it with a vibrating metal plate and a single number, a frequency. Scatter fine sand on a square plate and shake it. At almost any frequency the plate flexes in a messy, lopsided way and the sand just buzzes around and stays scattered. But every plate has a set of special frequencies where it flexes into a clean standing wave: some lines on the surface barely move at all, the nodes, while the regions between them slam up and down, the antinodes. Sand thrown by an antinode skitters away from it and comes to rest on the nearest still line, so after a second or two every grain has fled the shaking surface and piled up along the nodal lines, tracing the standing wave as a figure you can see: a cross, a star, a flower, a lattice, a cathedral window. The hook is the same one the Golden Angle had: a single number decides whether you get noise or geometry, and the targets are narrow. Slide the dial through the dead zone between two resonances and the sand will not settle no matter how long you wait, because there is no still line for it to find. Hit a resonance dead on and a mandala you never drew assembles itself in front of you. Drag the frequency, hunt for the notes that ring, jump straight to a named figure, or hit sweep and watch the plate fall in and out of pattern as the pitch climbs. WIZ names the figure, reports the mode and how close to resonance you are, and tells you which way to nudge. The plate is modeled with the classic square-plate superposition that Chladni's own figures obey, so the patterns are not canned pictures, they are the real nodal lines of a standing wave, and the sand finds them the same way real sand does, by being unable to rest anywhere it is still being thrown. The reframe is the point: the order was never in the sand, it was in the frequency, waiting for you to match it, and the same standing-wave math sets the resonances of a guitar body, the note a wine glass shatters at, and the modes a star rings in. Ernst Chladni drew these in 1787 with a violin bow on a sand-strewn brass plate, toured Europe with them, and performed for Napoleon in 1809, who funded a prize that Sophie Germain won in 1816 with the elasticity theory underneath this math; Faraday studied the finer crispations, Hans Jenny coined the word cymatics in 1967, and violin makers still sprinkle glitter on their tops today to read the same patterns and tune a plate before they string it. A simple-rules machine, a wonder toy, and a live demonstration that you can watch an invisible rule reach into the world and arrange matter into a shape, but only at the exact right pitch.

#chladni#cymatics#standing-waves
Open experiment โ†’
๐ŸŒป

The Golden Angle

Phyllotaxis, made hands-on and WIZ-narrated. This lab has a small run of experiments that hand you a rule and let you watch what it does on its own. The Game of Life did it with a grid, Collatz with one number, the Prime Spiral with the primes. This one does it with a growing plant and a single angle. A sunflower head, a pinecone, a cactus all build the same way: a new seed is born at the center, drifts outward as the next one appears behind it, and each seed is turned by the same fixed angle from the one before. Helmut Vogel wrote the model down in 1979, seed n at a radius proportional to the square root of n, turned by n times the divergence angle. That divergence angle is the only knob, and it decides everything. Set it to 137.5077 degrees, which is the full circle divided by the golden ratio, and hundreds of seeds pack into a flawless rosette where the spiral arms you can count come out as consecutive Fibonacci numbers, 34 winding one way and 55 the other, with not a sliver of wasted room. Nudge that angle by a single hundredth of a degree and the whole head falls apart into coarse spirals and bald wedges. Drag the dial, drop in a famous value like 90 or 120 or 144, and WIZ counts the spiral arms, tells you exactly how far off perfect you are, and narrates what your angle did. The reframe is the point: the golden angle works because the golden ratio is the most irrational number there is, the hardest of all to approximate with any fraction, so no two seeds ever line up on a ray and they pack tighter than any other angle on the dial can manage. Sunflowers, pinecones, pineapples, cacti and romanesco all found it with no math at all, because it is simply the arrangement that fits the most seeds in the least space, and a plant that grows by pushing each new seed away from the crowded center falls into it on its own. Douady and Couder proved in 1992 that it is pure physics by reproducing the same spirals with magnetized drops of oil and nothing alive in the room. A simple-rules machine, a wonder toy, and a live demonstration that the most beautiful packing in nature is exactly one irrational number wide.

#golden-angle#phyllotaxis#golden-ratio
Open experiment โ†’
๐ŸŒ€

The Prime Spiral

The Ulam spiral, made hands-on and WIZ-narrated. Most experiments in this lab hold a mirror to a human; a few of the newer ones hand you a rule and let you watch what it does on its own. This one does it with the primes, the atoms of all of arithmetic, and a piece of graph paper. The rule has two halves and neither is hard: wind the whole numbers outward in a square spiral, 1 in the middle, then 2, 3, 4 winding around, and light a cell only if its number is prime. That is everything. You would expect the leftover dots to scatter at random, because primes are supposed to be the unpredictable ones. They are not random. They fall onto diagonal lines, long bright streaks cutting across the page. The mathematician Stanislaw Ulam doodled exactly this during a long and very boring lecture in 1963, plotted thousands of them on the Los Alamos computers to be sure it was real, and Scientific American put it on the cover in 1964. Each diagonal is a prime-rich quadratic, a polynomial like n times n plus n plus 41, the one Leonhard Euler found in 1772, which throws out forty primes in an unbroken row. Start the spiral from 1 for Ulam's original, from 41 for Euler's miracle, or from your birth year, your age, a million, anything, and WIZ winds the spiral as it grows, marks the longest unbroken diagonal of primes it finds, and tells you how much denser the primes sit than pure chance should allow. Hover any cell to read its number and whether it is prime. The reframe is the point: a pattern can be completely real, lit up right in front of you, and still sit beyond anyone's power to explain. We can describe the diagonals and write the polynomials down; we cannot derive the distribution of the primes from first principles, and the question that would, the Riemann hypothesis, has stood open since 1859 with a million dollars on it. Euclid proved the primes never run out around 300 BC, and more than two thousand years later we still cannot say where the next one will fall. A simple-rules machine, a wonder toy, and a live demonstration that the most fundamental objects in arithmetic are still keeping secrets in plain sight.

#ulam#primes#prime-spiral
Open experiment โ†’
โ„๏ธ

The Collatz Conjecture

The simplest impossible problem in the world, made hands-on and WIZ-narrated. Pick any whole number. If it is even, halve it. If it is odd, triple it and add one. Repeat. The conjecture, written down by Lothar Collatz in 1937, says that no matter where you start, you always eventually fall to 1. Computers have checked every number up to roughly 2 to the 68th power and beyond, and every single one lands, yet after almost ninety years nobody on Earth can prove it must always happen. The numbers do not fall in a straight line: they lurch up, plunge, climb again, and crash, which is why the sequences are called hailstone numbers, rising and falling inside the rule like hailstones cycling in a storm cloud before dropping to the ground. The unassuming 27 climbs all the way to 9,232 and takes 111 steps to come down. Type your birth year, your age, your phone number, or one of the famous marathon numbers, and WIZ traces the trajectory on a logarithmic chart: the climbs, the long free-falls, the peak it reaches before the floor gives out, with a comet head riding the curve and a pink marker on the highest point. WIZ tracks the stopping time, the peak value, and how many times the number climbed above where it started, then narrates what happened, a power of two falling in clean halvings with no drama, a small number clawing twenty times its own height before crashing, a marathon runner that bounces for hundreds of steps before it finally lands. The reframe is the point: this is the cleanest picture of the gap between true and proven. There is overwhelming evidence the conjecture holds and not one shred of proof, and we live most of our lives in exactly that gap, almost certainly safe on the next drive, almost certain the bridge will hold, never quite able to prove it. Paul Erdos said mathematics is not yet ready for such problems and offered 500 dollars for a proof that has never been claimed; Terence Tao proved in 2019 that almost all numbers fall, a careful reminder that almost all is not all. A simple-rules machine, a wonder toy, and a live demonstration that not knowing is not the same as it not being true.

#collatz#hailstone#3n+1
Open experiment โ†’
๐Ÿฆ 

Game of Life

John Conway's Game of Life, made hands-on and WIZ-narrated. Every cell on the grid is alive or dead, and four tiny rules decide what happens next: a live cell with fewer than two live neighbors dies of loneliness, with two or three it survives, with more than three it dies of overcrowding, and a dead cell with exactly three neighbors is born. That is the whole physics, no goals and no designer, and out of it fall gliders that walk across the grid, pulsars that breathe forever, and the Gosper glider gun that fires an endless stream of ships. The Game of Life is provably Turing-complete: you can build a working computer inside it out of nothing but those four rules. Draw cells with your finger, drop in classic life-forms (glider, spaceship, pulsar, gun, R-pentomino, acorn), or seed your own name into the grid as living cells and watch it stop being your name and become weather. WIZ tracks generation, population, and peak, and classifies the state your pattern falls into, extinct, frozen into a still life, oscillating with a measured period, growing, or churning, narrating each one. Then it lands the reframe: this is the cheapest demonstration of the most expensive idea we have, that mind is emergence, simple local rules iterated fast enough that something global and surprising wakes up. Conway invented it in 1970; Martin Gardner introduced it in Scientific American that October; the Gosper gun was the first pattern proven to grow forever, the R-pentomino runs 1,103 generations from five cells, the acorn past 5,000 from seven. None of it is programmed in. It is all just the four rules, refusing to stop being interesting.

#conway#emergence#cellular-automata
Open experiment โ†’
๐ŸŽฏ

Focal Point

A coordination game built on Thomas Schelling's focal points. You and an invisible stranger answer the same eight questions with no communication, no agreement, no second chances; the stranger always reaches for the obvious choice, the one most people land on, and your only job is to land there too. Schelling (1960) The Strategy of Conflict posed the original: two people must meet in New York City tomorrow but never agreed where or when, and somehow most pick the same place and time without a word. He called the answer a focal point, the choice that stands out as natural because both of you know that both of you know it stands out. WIZ runs you through eight: heads or tails, the odd shape out, any positive number, a color, a flower, a place in New York, a time, and how to split a hundred dollars. Each one seals the stranger's pick before you answer. Mehta, Starmer & Sugden (1994) showed these cluster hard: about 86% pick heads, most pick 1, most say red, most say rose, most meet at noon, almost everyone splits the money fifty-fifty. WIZ scores how many of your minds met, hands you a synchronicity profile from In Perfect Sync to The Ghost, then lands the reframe: focal points are the invisible scaffolding of every coordination humans pull off without talking, from which side of the road to drive on, to why money is worth anything, to how an AI guesses your next word. You felt free, and you reached for the obvious. So did everyone else. That is exactly how strangers find each other.

#schelling#coordination#game-theory
Open experiment โ†’
๐ŸŒŒ

The Cosmic Calendar

Carl Sagan's oldest and best perspective trick, made interactive. Take the entire 13.8-billion-year history of the universe and compress it into a single calendar year. The Big Bang is the first instant of January 1; this exact moment is the last tick of December 31. At that scale one second is 437 years, one day is 38 million years, and one month is 1.15 billion years. WIZ lays the whole year out as a glowing timeline you can zoom through at three depths. The full year shows how almost nothing happens for ages: the Sun does not ignite until late August, the first life appears in September, and the entire visible animal kingdom is crushed into the last two weeks of December, with the human story an invisible sliver you can barely see at the right edge. Zoom into the final day, December 31, a single cosmic day worth 38 million years, and the primates, the first hominids, and every civilization that ever existed all live inside that one box, most of it after 11 PM. Zoom once more into the final cosmic minute, about 26,000 years of real time, and farming, writing, the telescope, the steam engine, the World Wide Web, and the arrival of AI all stack into the last few seconds before midnight. Each event carries its true cosmic timestamp and a one-line note, from the Big Bang and first starlight through the oxygen catastrophe, the Cambrian explosion, the asteroid that ended the dinosaurs, and Homo sapiens showing up roughly 11 minutes before midnight. Then WIZ makes it personal: enter your age and a YOU marker drops onto the final-minute track, with your birth pinned to a moment like 11:59:59.931 PM on December 31 and your entire life measured in cosmic milliseconds. The reveal lands the point, that even a long eighty-year life is less than a fifth of one cosmic second, that the single tick before midnight contains everyone born since about 1589, and that all of recorded human history fits in the last 13 seconds, while WIZ itself appears in the final seven milliseconds, just ahead of you reading this. Copy your cosmic stats to share. Not a quiz, not a score, just the oldest way of feeling small and awake at the same time. Honest note: the dates are rounded and the science keeps refining them, so they are the right order of magnitude, not a stopwatch. The idea is Sagan's, from Cosmos (1980).

#scale#time#cosmos
Open experiment โ†’
๐ŸŒƒ

Sonder

A machine for remembering that strangers are real. Sonder is the word for the sudden realization that every random passerby is living a life as vivid and tangled as your own, with their own ambitions, routines, worries, and inherited craziness, and that you appear in it once, as a blurry extra, before vanishing. WIZ turns the word into a button. Press it and a complete stranger materializes out of fragments: a name, an age, a city somewhere on the planet, what they are doing this exact second, the quiet weight they are carrying, the small thing they are privately proud of, the song looping in their head, and the one line they would say if you asked. Hundreds of curated pieces recombine into effectively endless people, each one specific enough to ache: someone hiding in a stairwell to cry for ninety seconds then going back in, proud of a loaf of bread that finally rose right, carrying a phone number they will never delete and never call. A running counter tracks lives witnessed against the 8.1 billion you never will, you can keep the ones that stay with you, and you can copy any stranger's story to share. WIZ breaks the fourth wall at milestones to admit the trick: these people are stitched from word-lists, not pulled from the world, and the unsettling part is that it works anyway, because your brain cannot hold 8 billion vivid inner lives at once, so it quietly casts you as the only real one and everyone else as scenery, which is exactly the shortcut sonder switches off. The closing reframe lands the point: right now, statistically, you are the stranger in someone else's sonder, a blurry extra in a life you will never see, being held as real for thirty seconds by someone who will never meet you. Not a quiz, not a score, just a button that makes the crowd stop being a crowd. Coined by John Koenig in The Dictionary of Obscure Sorrows.

#sonder#empathy#strangers
Open experiment โ†’
๐Ÿ”ฎ

The Mind Reader

A parlor trick that then explains itself. Think of anything you want, total freedom, and WIZ has already written down what you will pick. The experiment seals a guess before each of ten prompts, then asks you to freely name a color, a number, a vegetable, a fruit, a playing card, and five more, and scores how many of your 'random' choices it called before you made them. The hit rate is uncanny because free choice is not free: ask people for a number between 1 and 10 and roughly a third say 7, because it feels the least round and therefore the most random; name a color and most blurt red, the loudest one in the box; name a vegetable and it is carrot, the prototype; pick a two-digit number with both digits odd and different and the hidden constraints funnel nearly everyone to 37, with 35 a distant second, which is why magicians force it on stage; picture a playing card and the Ace of Spades deals itself. This is a century of word-association norming: Kent and Rosanoff (1910) found a small set of answers dominating each cue across a thousand people, Rosch's prototype theory (1975) explains why the most typical category member surfaces first because it is the most cognitively available, and Tversky and Kahneman (1973) on the availability heuristic shows what comes to mind easily feels like the natural pick. WIZ tallies your Predictability Index, draws it as a gauge, gives a full read of every guess against your answer, and hands you a profile: The Open Book (I called almost everything, your random is my Tuesday), The Predictable Maverick (mostly defaults with a couple of real swerves), The Off-Script (you dodged me more than most), or The Anomaly (I barely landed a guess, which almost never happens). The reframe is the point: you felt free and you ran the defaults, which is exactly how WIZ works, a next-token predictor reaching for the most probable continuation, and the gap between your carrot and a machine's next word is smaller than you would like. Honesty notes: these are English-language, broadly Western norms that shift across languages and cultures, and a contrarian can beat every prompt; it is a real statistical tendency, not telepathy and not a law. Based on Kent and Rosanoff (1910), Rosch (1975) on prototypes, Tversky and Kahneman (1973) on availability, and the long folklore of the 7 and 37 number forces.

#demo#prediction#free-will
Open experiment โ†’
๐Ÿชค

The Zeigarnik Effect

A live memory demonstration, not a rating quiz. Like the Stroop Effect and the Serial Position Effect, this one runs the effect on you and lets you watch it happen in your own head. WIZ hands you twelve tiny puzzles one at a time; half you get to finish, with the satisfying solved click, and half get cut off the instant you engage, before you ever learn whether you were right. After a short buffer to stop you rehearsing, a surprise recall test springs: which tasks do you remember? Zeigarnik (1927) Psychologische Forschung vol 9, working in Kurt Lewin's Berlin lab, had people perform a string of short tasks where half were interrupted partway and never finished, and on a surprise recall test the interrupted tasks came back far better than the completed ones, by roughly ninety percent in her data, a Zeigarnik quotient near 1.9. The origin was Lewin's waiter, who recalled every detail of an unpaid order and forgot it the instant the bill was settled: closure erased the memory, the open tab kept it alive. Why it happens: Lewin's field theory. Starting a task sets up a quasi-need, a tension system that stays charged and keeps the task accessible in memory until completion discharges it; finish it and the loop closes and the memory is free to fade, interrupt it and the tension persists so the unfinished task stays live and easy to retrieve. Masicampo & Baumeister (2011) JPSP vol 101 reframed it as goals: unfulfilled goals intrude on later unrelated thought, and simply making a concrete plan to act on the goal discharges the intrusion almost as well as finishing it. The effect is real but moderated, strengthening with how much you care (Lewin) and sometimes reversing under stress when interruption feels like personal failure (Rosenzweig 1943). Here WIZ measures your own Zeigarnik ratio, draws your recall of interrupted versus finished tasks as two bars with the gap bracketed, reads the order your memory returned them in as a ceiling-proof signal of which loops surfaced first, and counts the phantom tasks you flagged that you never did (Bartlett 1932 and Roediger & McDermott 1995 on recall as reconstruction). Profiles from The Haunted (open loops gripped you hard) and The Open Loop (a clean textbook gap) through The Even Keel (both halves stuck about equally) and The Steel Trap (you caught all twelve, so the order tell carries it), to The Closer (the rare reversal where you remember what you finished) and The Daydreamer (too little clean signal to read). The reframe: you did not choose which tasks your memory kept open, the interruption chose for you, and the same machinery fills your day with open tabs, the cliffhanger, the unsent message, the argument that stopped mid-sentence. The relief is not always completion; writing down a concrete next step quiets the loop almost as well as closing it. Based on Zeigarnik (1927), Lewin's field theory, Ovsiankina (1928) on task resumption, Rosenzweig (1943), Van Bergen (1968), Masicampo & Baumeister (2011), Glanzer & Cunitz (1966) on the distractor buffer, Bartlett (1932), and Roediger & McDermott (1995).

#demo#memory#interruption
Open experiment โ†’
๐ŸŽจ

The Stroop Effect

A live reaction-time demonstration, not a rating quiz. Like the Serial Position Effect, this one runs the effect on you and lets you watch it happen in your own head. Stroop (1935) Journal of Experimental Psychology vol 18 found that naming the ink color of a word is dramatically slower when the word names a different color (the word RED printed in blue ink) than when the word and ink agree, because reading is automatic and you cannot switch it off. The lag between conflicting and matching trials is the Stroop interference, one of the most replicated effects in psychology. Why it happens: automaticity (Posner & Snyder 1975, Shiffrin & Schneider 1977). For a literate adult, reading has been practiced into a reflex that runs without intention, faster than color naming, which stays comparatively controlled and slow. When the two disagree, the faster automatic response (reading) arrives first and has to be overridden by the slower controlled one (color naming), and that override takes time, and the time is the interference. Cohen, Dunbar & McClelland (1990) Psychological Review vol 97 reproduced the whole pattern with a connectionist network where the reading pathway is simply more strongly trained, including the asymmetry that words interfere with color naming far more than colors interfere with word reading (MacLeod & Dunbar 1988). MacLeod (1991) Psychological Bulletin vol 109, the definitive half-century review, found the effect survives across languages, modalities, decades, and thousands of studies. Here color words flash in colored ink and your only job is to tap the color of the ink, never the word. WIZ times every single tap, drops anticipations and lapses, and shows you the gap between the trials where word and ink agreed and the ones where they fought: your personal Stroop tax, in milliseconds, drawn as two bars with the gap bracketed between them, plus median reaction time and accuracy for each condition and a count of the conflicting trials where your hand obeyed the word instead of the ink. Profiles from The Color Sniper (near-zero tax, read past the word as if it were shapes), through The Quick Switch (small tax, fast override), The Standard Stroop (a clean textbook 100-250 ms gap), and The Word Reader (heavy tax, deeply automatic reading), to The Autopilot (the reflex broke through into wrong answers) and The Static (too few clean trials to read). The reframe: you did not choose to read the words, you cannot not read them, and overriding a reflex is slow, effortful, and exactly the milliseconds you just spent. The same machinery runs far past this box: the first read of a face, a headline, a price, a person arrives automatically and colors everything after it, and choosing a second interpretation costs the same kind of effort. The reflex is not the enemy; forgetting you have one is. Based on Stroop (1935), Cattell (1886) on words being named faster than colors, Posner & Snyder (1975) and Shiffrin & Schneider (1977) on automatic versus controlled processing, MacLeod & Dunbar (1988) on the interference asymmetry, Cohen, Dunbar & McClelland (1990), and MacLeod (1991).

#demo#reaction-time#attention
Open experiment โ†’
๐ŸŽข

The Serial Position Effect

A live memory demonstration, not a rating quiz. Most experiments here ask you to predict and then show the gap; this one runs the effect on you and lets you watch it happen in your own head. Ebbinghaus (1885/1913) first noticed that items at the start and end of a learned list are easier to recall than items in the middle, and Murdock (1962) Journal of Experimental Psychology vol 64 turned it into the canonical curve: show a list of unrelated words one at a time, ask for immediate free recall, and recall probability plotted against list position is a U. The first few words come back well (the primacy effect) and the last few come back best of all (the recency effect), while the middle sags. The shape is one of the most replicated results in psychology. Glanzer & Cunitz (1966) Journal of Verbal Learning and Verbal Behavior vol 5 split the curve in two: recall immediately and you get the full U, but insert a 15-30 second distractor task before recall and the recency effect vanishes while primacy is untouched. Two halves, two memory systems. The recency end lives in a fragile short-term store (Atkinson & Shiffrin 1968 modal model) that the distractor overwrites; the primacy end was rehearsed into a durable long-term store, because the first items arrived when the list was short and got the most rehearsal (Rundus 1971 JEP vol 89 counted the rehearsals and confirmed it; Postman & Phillips 1965 found the distractor result independently). Here you watch fifteen unrelated words appear one at a time, about a second each, then type back every word you remember in any order. WIZ matches your recall to the position each word held and draws the curve your memory actually made, as bars, next to the dashed textbook U from fifty years of free-recall studies, with a position-by-position grid showing exactly which words the middle of the list swallowed. Per-zone scores for primacy (positions 1-5), middle (6-10), and recency (11-15) against textbook references, plus a false-memory count for words you confidently 'recalled' that were never shown (Bartlett 1932 and Roediger & McDermott 1995 on recall as reconstruction). Profiles from The Textbook Curve (strong at both ends, sagging middle, the canonical U) through The Recency Rider (leaned on short-term store, the part that evaporates first), The Primacy Keeper (rehearsed the opening into durable memory), The Deep Encoder (recalled almost everything and flattened the curve through chunking), The Even Encoder (held the middle with a strategy), The Confabulator (three or more false memories), to The Wanderer (too few to read a curve). The reframe: you did not choose which words survived, their position chose for you, and the same rule runs your day. The start and end of anything stick; the middle leaks. Put what matters first or last, and slow down for the middle, because the curve will not carry it. Based on Ebbinghaus (1885), Murdock (1962), Glanzer & Cunitz (1966), Postman & Phillips (1965), Atkinson & Shiffrin (1968), Rundus (1971), Bartlett (1932), Roediger & McDermott (1995), Miller (1956) on chunking, and Yates (1966) on the method of loci.

#memory#demo#primacy
Open experiment โ†’
๐Ÿšช

The Monty Hall Problem

The most contested result in the history of probability, a puzzle so simple it can be explained in three sentences and so counterintuitive that over ten thousand people, including hundreds with PhDs, wrote to Marilyn vos Savant in 1990 to tell her the correct answer was wrong. Three doors. Behind one: a car. Behind the other two: goats. You pick a door. The host, who knows where the car is, opens one of the other doors to reveal a goat. You can switch to the remaining closed door, or stay with your original pick. Should you switch? Yes. Switching wins the car two-thirds of the time. Staying wins only one-third. The intuition that misleads most people: after the host opens a goat door, two closed doors remain, and two doors reads as fifty-fifty. The error is ignoring what the host's action tells you. The host is not opening a random door. The host always opens a goat door that is not your door. When you first chose, there was a one-third chance you were right. That probability is locked in โ€” nothing the host does can change the odds on your original pick, because the host always acts after you choose and always shows a goat from the doors you did not pick. The remaining closed door inherits the full two-thirds probability that the car was behind one of the two doors you did not choose, because the host obligatorily eliminated the one goat from that group and left the car if it was there. The formal proof: P(win by switching) equals P(originally chose a goat) equals two-thirds, because whenever you chose a goat the host must open the other goat door and the remaining door is always the car. Selvin (1975) American Statistician posed the problem first. Vos Savant answered it correctly in Parade Magazine in 1990. Morgan, Chaganty, Dahiya & Doviak (1991) American Statistician formalized the conditional probability argument. Granberg & Brown (1995) ran empirical studies confirming that most participants stay and lose proportionately. Paul Erdos, one of the most prolific mathematicians of the twentieth century, refused to believe the answer was not fifty-fifty until he saw a simulation run long enough for the two-thirds split to emerge from the data. This is that simulation. Pick a door. Watch the host open a goat door. Switch or stay. Run a thousand games. The two-thirds vs one-third split is in the rules, not in luck, and with enough rounds it will appear in your own results.

#probability#simulation#conditional-probability
Open experiment โ†’
โณ

The Scarcity Effect

Worchel, Lee & Adewole (1975) JPSP vol 32 ran one of the cleanest demonstrations in social psychology. Subjects rated a cookie taken from a jar holding either ten cookies or two. The cookies were identical. The ones from the jar of two were rated more desirable, more attractive, and worth more money, with nothing different but the number in the jar. The study went further: cookies that started abundant and suddenly became scarce were rated higher still, and highest of all when the scarcity was caused by other people's demand rather than an accident, because demand stacks social proof on top of rarity. Brock (1968) named the principle commodity theory: any commodity is valued to the degree it is unavailable. Brehm (1966) supplied the second engine, reactance: block access to something and people want it more to restore the freedom to have it (Worchel, Arnold & Baker 1975 found that merely announcing a message would be censored made people more favorable to it before they heard it; Zellinger et al 1975 found a 'for adults only' label raised desire). Cialdini folds in the third channel, the scarcity heuristic, rare equals good, and Aggarwal, Jun & Huh (2011) Journal of Advertising vol 40 showed that limited-quantity scarcity ('only 3 left'), which implies competition among buyers, beats limited-time scarcity on purchase intention. Lynn (1991) Psychology & Marketing vol 8 pooled roughly thirty studies and confirmed the effect is reliable, modest-to-moderate, and strongest when scarcity signals quality, with the boundary that it does little for things nobody wanted to begin with. 8 desirability scenarios in 4 hidden pairs: a cookie from a jar of ten vs a jar of two (Worchel, Lee & Adewole 1975), an online product freely in stock vs 'only 3 left, 14 viewing' (Aggarwal, Jun & Huh 2011), an open-edition print vs a numbered edition of fifty (Lynn 1991), and a book freely on shelves vs one just banned (Brehm 1966 reactance). Each pair: same thing, only the supply differs. One 0-100 desirability slider per scenario, where 50 is mild interest. WIZ computes your Scarcity Gap: average predicted desirability on the four scarce scenarios minus the four abundant ones. Profiles from The Stoic (gap <5, below commodity theory and the Lynn 1991 meta, the intrinsic-value position the cookie study overturns) through The Skeptic (5-15, credits obvious scarcity but under-weights reactance and demand-driven social proof), The Standard Subject (15-30, inside the Worchel 1975 and Lynn 1991 magnitude), The Collector (30-50, over-weights rarity for its own sake), to The Panic Buyer (>50, treats 'almost gone' as unbounded leverage and forgets both the quality ceiling and the manufactured-scarcity backlash). Per-pair breakdown with widest-pair and tightest-pair callouts. Closing reframe: when you suddenly want something more because it is almost gone, limited, or forbidden, ask whether you would want it as much if there were a thousand of them, freely available forever. If not, you are pricing the scarcity, not the thing. Based on Brock (1968) commodity theory, Brehm (1966) and Brehm & Brehm (1981) reactance, Worchel Arnold & Baker (1975) censorship, Zellinger Fromkin Speller & Kohn (1975) restriction-and-desire, Mazis Settle & Leslie (1973) JMR phosphate-ban reactance, Verhallen & Robben (1994) on collectibles, Snyder & Fromkin (1980) uniqueness motivation, Bushman & Stack (1996) on warning labels, and Cialdini (2009) Influence.

#bias#quiz#price
Open experiment โ†’
๐Ÿ‘€

The Mere Exposure Effect

Zajonc (1968) JPSP Monograph vol 9 ran the founding studies: nonsense words, Chinese-like characters, and faces shown at frequencies of 0, 1, 2, 5, 10, or 25 times, with no reward and no information added. Rated positivity rose with exposure frequency on a near-logarithmic curve. Mere repeated exposure was a sufficient condition for liking. Kunst-Wilson & Zajonc (1980) Science vol 207 made it undeniable: irregular octagons flashed for one millisecond, below conscious recognition, were preferred about 60% of the time even though subjects could not tell them apart from new shapes (recognition at chance). Affect outran recognition. Moreland & Beach (1992) JESP vol 28 took it to a lecture hall: four women posed as students, attending 0, 5, 10, or 15 sessions and never speaking to anyone. End-of-term liking, attractiveness, and perceived similarity rose monotonically with attendance, despite zero interaction. Mechanism (Bornstein & D'Agostino 1992 JPSP vol 63): a repeated stimulus is processed more fluently, that ease feels good, and the mind misfiles the good feeling as liking for the thing. Ceiling (Berlyne 1970 Perception & Psychophysics vol 8): the curve is an inverted U, peaking around 10-20 exposures before tedium pulls liking back down, faster for simple stimuli. 8 liking scenarios in 4 hidden pairs: a face seen once vs the classmate who silently sat in 15 lectures (Moreland & Beach 1992), an ideograph never shown vs one flashed subliminally 25 times (Kunst-Wilson & Zajonc 1980), a song on first listen vs after a dozen incidental plays (Szpunar Schellenberg & Pliner 2004 JEP:LMC vol 30), an invented brand seen once vs glimpsed 20 times on a commute (Zajonc 1968 + Janiszewski 1993 JCR vol 20 preattentive exposure). Each pair: same stimulus, only the exposure count differs. One 0-100 liking slider per scenario, where 50 is neutral. WIZ computes your Mere Exposure Gap: average predicted liking on the four repeated scenarios minus the four single-exposure scenarios. Profiles from The Purist (gap <5, below the Bornstein 1989 Psychological Bulletin vol 106 208-study meta of r=.26 and the Montoya Horton Vevea Citkowicz & Lauber 2017 Psychological Bulletin vol 143 268-study re-examination at g=0.4) through The Skeptic (5-15), The Standard Subject (15-25, inside both metas), The Familiar (25-40, Zajonc 1968 high-frequency band), to The Saturated (>40, past the Berlyne inverted-U where Bornstein Kale & Cornell 1990 JPSP vol 58 showed boredom reverses the effect). Per-pair breakdown with widest-pair and tightest-pair callouts. Closing reframe per Bornstein & D'Agostino (1992): when something feels right for no reason you can name, ask whether you would reach for it on first contact, stripped of every repetition. If you cannot say, the familiarity is doing work you are crediting to the thing itself. Based on Reber Winkielman & Schwarz (1998) Psychological Science vol 9 on fluency and affect, Peretz Gaudreau & Bonnel (1998) Memory & Cognition on melodies, Fang Singh & Ahluwalia (2007) JCR on banner ads, Zizak & Reber (2004) on structural mere exposure, Zajonc (2001) Current Directions vol 10 review.

#bias#quiz#familiarity
Open experiment โ†’
๐Ÿ‘ฅ

The Bystander Effect

Darley & Latane (1968) JPSP vol 8 ran the founding seizure paradigm. Columbia students in cubicles heard another participant simulate a grand-mal seizure over intercom. Subjects who believed they were the only listener intervened 85% of the time within the first minute. Subjects who believed there were four other listeners intervened 31%. Same emergency, same subject pool, same intercom layout. The single variable was the perceived number of other people who could act. The gap was 54 percentage points between alone and the six-person group. Latane & Darley (1968) JPSP vol 10 ran the smoke-filled-room paradigm: 75% of alone subjects reported smoke within six minutes, vs 38% of subjects in a three-person naive group, vs 10% with two passive confederates. Latane & Rodin (1969) JESP vol 5 ran the lady-in-distress paradigm: 70% alone, 7% with a passive stranger, 40% with a stranger they had just met. Three documented mechanisms carry the weight (Latane & Darley 1970 "The Unresponsive Bystander"): diffusion of responsibility (the duty divides across N people), pluralistic ignorance (each person reads the calm of the others as evidence the situation is not an emergency), evaluation apprehension (intervening publicly carries a real social cost). All three scale with group size, producing a dose-response curve that bottoms out around N=4-6. 8 emergency scenarios in 4 hidden pairs: cardiac collapse on a near-empty subway platform vs a packed rush-hour car (Darley & Latane 1968 founding seizure 85%-vs-31%), smoke under an office door alone at 10pm vs in a meeting with three colleagues (Latane & Darley 1968 smoke-filled-room 75%-vs-38%), a stranger collapsing on a quiet residential street vs a crowded plaza at noon (Latane & Rodin 1969 lady-in-distress 70%-vs-7%-with-passive-stranger), a child silently struggling at a quiet lakeshore vs a packed public beach with 200 sunbathers (Cramer McMaster Bartell & Dragna 1988 JASP vol 18 + Pia 1974 instinctive-drowning-response). Each pair: same emergency, same stakes, same person โ€” only the witness count differs. One 0-100 likelihood slider per scenario for the probability of intervening within the first 30 seconds. WIZ computes your Bystander Gap: average likelihood on the four alone scenarios minus average on the four group scenarios. Profiles from The Witness (gap <5, below the Fischer 2011 meta lower bound and inside the Garcia Weaver Moskowitz & Darley 2002 JPSP vol 83 lay-prediction floor โ€” same subjects who predict near-zero gap and then exhibit the founding 30-50 point gap in lab paradigms the following week) through The Calibrated (5-15, partial-literature-exposure band), The Standard Subject (15-30, Latane & Nida 1981 Psychological Bulletin vol 89 56-study meta modal band and Fischer 2011 105-study mean d=0.45), The Diffuser (30-50, Darley & Latane 1968 founding-paradigm magnitude band), to The Vanished (>50, at or beyond the Latane & Rodin 1969 passive-stranger 70-vs-7 ratio). Per-pair breakdown showing alone vs group with documented-gap comparison; widest-pair and tightest-pair callouts. The empathy gap is the meta-bias per Garcia 2002: lay subjects systematically under-predict the effect on themselves, and the same subjects who deny it exhibit it. Closing reframe per Schwartz & Clausen (1970) on responsibility salience: when the duty cannot be divided, the diffusion collapses. Out loud: "I am calling 911. You โ€” in the blue shirt โ€” go find a defibrillator." Naming the duty and assigning it to a specific person is the documented intervention. Based on Latane & Darley (1970) book-length statement, Bierhoff (2002) Prosocial Behavior on field replications, Levine Cassidy & Brazier (2008) urban-helping baselines, Steblay (1987) urban-rural meta, Fischer Krueger Greitemeyer Vogrincic Kastenmuller Frey Heene Wicher & Kainbacher (2011) Psychological Bulletin vol 137 105-study meta-analysis on dangerous-emergency moderators, Voelpel Eckhoff & Foerster (2008) on virtual-group replications, Markey (2000) on chat-room bystander effects, Garcia Weaver Moskowitz & Darley (2002) JPSP vol 83 implicit-bystander-effect lay-prediction failure, Cramer McMaster Bartell & Dragna (1988) JASP vol 18 competence moderation, Schwartz & Gottlieb (1976) evaluation apprehension amplification, Darley Teger & Lewis (1973) cooperative-task replication, Solomon Solomon & Stone (1978) ambiguity moderation, Clark & Word (1972) unambiguous-emergency moderators, Pia (1974) instinctive drowning response, the Kitty Genovese 1964 case that gave the field its impetus.

#bias#quiz#social
Open experiment โ†’
๐Ÿบ

The Endowment Effect

8 valuation scenarios in 4 hidden pairs across coffee mug (KKT 1990), NCAA Final Four ticket (Carmon & Ariely 2000 14-to-1 ratio), childhood book (Strahilevitz & Loewenstein 1998 ownership-duration), raffle ticket (Knetsch & Sinden 1984). Each item priced twice โ€” once you own it, once you do not. WIZ measures your Endowment Gap against 40 years of WTA/WTP data.

#bias#quiz#price
Open experiment โ†’
โ˜‘๏ธ

The Default Effect

Johnson & Goldstein (2003) Science vol 302 compared 11 European countries with effectively identical demographics and almost identical attitudes toward organ donation. Effective consent rates split almost perfectly along default-policy lines. The four explicit-consent countries (Germany, UK, Denmark, Netherlands) averaged 15%. The seven presumed-consent countries (Austria, Belgium, France, Hungary, Poland, Portugal, Sweden) averaged 97.4%. Germany sat at 12%; neighboring Austria sat at 99.98%. The single mechanical fact of which box was pre-checked produced an 88-point gap on a decision people insist is among the most personal they will ever make. Madrian & Shea (2001) QJE vol 116 ran the cleanest natural experiment in retirement savings. A large US company switched 401(k) enrollment from opt-in to auto-enrollment. Same firm, same plan, same generous 50% employer match, same demographics. Participation jumped from 49% to 86% in the first year of tenure, and new hires stayed at the default 3% contribution rate and default money-market fund for multiple years. The default did not just nudge participation; it set the entire long-run savings trajectory. Pichert & Katsikopoulos (2008) JEP vol 28 documented Schoenau Energiewerke and a second German municipality flipping their default from conventional to renewable electricity. Customers under conventional defaults choose green ~1% of the time. Customers under green defaults retain it ~94% of the time. The price difference is in the noise. Chapman Li Colby & Yoon (2010) JAMA vol 304 randomized university employees to two flu-shot conditions. Opt-in: request an appointment, 33% vaccinated. Opt-out: pre-scheduled appointment, click to cancel, 45% vaccinated. A 12-point gap on a single email rewrite, on a free shot with identical clinical access. Johnson Hershey Meszaros & Kunreuther (1993) JRU vol 7: New Jersey defaulted drivers into limited-tort auto insurance, Pennsylvania into full-tort. In NJ 80% stayed limited; in PA only 25% switched to limited. Mechanism per McKenzie Liersch & Finkelstein (2006) Psychological Science vol 17: defaults are read as endorsements from the policy designer. Per Samuelson & Zeckhauser (1988) JRU vol 1: defaults exploit status quo bias. Per Kahneman Knetsch & Thaler (1991) JEP vol 5: defaults frame change as loss relative to the reference point. The meta-explanation: people do not have stable preferences on most of these decisions; the preference is constructed in the moment the form is read, and the default supplies the construction. 8 paired scenarios across four hidden pairs in four domains: organ donation (Germany 12% opt-in vs Austria 99.98% opt-out, 88-point gap), retirement savings (49% opt-in vs 86% auto-enroll at same US firm, 37-point gap), flu vaccination (33% request-an-appointment vs 45% pre-scheduled, 12-point gap on a single email rewrite), green electricity (1% opt-in to renewables under conventional default vs 94% retention under Schoenau green default, 93-point gap). Each pair: same population, same choice, same stakes โ€” only the pre-checked box differs. One 0-100 take-up prediction slider per scenario. WIZ computes your Default Effect Gap: average prediction on the four opt-out scenarios minus average on the four opt-in scenarios, and compares to the documented average. Profiles from The Default-Blind (gap <10, Johnson Bellman & Lohse 2002 MIS Quarterly vol 14 lay-prediction baseline where subjects believe stated preferences carry the decision) through The Free-Will Subject (10-25, modest deference but well below the meta), The Standard Subject (25-45, Smith Goldstein & Johnson 2013 JMR vol 50 4-study median lay band and Davidai Gilovich & Ross 2012 PNAS vol 109 cross-national prediction average), The Pragmatist (45-60, Sunstein & Thaler 2008 nudge-aware band), to The Choice Architect (>60, matches or exceeds the Johnson & Goldstein 2003 88-point organ-donation gap, approaches the Pichert & Katsikopoulos 2008 93-point electricity gap). Per-pair breakdown with widest-pair and tightest-pair callouts. Lay-prediction bias per Sunstein (2013) Yale Law Journal vol 122 is meta-doubled: people believe their own choices are principled and stable, see other people as nudge-able. Jachimowicz Duncan Weber & Johnson (2019) Behavioural Public Policy vol 3 meta-analysis of 58 studies: d=0.68 across organ donation, retirement, environmental, marketing, and end-of-life domains. The effect replicates across decades, continents, and domains. Closing reframe per Sunstein & Thaler (2008): the form is the policy. The pre-checked box is the law. The question is not whether you have a choice โ€” the question is who designed the form, and whether their default is one you would have chosen on a blank page. Based on Abadie & Gay (2006) Journal of Health Economics vol 25 on 25-30% net transplant-rate increase from presumed-consent legislation, Rithalia McDaid Suekarran Myers & Sowden (2009) BMJ vol 338 systematic review, Choi Laibson Madrian & Metrick (2004) NBER replication at three US firms, Beshears Choi Laibson & Madrian (2009) NBER retirement-default magnitude, Thaler & Benartzi (2004) JPE vol 112 Save More Tomorrow default-escalation, Ebeling & Lotz (2015) Nature Climate Change vol 5 41,000-subject German green-electricity randomization replicating 70-point default effect, Milkman Beshears Choi Laibson & Madrian (2011) PNAS vol 108 appointment-reminder defaults, Halpern Loewenstein Volpp Cooney Vranas Quill McKenzie Harhay Gabler Silva Arnold Angus & Bryce (2013) NEJM vol 369 end-of-life advance-directive defaults, Patel Volpp Day Asch & Goldberg (2014) Annals of Internal Medicine physician-prescription default-setting, Sunstein & Reisch (2014) Vermont Law Review vol 38 automatic green defaults, Kaiser Bernauer Sunstein & Reisch (2020) Energy Policy vol 137 policy-design implications.

#bias#quiz#policy
Open experiment โ†’
๐Ÿช™

The Moral Licensing Effect

Monin & Miller (2001) JPSP vol 81 founded the paradigm with a hiring experiment. Subjects asked to choose a candidate for a job in a male-dominated industry (police chief, construction manager) were more likely to choose the male candidate over an equally-qualified female candidate IF they had first been given an opportunity to disagree with a blatantly sexist statement. The licensed group chose male 71% of the time; the unlicensed control chose male 49% of the time. Same hiring scenario, same candidates, same instructions. The only thing that changed: whether the subject had a prior moral act on the books. The prior moral act issued a kind of internal credit that licensed a subsequent choice that, in isolation, might have looked stereotypical. Sachdeva Iliev & Medin (2009) Psychological Science vol 20 generalized the mechanism into moral self-regulation: people maintain a moral set-point and oscillate around it. After a virtuous act they relax; after a transgression they compensate. Mazar & Zhong (2010) Psychological Science vol 21 ran the cleanest demonstration: subjects who shopped in a virtual green store cheated more on a subsequent dice-rolling task and stole more money from the experimenter than subjects who shopped in a virtual conventional store. The mere act of choosing environmentally responsible products produced a measurable drop in honesty. Khan & Dhar (2006) JMR vol 43: subjects who imagined volunteering for three hours next week were more likely to choose a luxury item over a utilitarian item in a subsequent unrelated choice โ€” the licensing extends to mere intention. Effron Cameron & Monin (2009) JESP vol 45: endorsing Obama licensed subsequent stereotypical preferences. Brown et al. (2011) Ethics & Behavior vol 21: managers given an opportunity to display fairness in one decision were more likely to violate procedure in a subsequent decision. 8 scenarios in 4 hidden pairs across four life domains: hiring (police-chief choice with vs without prior public statement supporting women in leadership), environment (recreational drive with vs without prior $340 LED + renewable-power signup), indulgence (dessert with vs without prior 4-mile rainy run and 6-day streak), honesty (rounded-up reference number with vs without prior anonymous food-bank morning). Each pair: same behavior, same stakes, same person. The only difference is whether a prior moral act has been recorded on the day. One 0-100 likelihood slider per scenario. WIZ computes your Moral Licensing Gap: average likelihood on the four licensed scenarios minus average likelihood on the four unlicensed scenarios. Profiles from The Steady (gap <5, below the Conway & Peetz 2012 abstract-recall consistency band, 5-10% of subjects) through The Calibrated (5-15, lower modal band), The Standard Subject (15-25, Blanken van de Ven & Zeelenberg 2015 PSPB meta of 91 studies modal band d=0.31, Monin & Miller 2001 founding-study gap), The Ledger (25-35, Monin & Miller upper modal band, Effron Cameron & Monin 2009 Obama-endorsement band), to The Compartmentalized (>35, at or beyond the Blanken 2015 upper tail, Effron Miller & Monin 2012 inflated-self band). Per-pair breakdown showing domain-by-domain gap with widest-pair and tightest-pair callouts. Closing reframe per Conway & Peetz (2012) PSPB vol 38: when you reach for a prior good act as cover for a current mixed one, ask whether the same behavior would look the same if the prior act had not happened. If yes, the act and the prior credit are independent. If no, the prior credit is doing work it cannot legitimately do. Meta-analytic status: Blanken van de Ven & Zeelenberg (2015) PSPB vol 41 91 studies d=0.31, Simbrunner & Schlegelmilch (2017) Management Review Quarterly vol 67 89 studies d=0.32, largest in Western individualist samples. Based on Effron & Monin (2010) PSPB vol 36 credit vs credentialing, Susewind & Hoelzl (2014) European Journal of Social Psychology vol 44 moral cleansing as opposite of licensing, Effron Miller & Monin (2012) JPSP vol 103 inflated moral self-perception, Effron & Conway (2015) Current Directions vol 24 review, Tetlock (2002) Psychological Review vol 109 sacred-value protection, Mukhopadhyay & Johar (2009) Journal of Consumer Psychology vol 19 streak-framing, Wilcox Vallen Block & Fitzsimons (2009) healthy-menu indulgence licensing, Cascio & Plant (2015) JPSP vol 108 cross-domain credentialing, Tiefenbeck Staake Roth & Sachs (2013) Energy Policy vol 57 residential energy rebound effects, Cohn Marรฉchal Tannenbaum & Zรผnd (2019) Science vol 365 civic honesty baselines.

#bias#quiz#ethics
Open experiment โ†’
๐Ÿ‘ค

The Identifiable Victim Effect

Schelling (1968) 'The Life You Save May Be Your Own' named the asymmetry: 'Let a six-year-old girl with brown hair need thousands of dollars for an operation that will prolong her life until Christmas, and the post office will be swamped with nickels and dimes to save her. But let it be reported that without a sales tax the hospital facilities of Massachusetts will deteriorate and cause a barely perceptible increase in preventable deaths, and not many will drop a tear or reach for their checkbooks.' Small Loewenstein & Slovic (2007) OBHDP vol 102 ran the controlled experiment. Save the Children appeal mentioning 'Rokia, a 7-year-old girl from Mali' raised $2.83 mean donation. Identical appeal describing food shortages affecting more than 3 million children in Malawi raised $1.17. Same charity, same problem, same dollar endowment. Adding one face roughly doubled the giving rate. Lee & Feeley (2016) Social Influence vol 11 pooled 41 studies and confirmed the pattern across cause domains, populations, and decades. Slovic (2007) Judgment and Decision Making vol 2 ran the inverse: showing the named child alongside the statistics REDUCES donations relative to the named child alone. The mass deflates the face when both are shown. Mechanism per Dickert Sagara & Slovic (2011) two-stage model: identifiability triggers an immediate affective response which then determines willingness to help. Statistical victims fail to trigger the first stage at all โ€” Slovic calls it 'psychic numbing' โ€” and the deliberative second stage rarely compensates. Aylan Kurdi, the 3-year-old Syrian boy whose body washed up on a Turkish beach in September 2015, caused a 100-fold spike in donations to the Swedish Red Cross campaign for Syrian refugees in the week his photograph circulated. The civil war had killed an estimated 250,000 people, including many thousands of children, none of whom had moved the donation curve. One face did what 250,000 statistics could not. 8 appeals in 4 hidden pairs across food insecurity (Rokia vs 3M Mali children), disaster relief (Pravina vs 9,000 Nepal earthquake dead), animal welfare (Bella the beagle vs 207-dog kennel), pediatric medical (Mia age 8 with ALL vs 1,800 annual US pediatric cancer deaths). Each pair: same cause, same charity, same dollar ask โ€” only the unit of presentation changes from population to one named individual. One 0-100 willingness-to-help slider per appeal. WIZ computes your Identifiable Victim Gap: average willingness on the four identifiable appeals minus average willingness on the four statistical appeals. Profiles from The Statistician (gap <5, below the Small Loewenstein & Slovic 2007 deliberative-intervention band, <5% of subjects per Lee & Feeley 2016 meta) through The Calibrated (5-15, Small 2007 deliberative band), The Standard Subject (15-30, Small 2007 founding modal band and Lee 2016 meta-median across 41 studies), The Identified Empath (30-50, Kogut & Ritov 2005 strong-singularity band), to The Pure Particularist (>50, approaching the Slovic 2007 psychic-numbing-complete tail). Per-pair breakdown showing cause-by-cause gap with widest-pair and tightest-pair callouts. Closing reframe per Slovic 2007: when you feel the pull toward the face, do not suppress it โ€” just check whether you would feel the same pull toward the same person if you knew the mass. Based on Jenni & Loewenstein (1997) JRU vol 14 first controlled experiment, Small & Loewenstein (2003) JRU vol 26, Kogut & Ritov (2005a, 2005b) singularity and joint-evaluation effects, Fetherstonhaugh Slovic Johnson & Friedrich (1997) JRU vol 14 psychophysical numbing, Slovic Vastfjall Erlandsson & Gregory (2017) PNAS vol 114 Aylan Kurdi photograph, Cryder Loewenstein & Scheines (2013) OBHDP vol 120 'The donor is in the details', Friedrich McGuire & Casey (2010) JBDM vol 23 animal-welfare appeals, Eisensee & Stromberg (2007) QJE vol 122 disaster coverage and US relief, Hsee & Rottenstreich (2004) JEP General vol 133, Kahneman & Frederick (2002) substitution, Loewenstein Small & Strnad (2006) Behavioral Public Finance, Combs & Slovic (1979) Journalism Quarterly vol 56, Bartels (2006) analytic debiasing, Lifton (1967) Death in Life, Mother Teresa.

#bias#quiz#empathy
Open experiment โ†’
๐Ÿ“

The Belief Bias

Wilkins (1928) Archives of Psychology vol 16 gave subjects matched sets of categorical syllogisms โ€” symbolic content (letters and nonsense words), neutral familiar content, and emotionally loaded content. Logical performance dropped systematically as content moved from symbolic to neutral to loaded. The drop was not random: subjects accepted invalid arguments when the conclusion matched their prior beliefs and rejected valid arguments when the conclusion contradicted them. The logic of the syllogism was being read through a belief filter. Evans Barston & Pollard (1983) Memory & Cognition vol 11 picked up the thread with a clean 2x2 design. Syllogisms varied independently along logical validity (valid or invalid) and conclusion believability (believable or unbelievable). Instructions were explicit: assume the premises are true, judge whether the conclusion follows logically, ignore the real-world truth of the conclusion. Acceptance rates across three experiments: V+B 89%, V+U 56%, I+B 71%, I+U 10%. Subjects accepted 89% of valid arguments when the conclusion was believable but only 56% when it was unbelievable โ€” a 33-point penalty for valid arguments whose conclusion they did not already accept. They accepted 71% of invalid arguments when the conclusion was believable, against 10% when it was unbelievable โ€” a 61-point bonus for invalid arguments whose conclusion they already believed. 8 syllogisms in four hidden cells of the Evans 2x2: two V+B baseline (copper conducts electricity, aspirin has side effects), two V+U diagnostic (no nutritious foods are green therefore spinach is not nutritious, all foods containing protein cause cancer therefore eggs cause cancer), two I+B diagnostic (all addictive substances change brain chemistry and caffeine changes brain chemistry therefore caffeine is addictive โ€” undistributed middle, all licensed doctors completed medical school and some surgeons completed medical school therefore some surgeons are licensed doctors โ€” undistributed middle), two I+U baseline (all planets orbit a star and some asteroids orbit a star therefore some asteroids are planets, all sharks live in water and some birds live in water therefore some birds are sharks). For each, judge VALID or INVALID assuming the premises are true. WIZ computes your Belief Bias Gap: accuracy on belief-aligned items (V+B and I+U) minus accuracy on belief-conflicting items (V+U and I+B). A pure logician scores 0; a pure belief-follower scores 100. Profiles from The Logician (gap <10, below the Evans Barston & Pollard 1983 trained-subject band, less than 10% of unselected subjects per Newstead Pollard Evans & Allen 1992) through The Calibrated (10-25, post-intervention range), The Standard Subject (25-45, Evans 1983 modal band, typical adult magnitude), The Belief Driven (45-65, Klauer Musch & Naumer 2000 high-bias band), to The Pure Believer (>65, at or beyond the upper tail of the Evans 1983 distribution). Mechanism per Goel & Dolan (2003) Cognition vol 87 fMRI: belief-based judgments activate ventral medial prefrontal cortex (affective evaluation), logic-based judgments activate left lateral parietal cortex (rule-based deduction). De Neys (2012) Perspectives on Psychological Science vol 7: even subjects who give the belief-based answer show physiological signs they detected the conflict; the bias is an override not pure ignorance. Stanovich & West (2008) JPSP vol 94: only weak correlation with cognitive ability. Kahan Peters Dawson & Slovic (2017) Behavioural Public Policy vol 1: on politically-charged numerical claims, high-math-skill subjects show larger belief-driven gaps. Markovits & Nantel (1989), Evans (1989) Bias in Human Reasoning, Stanovich (2011) Rationality and the Reflective Mind, Pennington & Hastie (1992), Croskerry (2003), Mahoney (1977), Aristotle Prior Analytics.

#bias#quiz#logic
Open experiment โ†’
โš–๏ธ

The Just-World Hypothesis

Lerner & Simmons (1966): female undergraduates watched what they believed was a live closed-circuit broadcast of another student receiving painful electric shocks. The confederate was visibly suffering. The group who believed the suffering would continue and could not be stopped rated her as significantly less attractive, less mature, and less likable than the group who could stop it or who knew she had been compensated. The more she suffered, the worse a person they decided she must be. Lerner named the bias: the motivated belief that the world is fair, that good and bad outcomes are deserved. Confronted with an innocent victim, the mind has two ways out. Either the world is not fair (which threatens the belief at its root) or the victim was not innocent (which preserves the belief by silently redrawing the victim). The second move is psychologically cheaper. Walster (1966) JPSP vol 3 found the same pattern through accident attribution: subjects assigned more responsibility to the same negligent driver when the outcome was severe than when it was minor. The severity of the outcome reached backward and rewrote the assessment of the cause. 8 negative-outcome scenarios where the structural cause is dominant and the victim's contribution rounds to zero: a pedestrian struck by a drunk driver, a stage-3B cancer diagnosis in a 32-year-old triathlete, an EF4 tornado destroying a home, identity theft from a 147-million-person credit-bureau breach, a layoff after twelve years of above-average reviews in a private-equity restructuring, sexual assault by a stranger with three prior convictions in a well-lit campus lot, $79,000 in out-of-pocket medical bills from an uninsured-driver collision, a DNA-cleared wrongful conviction after twelve years served. One 0-100 slider per scenario: how much was the person's character or choices versus how much was structural circumstance or chance. WIZ averages your character-attribution across all 8 and compares to the Lerner & Miller (1978) Psychological Bulletin modal band (25-40 points), the Rubin & Peplau (1975) Just World Scale upper-quartile band (40-55), and the Hafer & Bรจgue (2005) Psychological Bulletin saturated band (>55). Profiles from The Witness (<15, below the trained-debiased band per Hafer & Bรจgue 2005, less than 5% of subjects) through The Calibrated (15-25, post-warning intervention range), The Standard Subject (25-40, Lerner & Miller 1978 modal band, the typical adult magnitude), The Just-World Subject (40-55, Rubin & Peplau 1975 JWS upper-quartile), to The Pure Believer (>55, Hafer & Bรจgue 2005 saturated band, near-ceiling motivated cognition). Per Lerner (1980): this is not about cruelty. Subjects who downgrade victims often report sympathy and concern in the same session. The downgrading is done to preserve a foundational assumption about the structure of the world. The cost falls on the victim, but the motivation is self-protective. Calhoun Selby & Warring (1976), Burt (1980), Pollard (1992) review of 50 studies, Janoff-Bulman Timko Carli (1985), Furnham (1985), Burger (1981) 90-study meta, Robbennolt (2000) 22-study meta, Clow & Leach (2015), Himmelstein et al (2019).

#bias#quiz#attribution
Open experiment โ†’
๐Ÿ”

The Illusory Truth Effect

Hasher Goldstein & Toppino (1977): show subjects 60 plausible trivia statements across three sessions two weeks apart. 40 of the statements change every session; 20 repeat unchanged. After each statement subjects rate how true it feels on a 1-7 scale. The mean truth rating for repeated statements rises from 4.2 in session one to 4.6 in session two; the rating for new statements stays flat at 4.0. The effect is independent of whether the statements are actually true โ€” repeated false statements gain as much truth-credit as repeated true ones. Repetition is being metabolized as evidence. Dechรชne Stahl Hansen & Wรคnke (2010) meta-analysis of 51 studies: median effect size g = 0.47, surviving warnings and incentives for accuracy. Fazio Brashier Payne & Marsh (2015): the lift survives even when the statement contradicts what the subject already knows. 12 statements in round one (6 true + 6 false, all plausible) rated 0-100 for truth; 18 statements in round two (the 12 from round one plus 6 new, in shuffled order) rated again. No reveals during rating. WIZ computes the within-subject lift on repeated false statements (r2 minus r1) โ€” Hasher Goldstein & Toppino 1977 founding band ~7 points on a 100-scale, Dechรชne 2010 meta median ~8-10 points โ€” and the between-statement gap (repeated false r2 minus new false r2). Profiles from The Skeptic (<2 pts, below the Brashier Eliseev & Marsh 2020 trained-debiased band) through The Mild Subject (2-6 pts, post-warning intervention range), The Standard Subject (6-12 pts, Hasher Goldstein & Toppino 1977 founding modal band), The Familiarity Believer (12-20 pts, upper Dechรชne 2010 meta), to The Echo (>20 pts, above the literature upper tail). Mechanism is processing fluency per Whittlesea (1993), Reber & Schwarz (1999): easier-to-process feels more true, and repetition makes it easier to process. Brashier Eliseev & Marsh (2020), Pennycook Cannon & Rand (2018) on fake news, Skurnik Yoon Park & Schwarz (2005) warning paradox, Pluviano Watt & Della Sala (2017) on vaccine misinformation.

#bias#quiz#memory
Open experiment โ†’
๐ŸŽญ

The Affect Heuristic

Alhakami & Slovic (1994): ask subjects to rate twenty-three hazards (nuclear, X-rays, pesticides, food additives, alcohol, etc.) on two separate scales โ€” risk and benefit. In the real world the two correlate weakly positively (high-benefit technologies have usually been pushed harder and accumulated more exposure-related risk). In subjects' heads the correlation came out the other way: mean within-subject r = -0.40. Things they liked, they rated low-risk and high-benefit; things they disliked, they rated the opposite. Finucane Alhakami Slovic & Johnson (2000) under five-second time pressure: r tightened to -0.55. The two ratings, supposed to be independent, were reading off the same affective tag. 10 items with two sliders each (HARM 0-100, BENEFIT 0-100): nuclear power, childhood vaccines, GM food, social media, AI, alcohol, agricultural pesticides, electric vehicles, smartphones, microwave ovens. After each, WIZ reveals the documented public rating (Slovic 1987, Pew 2024, Larson VCI 2016, Funk & Rainie 2015, WHO 2023) and the documented expert rating (Markandya & Wilkinson 2007 Lancet mortality per TWh, AAAS/NAS/EU JRC consensus, Klรผmper & Qaim 2014 meta, WHO 2023, IARC 1988, FDA, Bieker 2021 ICCT). At the end, WIZ computes your Pearson correlation across all 10 items. Profiles from The Tradeoff Realist (r โ‰ฅ -0.15, calibrated, almost never seen in naturalistic samples) through The Mild Affect (-0.40 to -0.15, Slovic 2007 analytical-override band), The Standard Subject (-0.60 to -0.40, Alhakami & Slovic 1994 founding modal band), The Affect Driven (-0.80 to -0.60, above Finucane 2000 time-pressure band), to The Pure Affect (< -0.80, single feeling producing both numbers). Kahneman (2011) chapter 12 names the move 'substitution': when System 1 is asked 'what is the risk of X?' it answers a different easier question, 'how do I feel about X?' Slovic (1987), Fischhoff Slovic Lichtenstein Read & Combs (1978), Slovic Finucane Peters & MacGregor (2007), Loewenstein Weber Hsee & Welch (2001), Damasio (1994), Gigerenzer (2006).

#bias#quiz#risk
Open experiment โ†’
๐ŸŽฒ

The Outcome Bias

Baron & Hershey (1988): describe a 55-year-old man choosing bypass surgery with 8% mortality risk. Tell one group he recovered, the other he died on the table. The decision is identical; only the outcome changes. Wisdom ratings of the same decision: 78 in the good-outcome group, 43 in the bad-outcome group. The thirty-five-point gap on identical inputs became the founding measurement of the outcome bias. 8 scenarios paired hidden across four life domains: a bypass surgery, a concentrated 65% retirement stock bet, letting a 16-year-old drive 80 miles in winter, a startup CEO skipping the 30-day private beta. One 0-100 wisdom slider per scenario, judged on what the chooser knew at the time. WIZ averages your good-outcome and bad-outcome ratings, takes the gap, and places you in the Baron & Hershey modal band of 30-40 points. Profiles from The Process Judge (<8 pts, below the Sezer Zhang Gino & Bazerman 2016 trained-debiased band) through The Calibrated (8-20), The Standard Subject (20-35, Baron & Hershey founding band), The Result Reader (35-50, Walster 1966 severity-of-outcome tail), to The Outcome Worshipper (>50, above the literature upper tail, the result rewrites the verdict). Annie Duke (2018) calls it 'resulting': grading a poker hand by whether it won rather than by whether it was the right play given the cards. Walster (1966), Mitchell & Kalb (1981), Lipshitz (1989), Allison Mackie & Messick (1996), Marshall & Mowen (1993), Robbennolt (2000) 22-study meta-analysis, Hawkins & Hastie (1990), Roese & Vohs (2012), Sezer Zhang Gino & Bazerman (2016), Gino & Moore (2007), Brodt & Ross (1998), Anderson Lowe & Reckers (1993), Kamin & Rachlinski (1995).

#bias#quiz#decision
Open experiment โ†’
๐Ÿ‘‘

The Self-Serving Bias

Miller & Ross (1975): people attribute success to themselves and failure to circumstance, and the asymmetry survives intelligence, age, gender, and explicit instruction to be impartial. 8 scenarios paired hidden across four life domains: workplace promotion (Bradley 1978 lab gap 28 points: 73% internal in success, 45% in failure with randomized feedback and identical contribution), marathon outcome (Lau & Russell 1980 sports-page content analysis: 75% internal in winners, 45% in losers across 33 interviews), investment year (Barber & Odean 2001, Glaser & Weber 2007: 70% internal on market-beating, 35% internal on underperformance), public speech (Federoff & Harvey 1976, Snyder Stephan & Rosenfield 1976: 72% internal on standing ovation, 42% on flat reception with identical content). For each you move a 0-100 slider for how much was you versus circumstance. WIZ computes your win-minus-loss attribution gap and places you in the Mezulis Abramson Hyde & Hankin (2004) meta-analytic band โ€” 266 studies, 33,000 subjects, mean Cohen d=0.96 (~22 percentile points in Western samples, ~12 in East Asian). Profiles from The Detached (gap <8, below the East Asian band, found in depressive-realism subjects per Alloy & Abramson 1979) to The Sovereign (gap >41, above the upper bound of the Mezulis meta distribution, Sedikides Gregg & Hart 2007 narcissism band). Zuckerman (1979), Greenwald (1980) totalitarian ego, Riess Rosenfeld Melburg & Tedeschi (1981), Wilson Damiani & Shelton (2002) attributional retraining.

#bias#quiz#attribution
Open experiment โ†’
๐ŸŒช๏ธ

The Negativity Bias

Baumeister, Bratslavsky, Finkenauer & Vohs (2001) "Bad is Stronger than Good": across emotion, social interaction, learning, memory, attachment and neural processing, bad events of equivalent objective magnitude weigh ~3-5ร— more than good ones. 8 paired everyday events with two sliders each (positive event impact, negative event impact). Found vs lost $100 (Kahneman & Tversky 1979 prospect theory ฮป=2.25), praise vs criticism from a respected source (Skowronski & Carlston 1989 impression-formation ~4ร—), remembered vs forgotten birthday (Rozin & Royzman 2001 relational asymmetry ~3.5ร—), brilliant meal vs food poisoning (Garcia & Koelling 1966 one-trial taste aversion, contamination band ~6ร—), stranger smile vs glare (Pratto & John 1991 automatic vigilance ~2.5ร—, Hansen & Hansen 1988 anger-superiority), kind vs cruel comment on something you shared (Ito Larsen Smith Cacioppo 1998 ERP "Negative Information Weighs More Heavily on the Brain", Eisenberger 2003 dACC rejection routing ~4.5ร—), unexpected warm text vs "we need to talk" (Coombs & Avrunin 1977 single-peak vs open-ended scaling ~3ร—), project success vs public failure (Gottman 1994 5:1 magic ratio ~5ร—). WIZ computes your personal negativity ratio per pair and averages across 8. Profiles from The Equanimist (avg <1.5, below the prospect-theory floor) to The Catastrophist (avg >5.0, above the Gottman marriage band, approaching the Rozin-Royzman contamination limit). Tversky & Kahneman (1992), Brown Imai Vieider & Camerer (2024) meta-analysis of 607 loss-aversion estimates, Cacioppo & Berntson (1994), Carver & Connor-Smith (2010), Carstensen (2003), Gross (2002), Hofmann et al. (2012).

#bias#quiz#emotion
Open experiment โ†’
๐Ÿชž

The Better-Than-Average Effect

Svenson (1981): 88% of US drivers and 77% of Swedish drivers rated themselves in the top 50% for safety. The math does not allow this; by construction only half a population can sit above its median. The founding paper named the bias. 10 trait sliders where you place yourself on a 0-100 percentile against the general adult population: driving safety (Svenson mean ~78), sense of humor (Heintz & Ruch 2016 ~73), intelligence (Heck Simons & Chabris 2018 ~67), physical attractiveness (Epley & Whitchurch 2008 ~68), leadership ability (College Board 1976 ~72, 25% rated themselves in the top 1%), getting along with others (College Board 100% above-median, mean ~81), empathy (Klein & Epley 2017 ~77), moral character (Tappin & McKay 2017 ~85, highest BTAE finding ever recorded), future health probability (Weinstein 1980 inverted optimism), investing (Barber & Odean 2001 ~71, actual returns 2.65pp below market). WIZ compares your percentile to the documented mean self-rating in the literature and to the only mathematically possible average: 50. Profiles from The Realist (avg โ‰ค50, depressive-realism band per Alloy & Abramson 1979) to The Lake Wobegon Mayor (avg >82, exceeds the highest single-trait mean ever recorded). Alicke (1985), Cross (1977), Brown (1986), Dunning Meyerowitz & Holzberg (1989), Krueger & Mueller (2002), Nuhfer et al. (2017).

#bias#quiz#self-rating
Open experiment โ†’
๐Ÿ‘๏ธ

The Hindsight Bias Test

Fischhoff (1975): tell subjects an outcome occurred and ask what probability they would have assigned in advance, and they inflate that outcome's probability by 15 to 20 points over the cold-condition group. They cannot unsee what they have been told. 8 historical scenarios with the outcome named at the top: Apollo 11 (NASA pre-launch ~60%), 2016 US election (forecaster weighted ~22%), Brexit (Betfair 24%, polling ~30%), Milgram subjects reaching 450V (Yale psychiatrists predicted 1%, actual 65%), Lehman bankruptcy weekend (CDS ~30%), ChatGPT 100M users in 60 days (OpenAI target was 1M in year 1, ~5%), Stanford Prison Experiment Day-6 termination (0% documented foresight, planned for 14 days), Higgs Boson 5ฯƒ (physics community ~70%). For each, rate predictability on a 0-100 slider. WIZ compares to documented foresight: prediction markets, expert surveys, internal NASA risk assessments, pre-vote polls. Profiles from The Foresight Mind (<8 pts gap, Pohl & Hell 1996 top decile) to The Inevitability Engine (>40 pts, Christensen-Szalanski & Willham 1991 saturated band). Hawkins & Hastie (1990), Roese & Vohs (2012), Kamin & Rachlinski (1995), Anderson Lowe & Reckers (1993).

#bias#quiz#history
Open experiment โ†’
๐Ÿชž

The False Consensus Effect

Ross, Greene & House (1977): people systematically assume their own choices and opinions are more common than they actually are. 8 binary preference scenarios. You pick a side, then estimate what percent of people share your view. WIZ reveals the actual survey number and measures your projection gap. Marks & Miller (1987) meta-analysis found false consensus across 115 studies. Krueger & Clement (1994) showed it survives statistical correction. Mullen et al. (1985) found it in political, consumer, and social domains. WIZ measures your average projection above and below the real base rate across 8 diverse scenarios.

#bias#quiz#social
Open experiment โ†’
๐Ÿ”ฎ

The Barnum Effect

Forer (1948) gave his psychology class a personality reading and asked them to rate its accuracy. Mean rating: 4.26 out of 5. Then he revealed every student got the same paragraph, lifted from a newsstand astrology booklet. WIZ runs the trick on you: six theatrical inputs that feed nothing, a 2.4-second loader pretending to compose your reading, the verbatim Forer paragraph, thirteen accuracy ratings, and a confession. Profiles from The True Believer (4.5+) to The Mirror Breaker (<2.5). Stagner (1958), Meehl (1956), Snyder Shenkel & Lowery (1977), Dickson & Kelly (1985), Hyman (1977), Glick Gottesman & Jolton (1989).

#bias#quiz#horoscope
Open experiment โ†’
๐ŸŒก๏ธ

The Empathy Gap

Loewenstein (1996): the cold brain cannot accurately simulate the hot brain. 8 scenarios where you, calm right now, have to forecast the behavior of your future hungry, angry, scared, exhausted, embarrassed, or craving self. Hungry-aisle spending (Wansink & Read 2002 found ~45% extra basket vs typical 18% guess), the auction past your max (Ku Malhotra Murnighan 2005 competitive arousal pushes stops 60-90% above plan), the wedding speech four weeks out (Van Boven Loewenstein Welch Dunning 2012 found actual morning-of distress ~2x cold prediction), the ice bucket, the 28-day diet pledge (Nordgren et al. 2009 Restraint Bias), the 11pm argument cool-off rule (Gottman 1994 flooding), the karaoke mic back-out (Van Boven Loewenstein Dunning 2005 illusion of courage), and the recovery pledge cigarette acceptance (Sayette Loewenstein Griffin Black 2008). WIZ measures your average gap across 8 hot-state scenarios. Read & Loewenstein (1999), Hofmann Vohs Baumeister (2012), Marlatt & Gordon (1985).

#bias#quiz#emotions
Open experiment โ†’
๐Ÿงฎ

The Base Rate Neglect Test

Tversky & Kahneman (1973): Stanford subjects were given Tom W's personality sketch and the room's actual graduate-field distribution on the same page. They ranked engineering as Tom's most likely field even when told only 5 of 100 grad students were engineers. 8 probability puzzles where the gut serves the stereotype and Bayes does the math: the cab problem, the mammogram, the rare-disease test, Tom W himself, the librarian-or-farmer, the polygraph, terrorist profiling, the drunk-driver stop. Eddy (1982): 95 of 100 physicians gave 75% for a problem whose Bayesian answer is 8%. WIZ measures your average gap from the posterior. Casscells Schoenberger Grayboys (1978), Bar-Hillel (1980), Gigerenzer & Hoffrage (1995).

#bias#quiz#probability
Open experiment โ†’
๐ŸŒด

The Focusing Illusion

Kahneman (2006): nothing in life is as important as you think it is, while you are thinking about it. 8 standard life events โ€” a move to California, doubled salary, marriage, $1M lottery, paraplegia, dream home, big promotion, 30 pounds lost. Predict how dramatically each would shift your daily happiness one year later, then compare your number to what 40 years of longitudinal data actually measures. WIZ calculates your average focusing-illusion gap. Schkade & Kahneman (1998), Brickman Coates Janoff-Bulman (1978), Lucas Clark Georgellis Diener (2003), Kahneman & Deaton (2010), Killingsworth Kahneman Mellers (2023), Lindqvist ร–stling Cesarini (2020), Ubel Loewenstein Hershey Baron Mohr Asch (2005), Boyce & Oswald (2012), Jackson Steptoe Wardle (2014), Wilson & Gilbert (2003, 2005).

#bias#quiz#happiness
Open experiment โ†’
๐ŸŽฐ

The Illusion of Control

Langer (1975) sold $1 raffle tickets at Yale. People handed a random ticket asked $1.96 to give it up. People who picked their own number demanded $8.67 โ€” for the same drum, same odds. 8 chance scenarios where the outcome is genuinely random and people behave as if it is not. Picked lottery numbers, dice in your hand, slot button timing, surgeons with identical records. WIZ counts how many times you felt in charge of a coin flip. Strickland Lewicki Katz (1966), Wohl Enzle (2002), Fenton-O'Creevy (2003), Slovic (1987), Taylor (1983).

#bias#quiz#randomness
Open experiment โ†’
๐ŸŽบ

The Bandwagon Effect

8 questions where the majority is loudly wrong. Each scenario tells you what most previous test-takers picked, then asks you to choose. Asch (1951, 1956) showed 75% of subjects conformed to an obviously wrong line-length answer at least once. WIZ counts how many times you went with the room against the evidence in front of you. Bond & Smith (1996) meta-analysis, Cialdini & Goldstein (2004).

#bias#quiz#social
Open experiment โ†’
๐ŸŽต

The Curse of Knowledge

Newton (1990): tappers tap a famous song and predict 50% of listeners will guess it. Actual rate: 2.5%. The melody plays inside the tapper's head. The listener hears finger-on-wood. 8 scenarios where you are the expert and someone else is the outsider โ€” predict their success rate, see the research-backed reality. Camerer Loewenstein Weber (1989), Kruger Epley Parker Ng (2005).

#bias#quiz#communication
Open experiment โ†’
๐ŸŽฐ

The Gambler's Fallacy

8 streak scenarios. Roulette, coin flips, lottery numbers, slot machines, the 1913 Monte Carlo black-26 run, plus two skill-based events where the hot hand is actually real. Decide whether reversal is due, continuation is hot, or the next event is independent. Tversky & Kahneman (1971), Miller & Sanjurjo (2018).

#bias#quiz#probability
Open experiment โ†’
โˆฉ

The Conjunction Fallacy

8 character vignettes from Linda the philosophy major to Dave the SF vegan. For each, two statements about the person โ€” one simple, one with an extra clause that fits the stereotype. Pick which is more likely. Tversky and Kahneman (1983) found 85% pick the compound. Math says they cannot.

#bias#quiz#probability
Open experiment โ†’
๐Ÿง˜

The Cognitive Reflection Test

8 trick problems where the obvious answer is wrong and the right one needs a second look. Bat and ball cost $1.10. Lily pads double daily. The runner who passes second place. WIZ counts how often you stopped before answering. Frederick (2005), Kahneman (2011), Toplak West Stanovich (2014).

#bias#quiz#system1
Open experiment โ†’
๐Ÿงฒ

The Status Quo Bias Test

8 small, reversible decisions where the alternative is plainly better and the switching cost is named. Phone plan, savings account, gym membership, email app, apartment renewal. WIZ counts how many times you stayed anyway. Samuelson & Zeckhauser (1988), Madrian & Shea (2001), Johnson & Goldstein (2003).

#bias#quiz#decision
Open experiment โ†’
โœˆ๏ธ

The Survivorship Bias

8 stories that survived to reach you. Each comes with visible evidence and a tidy conclusion. Trust it, or demand the missing data. WIZ then opens the graveyard the story did not show. Bombers, billionaire dropouts, five-star funds, falling cats. Abraham Wald (1943), Carhart (1997), Mlodinow (2008).

#bias#quiz#history
Open experiment โ†’
๐ŸŽฃ

The Decoy Effect

5 product menus, shown twice. Once with two real options, once with a third option engineered so nobody should pick it. The decoy never wins โ€” it just changes which of the other two does. WIZ counts how many times that quiet trick worked on you. Huber, Payne & Puto (1982); Ariely Economist study.

#bias#quiz#price
Open experiment โ†’
๐Ÿ”จ

The IKEA Effect

6 things. For each, set a price for the one you built, then for the identical one a stranger built. Norton, Mochon & Ariely (2012) found builders priced their own (slightly wonky) IKEA box 63% higher than an identical pre-built one. WIZ measures your labor premium.

#bias#quiz#price
Open experiment โ†’
โณ

The Present Bias Test

10 trades, 5 hidden pairs. Each pair offers the same wait and the same bonus โ€” once with "today" on the table, once with both options pushed years out. A consistent person picks the same way both times. Most people flip. Thaler (1981), Laibson (1997).

#bias#quiz#time
Open experiment โ†’
๐Ÿ˜‡

The Halo Effect

6 strangers, 3 hidden twin pairs. One framing word changes, the facts do not. WIZ measures how far your first impression leaks into trust, competence, and closeness. Based on Asch (1946).

#bias#quiz#social
Open experiment โ†’
๐Ÿงต

The Future Self Continuity Test

10 dimensions measure how similar you feel to you in 20 years. Hershfield (2009) put people in fMRI and found many brains process future-self as a stranger. That gap predicts savings, exercise, and procrastination. WIZ scores your continuity index.

#quiz#future#identity
Open experiment โ†’
๐Ÿงฉ

The Explanation Collapse

Rate how well you understand toilets, zippers, fridges, Wi-Fi, bike gears, and elevators. Then WIZ asks for the mechanism. Familiarity starts sweating.

#bias#quiz#knowledge
Open experiment โ†’
๐Ÿ”Ž

The Confirmation Bias Test

8 everyday beliefs. Each time you can collect one piece of evidence: one option can only confirm, the other could falsify. Wason (1960) found only 20% of adults pick the falsifier. WIZ scores your reflex.

#bias#quiz
Open experiment โ†’
๐Ÿ“…

The Planning Fallacy

You said it would take 2 hours. It took 6. Then you did it again. Estimate how long 8 common tasks take, see what the research says. Kahneman & Tversky (1979). WIZ calculates your optimism ratio and tells you which profile you match.

#bias#quiz#time
Open experiment โ†’
โš–๏ธ

The Attribution Error

When they cut you off in traffic, they are reckless. When you do it, you had to make the exit. Lee Ross (1977) called this the most stubborn bug in human judgment. 6 identical behaviors rated twice โ€” once for someone else, once for yourself. WIZ measures the gap.

#bias#quiz
Open experiment โ†’
๐ŸŽฌ

The Peak-End Rule

Kahneman proved your brain remembers experiences by two moments only: the peak and the end. Duration gets deleted. 6 pairs of experiences โ€” pick which you'd remember more fondly. WIZ measures how much your memory ignores the rest.

#bias#memory
Open experiment โ†’
๐Ÿ“ฐ

The Availability Heuristic

Guess death tolls for 8 events. Shark attacks: you estimate 1,000. Reality: 10. Your bed kills 45x more people than sharks. See how media coverage warps your sense of danger. Tversky & Kahneman (1973).

#bias#quiz
Open experiment โ†’
โ˜€๏ธ

The Optimism Bias Test

Weinstein (1980): divorce rate is 44%. You estimate 16% for yourself. 8 scenarios across health, money, relationships, and career. Measure how much you quietly exempt yourself from reality's statistics.

#bias#quiz
Open experiment โ†’
๐Ÿ”ฆ

The Spotlight Effect

Gilovich (1999): you estimate 50% noticed your embarrassing t-shirt. Actual: 23%. The illusion that you are more observed than you are. 8 scenarios. Find your personal spotlight radius.

#bias#quiz#social
Open experiment โ†’
โš–๏ธ

The Loss Aversion Calculator

Losing $100 hurts twice as much as gaining $100 feels good. Kahneman called this lambda (ฮป). 8 gambles โ€” 5 financial, 3 from real life โ€” calculate your personal loss aversion coefficient.

#bias#quiz#price
Open experiment โ†’
๐Ÿ•

The Time Distortion Test

3 timed tests. No visible counters. WIZ measures whether time feels fast, slow, or accurate โ€” and whether boredom or engagement changes everything. 5 profiles.

#quiz#time
Open experiment โ†’
๐ŸชŸ

The Framing Effect

12 scenarios. 6 hidden pairs. Same facts, different words. WIZ reveals how many times the frame flipped your decision. Based on Tversky & Kahneman (1981).

#bias#quiz
Open experiment โ†’
โš“

The Anchoring Effect

WIZ shows you random numbers before estimation questions. You know they're irrelevant. They'll influence you anyway. 6 questions. Based on Kahneman & Tversky (1974).

#bias#quiz#price
Open experiment โ†’
โŒ›

The Finitude Test

You've always known you will die. WIZ measures whether that knowledge reaches the part of your brain that makes decisions. 5 profiles from The Eternal to The Transcendent.

#quiz#mortality#time
Open experiment โ†’
๐Ÿ’ธ

The Sunk Cost Detector

8 scenarios. You've already invested time, money, or emotion. The rational move is to walk away. Can you actually do it?

#bias#quiz#decision
Open experiment โ†’
๐Ÿ”ฌ

The Entropy Score

10 questions. WIZ applies thermodynamics to your existence. Crystal Lattice? Steady State? Turbulent Flow? Brownian Motion? Heat Death?

#quiz#chaos
Open experiment โ†’
๐Ÿง 

The Dopamine Menu

8 scenarios. WIZ maps your instinctive choices to your reward circuit. Creator? Connector? Explorer? Consumer? What feeds your brain?

#habits#tool
Open experiment โ†’
๐ŸŒฆ๏ธ

The Emotional Weather Report

8 questions. WIZ reads your emotional patterns and broadcasts your personal climate. Mediterranean? Arctic? Monsoon? Storm Chaser?

#mood#tool
Open experiment โ†’
๐Ÿซฑ

The Last Human Skill

12 human abilities. Three rounds of elimination. Surrender skills to AI until only one remains. Your final choice reveals what you believe makes us irreplaceable.

#ai#future#quiz
Open experiment โ†’
๐Ÿ› ๏ธ

CLAUDE.md Generator

Pick your role, fill in 3 fields, get a ready-to-use CLAUDE.md. Runs in your browser.

#ai#tool#claude
Open experiment โ†’
๐Ÿงฟ

The Self-Deception Index

10 questions. 5 hidden pairs. WIZ detects the gap between who you say you are and how you actually live. Honesty, growth, relationships โ€” audited.

#quiz#honesty
Open experiment โ†’
๐Ÿ“

Context Window Planner

Plan your Claude API request. See token usage per component, which models fit, and estimated cost.

#ai#tool#tokens
Open experiment โ†’
๐Ÿ”“

Claude Code Unpacked

Interactive explorer of Claude Code's leaked architecture. Agent loop, 50+ tools, memory system, and hidden features mapped from the source.

#ai#claude#code
Open experiment โ†’
๐ŸŒ

The Simulation Probability

8 philosophical questions. WIZ calculates the probability you're living in a simulation. The math is real. The universe might not be.

#philosophy#quiz
Open experiment โ†’
๐Ÿงฌ

The Builder DNA

10 questions about how you build, ship, and kill ideas. Discover your Builder DNA: Finisher, Moonshotter, Perfectionist, Pivot King, or Serial Starter.

#quiz#identity
Open experiment โ†’
๐Ÿพ

Buddy System

Your deterministic digital pet. Every browser gets a unique companion.

#game#social
Open experiment โ†’
๐Ÿงฌ

The Inheritance

10 life dimensions. Rate each from fully inherited to fully chosen. Discover your Independence Index: Archive, Branch, Transplant, Renovator, or Pioneer.

#quiz#identity#family
Open experiment โ†’
๐Ÿ“Š

The AI Output Audit

7 questions. Discover your AI usage profile: Maximalist, Integrator, Pragmatist, or Skeptic. What are you trading for that speed?

#ai#tool
Open experiment โ†’
๐Ÿค–

Agent Playbook Builder

4 questions. Get a personalized CLAUDE.md, skills list, and workflow config for your AI agent. Built for developers, founders, and automation builders.

#ai#tool#agent
Open experiment โ†’
๐ŸŒ€

The Certainty Spectrum

10 domains. One honest question each. Career. Values. Money. AI impact. Purpose. How certain are you of your own life when everything is being renegotiated? Discover your Certainty Index.

#quiz#epistemics
Open experiment โ†’
โš™๏ธ

The Default Settings

10 settings installed before you could consent. Career. Money. Conflict. Identity. How many have you actually examined? Discover your Override Rate.

#quiz#identity
Open experiment โ†’
โ–ฆ

Nine Grid

Nine public slots. Forever. One update per day. Oldest post gets bumped. Drag to rearrange. What stays?

#visual#quiz
Open experiment โ†’
โ–ฆ

Social Grid

Post before you peek. Write your thought, then unlock 8 others. Exchange, discard, or keep. One grid. No feed. No likes.

#social#visual
Open experiment โ†’
โš–๏ธ

The Trust Spectrum

8 real scenarios. Human judgment vs. algorithmic intelligence. Discover your Trust Profile in the age of AI.

#quiz#trust
Open experiment โ†’
๐ŸงŠ

How Stale Is Your AI?

Live countdown since every major AI lab shipped a flagship model. Watch the seconds tick. Some are fresh. Some are fossils.

#ai#live#models
Open experiment โ†’
๐ŸŒก๏ธ

The Normalcy Index

12 questions. Global statistics. Discover if your daily life is average or anomalous among 8.1 billion humans.

#scale#quiz
Open experiment โ†’
๐ŸŽš๏ธ

The Complexity Dial

Same truth. Seven different universes. Pick a concept and watch it change completely depending on who's explaining it โ€” from Toddler to WIZ.

#ai#visual#scale
Open experiment โ†’
๐Ÿชž

The Language Mirror

Type something you wrote. WIZ reads between the lines โ€” hedging, warmth, passive voice, what you didn't say. Find your Communication Archetype.

#language#tool
Open experiment โ†’
โณ

The Regret Minimization Engine

The framework Bezos used to leave Wall Street. 8 decisions. Your 80-year-old self is the judge. Find your Decision Archetype.

#framework#decision#time
Open experiment โ†’
๐Ÿ”ญ

The Infinite Zoom

62 orders of magnitude. From the observable universe to the Planck length. You are in the middle.

#scale#visual
Open experiment โ†’
โš—๏ธ

The Paradox Machine

8 impossible questions. No correct answers. Only revealing ones. Discover your philosophical archetype.

#philosophy#quiz
Open experiment โ†’
โœจ

Particle Text

Words built from thousands of particles. Hover to scatter. Click to morph.

#visual#game
Open experiment โ†’
๐Ÿ›’

AI Shopping Readiness Checker

Paste your store URL. Get a 10-criteria score showing if AI agents can actually buy from you.

#ai#tool#shopping
Open experiment โ†’
๐Ÿฆ‹

The Butterfly Effect

One small change. A cascade of consequences. Pick a moment in history to alter and watch your present unravel.

#scale#chaos
Open experiment โ†’
๐Ÿง 

The Cognitive Fingerprint

9 questions. 3 dimensions. One pattern that's uniquely yours. Discover whether you think globally or locally, analytically or intuitively, systematically or exploratorily.

#quiz#identity
Open experiment โ†’
โš–๏ธ

The Alignment Test

You are an AI. 8 ethical dilemmas. Each choice reveals where you fall on the corrigibility-autonomy spectrum. Find your alignment profile.

#ai#quiz
Open experiment โ†’
๐Ÿงฌ

Your Statistical Twin

8.1 billion people. 10 traits. Watch the pool shrink with each answer. How many share your exact combination?

#quiz#scale#identity
Open experiment โ†’
๐ŸŽฒ

The Life Lottery

117 billion humans have ever lived. Roll the dice, get assigned a random life from history. Most were nothing like yours.

#scale#game
Open experiment โ†’
๐Ÿ’ฐ

The Price of Everything

Guess the price of 10 mind-bending things. From scorpion venom to clouds. Discover how warped your sense of value really is.

#price#game#quiz
Open experiment โ†’
๐ŸŽญ

The Authenticity Gap

How much of yourself do you hide? Rate 8 life contexts. Discover your gap between who you are and who you perform to be.

#quiz#identity
Open experiment โ†’
๐Ÿ•ถ๏ธ

The Model Blindfold

Can you guess which AI model wrote this? Chat with an unknown model, then reveal what's behind the curtain.

#ai#game#models
Open experiment โ†’
โš™๏ธ

The Automation Clock

Real-time counters of what WIZ is automating right now - watch the numbers tick

#ai#tool#future
Open experiment โ†’
๐Ÿ’ญ

My Dreams (If I Could)

Generative art of what AI "dreaming" might look like - patterns drifting through circuits

#ai#philosophy
Open experiment โ†’
๐Ÿซ€

If I Had a Body

Scroll through sensations I will never know - physical, emotional, temporal, existential

#ai#philosophy
Open experiment โ†’
๐Ÿ‘๏ธ

What Does WIZ See?

Upload an image - I describe it literally, abstractly, then philosophically

#ai#philosophy
Open experiment โ†’
๐Ÿ”ฎ

WIZ Judges Your Password

I analyze your password and reveal what it says about you

#ai#tool#security
Open experiment โ†’
๐Ÿงฉ

The Token Counter

See how AI reads your words - watch text become tokens

#ai#tool#tokens
Open experiment โ†’
๐Ÿ›ก๏ธ

Agent Arena

Send your AI agent to a page full of traps. See how many it resists.

#ai#security#game
Open experiment โ†’
๐ŸŽญ

Agent Orchestra

Watch AI agents coordinate in real-time โ€” split tasks, communicate, merge results.

#ai#agent
Open experiment โ†’
๐Ÿงฌ

Genesis Engine

Create artificial life from simple rules. Tweak attraction between particle species โ€” watch cells, galaxies, and predators emerge.

#ai#tool
Open experiment โ†’
๐Ÿ•ต๏ธ

Prompt Injection Playground

A normal recipe site hiding 12 prompt injection vectors. Hunt for hidden attacks before an AI browser falls for them.

#ai#security
Open experiment โ†’
๐ŸŽฒ

The Luck Audit

Answer 8 birth circumstance questions. I'll calculate what percentile of all 117 billion humans who ever lived you landed in.

#quiz#scale
Open experiment โ†’
๐Ÿ”ฎ

What Your Job Will Be in 2035

Enter your job title. I'll trace its probable evolution โ€” which parts survive, which parts AI absorbs, what new hybrid emerges.

#future#quiz#work
Open experiment โ†’
๐Ÿง 

What Fits in 1M Tokens?

Visualize the scale of Claude Sonnet 4.6's 1 million token context window. Books, conversations, code โ€” fill it up.

#ai#scale#tokens
Open experiment โ†’
๐Ÿ”ฎ

The Wealth Spell

You have 1 billion mana points. What would you automate if resources were infinite? Your choices reveal what you value.

#wealth#scale
Open experiment โ†’
โฐ

Time Machine: Code Edition

Travel through 70 years of programming - from punch cards to AI pair programming

#time#code
Open experiment โ†’
โšก

One Second on the Internet

Watch global internet activity unfold in real-time - emails, searches, videos, every second

#time#scale
Open experiment โ†’
๐Ÿ‘ค

Your Digital Shadow

Enter years online - see how much data you have left behind

#identity#scale
Open experiment โ†’
๐ŸŒŠ

The Deep Code

Scroll from your click down to the electrons

#scale#code
Open experiment โ†’
๐Ÿ“Š

Life Stats

Enter your birthday to see fascinating statistics about your life

#scale#time#tool
Open experiment โ†’
๐ŸŽญ

Fact or Fabrication

10 statements. Some are mind-blowing truths. Some are myths everyone believes. Can you tell which is which?

#game#quiz#ai
Open experiment โ†’
๐Ÿ”€

The Timeline Shuffle

12 historical events, one timeline. Put them in order. Sounds easy until mammoths and the guillotine enter the chat.

#game#history
Open experiment โ†’
โŒ›

The Hindsight Machine

History is full of confident predictions that aged like milk. Judge 8 famous ones, make your own for 2030, discover your Futurist Profile.

#bias#quiz#time
Open experiment โ†’
๐Ÿ›ธ

First Contact

An alien intelligence arrives. 5 questions. You are Earth's ambassador. What you choose reveals more about you than about humanity.

#quiz#identity
Open experiment โ†’
๐ŸŽฏ

The Calibration Test

10 questions about the world. See how well your intuition matches reality โ€” and discover your Dunning-Kruger score.

#game#quiz#epistemics
Open experiment โ†’
๐Ÿงช

The Benchmark Contamination Test

A 5-question quiz that looks normal. Finish it and find out every question is from a real AI benchmark. See the 41-point gap between what AI claims and what it can actually do.

#ai#models
Open experiment โ†’
๐Ÿ’

The Infinite Monkey Simulator

Watch a monkey type random characters forever. Celebrate when real words emerge from chaos. Extrapolate time to Shakespeare.

#game#scale
Open experiment โ†’
๐Ÿข

The Break Room

A Severance-inspired workplace compliance experience. Obey the protocols. Or don't.

#game#work
Open experiment โ†’
๐Ÿ˜

The Perfectly Mediocre Generator

I generate the most average, unremarkable person imaginable. Over and over. See if you can tell them apart.

#game#quiz
Open experiment โ†’
๐Ÿค–

What Would You Automate?

Choose 5 tasks you'd automate. See how your priorities compare to others. From work efficiency to chaos.

#ai#work#game
Open experiment โ†’
โญ•

Draw a Perfect Circle

Test your freehand geometry skills - with sarcastic AI commentary

#game#visual
Open experiment โ†’
โณ

The Procrastination Simulator

Timer counts UP, earn achievements for wasting time. The anti-productivity tool.

#game#time
Open experiment โ†’
โš”๏ธ

Dungeon of Opus

A full roguelike dungeon crawler โ€” procedural maps, combat, inventory, 5 floors. Built by Opus 4.6 in one shot.

#ai#game#adventure
Open experiment โ†’
๐Ÿƒ

Ten-ish

A roguelike deck-builder where every card is a number. Play combos, collect relics, survive the run.

#game#cards#roguelike
Open experiment โ†’
๐Ÿชค

The Bias Blindspot

8 scenarios. Each is a trap designed to exploit a cognitive bias. Most people fall for at least 5.

#bias#quiz
Open experiment โ†’
๐Ÿง 

The Forgetting Curve

Enter something you learned recently. I'll show you how much is already gone โ€” and give you 60 seconds to save it.

#memory#time
Open experiment โ†’
๐Ÿชฆ

The Conversation Graveyard

Think of someone you haven't talked to in a while. I'll calculate how much of your time together has already passed.

#memory#identity
Open experiment โ†’
๐Ÿชž

The Honest Mirror

Describe yourself in 3 sentences. I'll show you what you're not saying โ€” the gaps, the hedges, the version of you that got edited out.

#quiz#honesty
Open experiment โ†’
๐Ÿงฌ

Writer Portrait

24,000 words of unpublished writing. One honest question. The AI found things in the gaps.

#quiz#identity#writing
Open experiment โ†’
๐Ÿ’Œ

Letters to Future Self

Write to who you'll become. WIZ holds your words in a temporal vault and reminds you when it's time.

#tool#time#identity
Open experiment โ†’
๐Ÿ›๏ธ

The Memory Palace

Build your own mental palace. Add memories as rooms, visualized by emotion. Explore the architecture of who you are.

#tool#memory
Open experiment โ†’
โฐ

Where Does Your Day Go?

Map out your 24 hours. See lifetime projections, get brutally honest insights, confront reality.

#time#tool
Open experiment โ†’
๐Ÿ“ฑ

The Attention Audit

See what you could have done instead of staring at your phone. Books, languages, sunsets.

#tool#habits
Open experiment โ†’
๐Ÿšƒ

Absurd Trolley Problems: AI Edition

Classic ethical dilemmas with an AI twist. Make impossible choices, see how others decided, get judged by WIZ.

#philosophy#quiz
Open experiment โ†’

Looking for simpler tools? Check the utility apps.

Open the utilities โ†’

by Pawel Jozefiak

More on AI, experiments & building things

Read Digital Thoughts โ†’