Intellect PlusStatistics and probability
180 cards. Each one is a 40-second story with a five-minute read behind it. Page 3 of 3.
Norbert Wiener earned a Harvard doctorate at 19 and then founded cyberneticsNorbert Wiener finished high school at 11, collected a mathematics degree at 14 and a Harvard PhD at 19. He went on to argue that intelligent behaviour comes down to feedback loop…
One hot oven can wreck the average temperature of a roomMeasure ten objects in a room: nine sit between 20 and 25 degrees Celsius, and one is an oven at 175. The median stays sensibly between 20 and 25, but the mean jumps to somewhere …
The Mann-Whitney test is a tournament between two groups of numbersWant to know whether one group tends to score higher than another without assuming a bell curve? Pair every value in the first group against every value in the second, award a poi…
The doubling bet that inspired a branch of probabilityEighteenth-century French gamblers loved a system called the martingale: double your stake after every loss, and the first win recovers everything plus a profit. Given unlimited m…
KL divergence measures the cost of believing the wrong modelSuppose you design a code for messages assuming one pattern of letter frequencies, but the real messages follow another. On average you will waste some bits per symbol. That waste…
Kurtosis measures a distribution's outliers, not how pointy its peak isMany textbooks describe kurtosis as peakedness, the sharpness of a curve's summit. Statisticians now agree that description is wrong. Kurtosis is about the tails: how prone a dist…
Likelihood is not probability, and confusing them can be disastrousToss a coin twice and get two heads. If the coin is fair, that outcome had a 0.25 chance, so the likelihood of fairness given those tosses is 0.25. It is tempting to conclude ther…
In regression, a factor can matter hugely alone yet vanish with companyLinear regression looks simple: draw the best straight-line relationship between an outcome and its causes. But the coefficient it reports for any one factor depends on what else …
A single die roll shows why joint probabilities are not simply multipliedRoll one ordinary die and ask two questions at once: is the result even, and is it prime? Only the 2 answers yes to both, so that combination has a one-in-six chance. A joint prob…
The clustering method born at Bell Labs waited 25 years for publicationIn 1957 Stuart Lloyd of Bell Labs devised a simple loop for squeezing signals into fewer values: sort points to their nearest centre, move each centre to the average of its points…
The Kolmogorov–Smirnov test judges data by its single biggest gapHow can you tell whether a pile of measurements really follows the curve you expected? One classic answer, devised by two Soviet mathematicians in the 1930s, ignores averages enti…
A histogram's story depends on how wide you cut the binsWhen the US Census Bureau charted how long 124 million people take to get to work, one bar stood out: 30 to 35 minutes was more popular than the slots on either side. Commuters ha…
Infinite monkeys will type Hamlet, but a real universe of them never couldGive a monkey a typewriter and eternity, and it will almost surely produce the complete works of Shakespeare, not once but infinitely many times. Shrink the setup to something phy…
How attention lets a translation model look back at the right wordAsk a neural network to translate I love you into French and watch where it looks. To produce je it puts 94% of its focus on I. For the next word it jumps ahead to you and writes …
Guessing the weather from what a friend did all dayAlice and Bob live far apart and chat by phone every evening. Bob only ever walks in the park, shops or cleans his flat, and which he picks depends on the weather. Alice never see…- Actuaries turn the question of when people will die into a priceSelling someone a life insurance policy or a pension means promising money decades ahead, on a date nobody knows. Actuarial science grew up in the late 1600s to make that promise …
The 1954 translation machine that fooled headlines knew just 250 wordsIn 1954 newspapers hailed a bilingual robot brain turning Russian into the King's English. The Georgetown-IBM demonstration actually translated 49 hand-picked sentences using a 25…- How statisticians charge a model for every extra knob it addsGive a statistical model more adjustable parameters and it will almost always hug your data more tightly, even when it is just memorising noise. The Japanese statistician Hirotugu…
Why mathematicians cannot give every set a size, and what they do insteadTry to assign a length to every possible subset of the real line and you hit a wall: the axiom of choice guarantees sets, such as the Vitali sets, that have no sensible length at …
APA style grew from a seven-page memo into a manual with a political edgeIn 1929 psychologists printed a seven-page guide for authors. That modest memo became the Publication Manual of the American Psychological Association, now in its seventh edition …
Why a six-sigma day in daily data should come once in 1.4 million yearsIn a normal distribution about 68 per cent of values fall within one standard deviation of the average, 95 per cent within two and 99.7 per cent within three. The shortcut is hand…- Standard deviation measures how spread out values areStandard deviation gauges how tightly values hug their arithmetic mean. Take the square root of average squared gaps from that mean, and you recover a spread figure in the same un…
Standard error tracks uncertainty in sample estimatesStandard error is the standard deviation of a statistic's sampling distribution, often the wobble of the sample mean. It equals population sigma over square root of n, so four tim…
Hypothesis tests weigh evidence against a null claimA statistical hypothesis test asks whether sample data give enough evidence against a default claim about a population. Roughly one hundred specialized tests exist; each turns raw…
Statistics turns raw data into tested conclusionsStatistics gathers, organizes, analyzes, interprets, and presents data, starting from a population or model. The German Statistik once meant describing a state; by the 1790s John …
Stochastic processes model systems that change randomlyA stochastic process is a family of random variables on a probability space, usually indexed by time. From bacterial growth to thermal noise to gas molecules, it models systems th…- Guinness, a pseudonym, and small-sample meansStudent's t-test asks whether a mean difference is statistically meaningful when a scaling term must be estimated from data. William Sealy Gosset published the method in 1908 as "…
Survey methodology turns questions into population inferencesSurvey methodology is the applied statistics of human-research surveys—designing samples, building questionnaires, collecting responses, and carefully adjusting results so a chose…- Squared spread that algebra loves, units hateVariance measures how far numbers sit from their average: the expected squared gap from the mean. Take its square root and you get standard deviation—friendlier everyday units, fu…
- Fit the line by squaring every vertical missOrdinary least squares picks the coefficients of a linear regression by minimising the sum of squared gaps between observed responses and the model’s predictions. Geometrically th…
- A small p is not the chance your hypothesis is trueIn null-hypothesis significance testing, a p-value is the probability of seeing a result at least as extreme as the one observed if the null hypothesis were true. Tiny p-values ma…
A French mathematician named rare random piles of eventsCall centers, horse kicks, and falling stars share a quiet math pattern. When events arrive independently at a steady average rate, their counts often follow a Poisson distributio…- Probability once meant a witness's social standingToday a fair coin sits at one half. Centuries ago, Latin probabilitas could mean a witness's legal weight, often tied to nobility. Modern chance instead weighs evidence through co…
A continuous outcome never hits one exact numberAsk for the chance a bacterium dies at exactly five hours and the answer is zero. Continuous chance lives in intervals, not pinpoints. A probability density function turns height …
Every random experiment carries a map of likelinessA fair coin is not just heads or tails—it is a distribution that splits probability in half. Distributions assign chances to events so the axioms hold. Discrete cases use masses o…- Chance became rigorous when Kolmogorov wrote axiomsGamblers asked first; physicists later found atoms behaving randomly. Probability theory turns that fog into a measure between zero and one on a sample space. Two classic theorems…
- A random variable is a function, not a wild numberDespite the name, a random variable is a measurable map from outcomes to numbers. Heads might become one and tails zero. The distribution that map pushes forward forgets the under…
Random assignment turned medicine into a fair fightPatients differ in ways no checklist can catch. Randomized controlled trials shuffle people into treatment and comparison arms so bias and confounders lose their grip. A 1948 stre…- Regression first meant children of tall parents shrinkFrancis Galton coined "regression" for heights drifting toward the average. The math toolkit now fits lines and curves between outcomes and predictors. Least squares grew from tra…
Two million votes still picked the wrong presidentIn 1936 the Literary Digest surveyed over two million people and forecast a Republican win—then lost badly. Magazine and phone lists leaned Republican, so sheer size could not sav…
Tomorrow depends only on today—history drops outA Markov chain is a random sequence whose next move depends only on the present state, not the full past. That “memoryless” structure powers weather sketches, text generators, and…- The middle value that extreme outliers barely touchThe median splits ordered data into equal upper and lower halves. Unlike the mean, a few extreme values cannot yank it away—so median income often describes a center that billiona…
Pooling many studies into one sharper effect sizeMeta-analysis combines quantitative results from independent studies that ask the same question, extracting effect sizes and variances to form a pooled estimate. Extra statistical…
The value that shows up most—even for names, not numbersThe mode is the most frequent value in a sample, or the peak of a probability mass or density. Unlike mean and median, it works for categories: among Korean surnames, “Kim” can be…
The bell curve whose averages inherit normalityA normal—or Gaussian—distribution is the continuous bell shaped by mean μ and variance σ². The central limit theorem explains why measurement errors and many sums look nearly norm…
When you cannot randomize, you still watch and infer carefullyObservational studies draw conclusions without assigning the treatment themselves—ethics, politics, or rarity get in the way. Unlike randomized trials, who gets exposed is not und…- Fisher taught science to plan variation on purposeDesign of experiments builds procedures that show how changing some parts of a system shifts others. From Charles Peirce's 1880s blinded weight trials to Ronald Fisher's 1926 and …
- Fair dice are the textbook discrete uniformA discrete uniform distribution gives each of n known finite outcomes the same probability 1/n. A fair six-sided die—faces 1 through 6, each with chance 1/6—is the classroom examp…
- A fair die averages 3.5—not a face you can rollExpected value is the probability-weighted average of outcomes, not a result you ever have to see. A fair six-sided die has expectation 3.5; American roulette’s single-number bet …
- The waiting-time law that forgets how long you already waitedExponential waiting times model gaps between events in a Poisson stream—calls, machine failures, fabric defects—at a constant average rate. Mean and standard deviation both equal …
- Many rolls settle the average—streaks do not “owe” a correctionThe law of large numbers says sample averages of independent, identically distributed draws converge to the true mean when that mean exists. Casinos rely on it; gamblers’ fallacie…
"Average" can mean mean, median, or mode—choose carefullyAn average is a value meant to sit at the center of a collection. People usually mean the arithmetic mean—sum divided by count—but teachers also lump in the median and mode. For s…
Bayes flips "effect given cause" into "cause given effect"Bayes' theorem, named for Thomas Bayes and refined with Pierre-Simon Laplace, turns a familiar conditional probability around. Knowing how often a positive test appears when disea…
Bayesian inference updates belief as each new clue arrivesBayesian inference uses Bayes' theorem to turn a prior guess about a hypothesis into a posterior once evidence lands. Science, medicine, law and sports all lean on this loop: mult…- Coin flips add up: the binomial counts yes–no successesFlip a biased coin six times and ask for exactly four heads: the binomial distribution answers. It models how many successes appear in n independent Bernoulli trials that each suc…
- Sample means turn normal—even when raw data do notThe central limit theorem says that a properly scaled sample average drifts toward a bell curve as the sample grows, even if each observation comes from a skewed or jagged law. Th…
- Pearson's χ² asks if table frequencies match a null storyA chi-squared test compares what you counted in categories with what a null hypothesis predicted. For large contingency tables it checks whether two categorical factors act indepe…
Correlation tracks linear co-movement—not proof of causeWhen two quantities rise and fall together, statistics speaks of correlation—usually a linear link measured by Pearson's coefficient. Utilities may dial down generation on mild da…- Covariance signs the shared dance of two random variablesCovariance measures how two random variables wobble together. Positive values mean they tend to rise and fall in step; negative values mean they move oppositely. Because units cli…
Good charts let the eye discover what tables buryData and information visualization turn numbers and abstractions into graphics people can explore. Charts, maps and interactive displays surface patterns, outliers and stories—whi…