Subexponential distribution (heavy-tailed) Created 2026-10-14 Updated 2026-11-05
For nonnegative independent and identically distributed random variables with unbounded support, this is the displayed two-summand tail asymptotic. The probability that the maximum exceeds is , so the sum and maximum have asymptotically the same tail. This is the basic single-large-arrival effect relevant to queue overflow. The concentration-theory sub-exponential random variable uses a separate light-tail convention.
Sub-Gaussian random vector Created 2026-10-06 Updated 2026-11-05
A centered random vector is sub-Gaussian with variance proxy when the displayed inequality holds for every deterministic vector . Equivalently every scalar inner product is a sub-Gaussian random variable with proxy . independence of coordinates is not required. For vectors with Euclidean norms at most one, the union bound gives with probability at least . This assumption alone is weaker than concentration for all Lipschitz functions.
Substitution of indifferent lotteries 2026-10-25
Apply the independence axiom for lottery preferences in both weak directions of the indifference relation. At the two mixtures are identical. Repeating this substitution and using transitivity lets one replace any finite collection of indifferent component lotteries, after which reduction of compound lotteries combines their outcome probabilities.
Success-failure stopping identity Created 2026-10-18 Updated 2026-10-20
In independent Bernoulli trials with success probability and failure probability , let be the trial of success and the trial of failure , for positive integers . If success occurs after failures, its final trial is a success and the preceding trials have successes and failures. The negative binomial stopping argument therefore gives the title sum.
Put . By trial , at least one target must have been reached, so the same event is that a binomial distribution with trials has at least successes. Conditioning instead on the final trial of a block of trials givesIndeed, more than successes in the full block guarantee the success target won; with exactly successes it won only when the final trial was a failure. These disjoint cases explain the extra boundary term. A zero success target wins immediately; a zero failure target has already been reached. To include in the displayed finite sum requires the explicit generalized convention ; an upper summation limit of denotes an empty sum.
Sudden oscillator frequency quench 2026-10-20
A jump from positive frequency to keeps an oscillator mode and its first derivative continuous. A normalized positive-frequency incident solution has outgoing coefficients and , satisfying . This temporal matching is not a sum-one spatial-scattering probability rule.
Sufficient statistic Created 2026-09-24 Updated 2026-10-14
A statistic is sufficient for a statistical model when the conditional distribution of the full data given has a version independent of the unknown statistical parameter . Once is recorded, the remaining uncertainty about the data's arrangement can be generated without knowing . [ref]
Sufficiency (Stat 210a, Fall 2025) by Will Fithian. Sufficiency, Will Fithian, “Sufficiency”: definition
For data and statistic taking values in standard Borel spaces, express this precisely with one Markov kernel , common to every statistical parameter value:for each measurable event . Thus is a regular conditional distribution of given . Equality is only required outside a null set for the distribution of under each ; the choice of version matters on impossible conditioning events. For a continuous probability distribution, the event may have zero probability even at meaningful observations, so the definition uses a regular conditional distribution, rather than division by .
A sufficient statistic is a property of a specified statistical model, rather than of a formula for summarizing data in isolation. It need not be an unbiased estimator or even shorten the record: the full data is always sufficient. An estimator instead targets a quantity such as a mean. An ancillary statistic has a parameter-independent marginal probability distribution, whereas a sufficient statistic has a parameter-independent conditional full-data law. A minimal sufficient statistic permits the greatest reduction among sufficient statistics. A complete statistic satisfies a separate condition about zero expectations; a complete sufficient statistic must satisfy both conditions.
Summable and harmonic event probabilities 2026-10-14
Take independent random variables , each with the continuous uniform distribution on , and set . Then , and the events are mutually independent events. The product probability space supplies a model for the entire infinite sequence; a finite list of simulated values is only an illustration.
For , the p-series is summable:The Basel problem supplies the total sum, and the integral test for convergence gives the displayed tail bound. The First Borel-Cantelli lemma makes the limsup of events have probability zero. In fact, the tail bound implies that the probability of even one alarm at or after is at most ; for , it is at most . This bound remains valid for dependent events with the same individual probabilities.
For , the harmonic series diverges, so the Second Borel-Cantelli lemma makes the limsup of events have probability one. There is also an exact finite-horizon calculation using independent events:For any fixed , this finite product of numbers tends to zero as . Every infinite tail therefore contains an alarm almost surely, even though the individual probabilities tend to zero.
Support of a probability distribution Created 2026-10-06 Updated 2026-11-05
For a probability measure on a topological space, its support isA point belongs to the support when observations can fall arbitrarily close to it with positive probability. This is a neighbourhood condition, not a requirement that the point itself have positive probability mass or that a chosen density be positive there.
For the uniform distribution on viewed as a law on the real line, the support is , including both endpoints. A density on vanishes at zero, but zero still lies in its support because every open neighbourhood of zero has a positive integral of . A discrete distribution with mass at , , likewise has zero in its support despite assigning it probability mass zero.
If a distribution's support equals an interval, every open subinterval of that interval has positive probability. The weaker condition does not imply that the support equals : a uniform law on also assigns probability one to , but assigns zero to its subinterval . This distinction matters when using support to describe feasible observations or randomized effort in a contest.
Surface-code decoding as a random-bond Ising model Created 2026-09-28 Updated 2026-11-05
For independent bit flips, summing the probabilities of all error chains in one syndrome and logical class equals, up to a class-independent factor, the partition function of a random-bond Ising model. A reference error fixes bond signs and products of star stabilizers become Ising spin configurations.
Surprise Created 2026-10-24 Updated 2026-11-05
Surprise is an emotion associated with an unexpected event or a discrepancy from an expectation. It does not determine pleasantness: an unexpected welcome visitor and an unexpected loss can both be surprising while having different emotional valence. Research can manipulate event probability or prior information and then measure reported surprise, orienting or changes in prediction. These are related operational measures rather than one interchangeable signal. [ref]
Start traps at a unit-intensity Poisson random measure in Euclidean space and attach mutually independent Brownian motions, independent of the initial measure. For a deterministic target path , independently thin initial positions according to whether their marked path comes within radius one of the target by time . The retention probability integrates, by the Tonelli theorem, to the expected volume of the sausage along . Symmetry of Brownian motion makes this the expected volume of the sausage along . The Poisson zero-count probability proves the displayed survival formula. This derives the formula directly from independent thinning, without assuming a separate marking or displacement theorem.
Survival function Created 2026-09-24 Updated 2026-10-14
For a nonnegative survival time , the survival function iswhere is the cumulative distribution function. Thus is the probability that the chosen occurrence is still in the future at time . We consistently use the strict inequality : a probability mass at has already left the survival function at itself. With this convention is a right-continuous function and a nonincreasing function taking values in . If , then ; if is finite with probability one, then as .
For and , conditional probability givesFor example, and mean that of those still event-free at year two remain event-free through year five. Their intervening event probability is , whereas the original population's probability of an occurrence in is . Conditioning changes the denominator.
The area under the survival function up to a horizon gives the restricted mean survival time. A positive terminal plateau in a plotted Kaplan–Meier estimator does not establish that the occurrence will never happen: terminal censoring and survival-mean identifiability explains why the unobserved tail remains undetermined.
Survival likelihood Created 2026-09-24 Updated 2026-11-05
For a continuous event time with probability density function , survival function and hazard function , right censoring records and . Suppose and are conditionally independent given the observed covariates, suppressed here. Write and for the censoring-time probability density function and survival function. The joint observed-data likelihood contribution isAn event contributes its event-time probability density function times the probability of remaining uncensored until then. A censored observation contributes its censoring-time probability density function times the probability of remaining event-free until then.
If the censoring model has separate parameters with no constraints linking them to , its factor can be omitted when maximizing over . The event-model part for independent subjects is thenwhere is the cumulative hazard function. Independent censoring alone does not permit dropping a censoring term that itself contains the event-model parameters. The familiar event-versus-censoring formula is thus a factorization under an ignorable observation mechanism, rather than an unconditional rule for every missing event time.[ref]
Survival prediction calibration 2026-10-07
Calibration compares predicted survival probabilities with observed survival experience at relevant horizons or over time. Censoring must be handled, for example by suitable survival estimates or inverse-probability weighting. Calibration should be assessed on validated predictions, because the fitting sample can hide optimism from variable selection and tuning. It is distinct from the risk-ranking information measured by a concordance index.
Survival probability Created 2026-09-28 Updated 2026-11-05
For a nonnegative first-passage time , possibly infinite, the survival probability isIt describes the Probability that absorption has not yet occurred. The tail-integral identityholds as an identity of extended nonnegative quantities, even when absorption is not almost sure or the expected time is infinite. Indeed, pathwise, and Tonelli's theorem exchanges expectation and integration. For an integer-valued hitting time, the corresponding formula is .
If the finite-time distribution is absolutely continuous, its first-passage-time density is . Where , its hazard function is . Constant hazard gives and mean .
The limit of as is the Probability of never being absorbed; a positive limit forces infinite mean. Conversely, zero limit does not guarantee a finite mean. Brownian first passage to a positive level occurs almost surely, but its survival tail is proportional to , whose integral diverges.
Survival probability of a branching process 2026-10-07
For a Galton-Watson process started from one ancestor, this is the probability that every generation is nonempty. It is the complement of the branching extinction probability. The binomial branching survival correction distinguishes finite-binomial reproduction from the Poisson branching process near mean one.
Survival probability of the contact process Created 2026-09-28 Updated 2026-11-05
The Survival probability is the probability that the Contact process begun from one infected vertex never reaches its absorbing empty state. Its critical infection rate is .
Survival probability under risk pooling Created 2026-10-06 Updated 2026-11-05
independent positively loaded portfolios with zero initial capitals have separate joint survival probability . Pool their claims and premium incomes to obtain zero-capital merged survival . This is the premium-weighted average of the individual survival probabilities and is at least their product. The event of aggregate solvency permits transfers of surplus between the original portfolios.
Swap test Created 2026-09-24 Updated 2026-11-05
The swap test estimates the overlap of two pure states without requiring their classical descriptions. Prepare an auxiliary qubit in , apply a Hadamard gate, a Fredkin gate controlled by that auxiliary qubit, and another Hadamard gate. Then measure the auxiliary qubit in the computational basis.
For target states , the auxiliary-zero branch has unnormalized target vector . Its squared norm is thereforeIdentical pure states give , while pure states with zero inner product give . A single zero outcome does not certify equality; repeated trials estimate the probability. The calculation is a direct projection onto the symmetric target subspace using the SWAP gate.
Symmetric stochastic attrition process Created 2026-10-14 Updated 2026-11-05
This finite-state Markov jump process models two opposing populations, each suffering losses at a rate proportional to the surviving opposing population. The displayed rates apply when and ; both coordinate axes are absorbing. Its embedded jump chain removes one individual at each step, with probabilities and . Before the first population is exhausted, the corresponding deterministic equations , conserve ; finite stochastic populations satisfy a near-invariant instead.
