Rayleigh moments Created 2026-10-18 Updated 2026-10-20
Using in the defining expected value integral givesThe last integral is the gamma function. Near zero, the integrand before substitution behaves as , explaining the threshold ; at infinity the Gaussian tail controls every power. In particular,[ref]
The Rayleigh Distribution. Random: Probability, Mathematical Statistics, Stochastic Processes by Kyle Siegrist; The General Rayleigh Distribution, Moments: mean, variance and nonnegative integer moments; extension to real q greater than -2 follows from the displayed integral
For the chosen , the mean radius is about , its median is , and its mode is . Its root mean square is . Different summaries answer different questions; the component scale is neither the radial mean nor the radial standard deviation.
Return period Created 2026-10-06 Updated 2026-10-14
A return period expresses how rare a threshold exceedance is in a specified time block. Let be the sample maximum in block , fix a threshold , and suppose is the same for every block. The conventional return period is blocks; blocks of duration give a time of . Thus an annual exceedance probability means a 100-year return period. This names a frequency scale, not a schedule or a guarantee of one exceedance in each century. The annual convention and the misleading “once every 100 years” interpretation are discussed in USGS: the 100-year flood.
To derive a waiting-time interpretation, assume the exceedance indicators are independent and identically distributed random variables with the Bernoulli distribution of parameter . Define , counting the successful block as well as all preceding blocks. The first Bernoulli trials must fail and the th must succeed, so independence gives the geometric distributionThe expected value can also be read directly from the geometric series: . This counts blocks from the starting boundary, not the exact time of an exceedance within its block. Counting only failed blocks instead would give expected value . The waiting law and its memorylessness of the geometric distribution are developed in Kyle Siegrist, Random Services: the geometric distribution.
For a horizon of complete years, at least one exceedance is the complement of no exceedances. Under the same independence and constant- assumptions,For the toy choice , this gives in 10 years, in 30 years, in 50 years, and in 100 years. Even after 100 years there is a chance of no exceedance. These numbers are exact model calculations rounded to one decimal place, not measured frequencies. The median waiting count is 69 years, since ; the expected value is 100 years. A mean and a median answer different questions.
The exceedance count has the binomial distribution under independence, with expected value . Thus has expected value one, although its probability of being at least one is only : some realizations contribute several exceedances. For , the memorylessness of the geometric distribution gives . A long quiet spell does not make the next independent year's exceedance overdue, and an exceedance this year does not force a quiet interval. Two specified consecutive years both exceed with probability in this toy model.
A reproducible tabletop experiment makes the assumptions tangible. Put 100 identical slips numbered 1 to 100 in a bag; after thoroughly mixing, draw one and record whether it is 100. Replace the slip and mix again before every draw. Treat each draw as one toy year, so the exceedance indicator is a Bernoulli random variable with , provided the sampling is uniform and successive draws are independent. Repeat complete 30-draw runs and record the fraction containing at least one 100; separately repeat runs stopped at the first 100 and record their waiting counts. The first fraction targets ; the waiting-count histogram targets . Replacement keeps fixed, mixing is intended to support independence, and repeated runs expose sampling variation. This is a proposed demonstration of the toy model; no experimental results are asserted here.
Stationarity supports a time-invariant annual probability distribution, but does not imply independence. For an extreme counterexample, draw just one Bernoulli random variable of parameter and set for every year. This is a stationary process, yet the probability of any exceedance in every finite horizon is ; from an arbitrary starting boundary the wait is infinite with probability . The reciprocal marginal frequency therefore need not equal that forward waiting expected value. With equal marginal probabilities and arbitrary dependence, only the general bounds follow, the upper bound being the union bound. Linearity of expectation still gives without independence. Mere zero correlation does not restore the product formula; even pairwise independence does not imply joint independence. If annual probabilities change to , independence instead gives , and one fixed return period no longer describes all years.
Robust estimator 2026-11-05
A robust estimator is an estimator whose behavior remains acceptably stable under specified departures from an assumed statistical model. “Stable” must identify a criterion, such as limited bias under contamination, bounded response to a single extreme observation, or resistance to replacing a fraction of the sample. It does not mean that the estimator is accurate under every possible probability distribution.
For example, consider data with . The sample mean is , whereas the sample median, using the midpoint convention for an even-sized sample, is . Sending to infinity makes the sample mean diverge without changing that sample median. The example concerns location estimation: the population median and population mean need not coincide outside a symmetric reference model.
One formal model is Huber contamination:where is a reference probability distribution, is an unspecified contaminating probability distribution, and is its proportion. A robustness claim can bound an estimator’s worst bias over allowed , rather than only calculating its variance when the reference model is exactly true. Peter J. Huber’s original location paper treats this tradeoff systematically.[ref]
Robust Estimation of a Location Parameter by Peter J. Huber. Introduction, contaminated normal location models; section 4(iii), clipped location criterion
A Huber location estimator illustrates how stability enters the calculation. With a specified positive scale , solveLarge residuals have bounded scores, so one increasingly extreme point cannot contribute an arbitrarily large term to this equation. By contrast, the squared-error score grows linearly with the residual and yields the sample mean. A practical analysis must also state how is obtained: a scale estimate that itself explodes under outliers can undermine the intended protection.
Different criteria answer different questions. An influence function studies an infinitesimal perturbation of a distribution, while a breakdown point concerns a finite amount of sample corruption. A robust estimator can sacrifice some model-specific statistical efficiency to improve performance away from that model, but the amount depends on the chosen tuning and departure. Reporting the target parameter, reference model and robustness criterion makes the claim testable.
Robust statistics Created 2026-09-24 Updated 2026-11-05
Robust statistics studies how estimators and other statistical procedures respond to departures from a reference statistical model. Such departures can include a small fraction of outliers, a heavier-tailed probability distribution, or an incorrect assumed shape. A robustness statement specifies both the permitted departure and the quantity whose change is controlled.
For data with , the sample median remains , whereas the sample mean is and grows without bound. This does not make the sample median an estimator of every desired expected value: the mean and median of an asymmetric probability distribution can differ. The estimand must therefore be specified before judging robustness.
Influence functions describe infinitesimal changes, breakdown points describe large finite-sample failures, and minimax asymptotic bias compares worst cases over a neighborhood. M-estimators connect these questions with an explicit loss function. Peter J. Huber's 1964 location paper developed a systematic treatment of estimation under contamination.[ref]
Robust Estimation of a Location Parameter by Peter J. Huber. Introduction and location estimation under contaminated normal models
Sample median Created 2026-09-28 Updated 2026-10-14
The sample median is a central value of an ordered statistical sample. The usual numerical convention selects the middle order statistic for odd sample size and the arithmetic mean of the two central order statistics for even sample size:NIST uses this convention. [ref]
https://www.itl.nist.gov/div898/handbook/eda/section3/eda351.htm. NIST: measures of location, median definition
. For , the sample median is ; for , it is , which need not be an observed value.More generally, every value between the two central order statistics in an even statistical sample is a median of its empirical distribution function. The generalized-inverse sample quantile at selects the lower central order statistic in that case, so it can differ from the midpoint convention. Ties are allowed in all these definitions.
The sample median measures the center using sorted positions. Changing the largest observation from to in the five-value example leaves the sample median at , while the sample mean rises from to . Its sampling distribution still depends on the population probability distribution: the mean found for five observations from a uniform distribution is a special calculation, and a sample median's expected value need not equal the population median in general.
For a concrete counterexample, three independent and identically distributed random variables with rate-one exponential distribution have sample median with expected value , by independent exponential order-statistic spacings. The population median instead solves , giving . Thus even the middle order statistic can have mean different from the population median.
Sign test Created 2026-09-24 Updated 2026-11-05
The sign test uses only the signs of independent differences . Under the Null hypothesis and no ties, the count has the binomial distribution . It can test a continuous population median against a specified value, or a median paired difference against zero, without specifying the magnitude distribution of the differences.
For a two-sided test, one conventional exact p-value is for . Seven positive differences among eight give . The test is robust to the sizes of extreme differences because only signs are used, but discarding magnitudes can lose power relative to a suitable parametric test.
With zero differences, deleting ties gives an exact conditional sign calculation when the nonzero signs are equally likely under the Null hypothesis. For a discrete population, merely declaring zero a median does not by itself imply that conditional sign symmetry. Dependence between pairs also invalidates the independent binomial calculation.
Talagrand concentration inequality for certifiable functions Created 2026-09-24 Updated 2026-11-05
For a nonnegative integer-valued certifiable function of independent random variables, suppose changing one coordinate changes by at most and a value has a certificate involving at most coordinates, with . Then for ,For a high-value input , every low-value input must disagree with its certificate on at least coordinates: replacing its certificate positions by those of would change the low value by less than but force a value at least . Uniform weights on certificate positions therefore give Talagrand convex distance at least from the low-value event. Apply Talagrand's convex distance inequality to that event.
The bound is useful when one of the two events has known nonnegligible probability, for example at a median. A certificate alone supplies no coordinate-change constant ; both hypotheses are used in the proof. The two-threshold concentration for certifiable functions article develops the resulting mean-median comparison.
Two-threshold concentration for certifiable functions Created 2026-10-14 Updated 2026-11-05
Suppose an integer-valued certifiable function is -Lipschitz under one-coordinate changes and a value at least has a certificate of at most coordinates. Any input with value at most must differ on at least certificate positions. Equal weights on those positions bound its Talagrand convex distance from the low-value event. Applying Talagrand's convex distance inequality gives the displayed product bound. For , median tails integrate to , and deviations of order have exponentially small probability.
Unimodal distribution 2026-11-05
A unimodal distribution, under the density or mass-function convention, has probabilities that rise toward one central region and then fall away. For an absolutely continuous probability distribution, this means it has a probability density function that is nondecreasing up to a mode and nonincreasing afterward. This is the single-maximum meaning of unimodality; the chosen density is a one-dimensional quasiconcave function. A plateau of maximizing points is allowed.
For a discrete original example on outcomes0,1,2and3, choose probabilities0.1,0.4,0.4and0.1. The probability mass function rises, has two adjacent maximizing values, then falls. It is unimodal under the weak monotonicity definition despite not having a unique mode. In contrast, separated high-probability regions with a genuine dip between them are not one mode plateau.
Unimodality does not imply symmetry or equality of arithmetic mean, median and mode. For the exponential distribution with density on , , the mode is0, the mean is and the median is . The density decreases from its boundary maximum. The NIST handbook gives the family and these summary characteristics, using a scale parameter reciprocal to this rate parameter.[ref]
Exponential Distribution in the NIST e-Handbook. §1.3.6.6.7, Probability Density Function and summary statistics
Nor is unimodality generally preserved by a mixture distribution. Equal mixing of uniform distributions on and gives density1/2 on those two intervals and zero in between. Each component has a flat unimodal density, while the mixture has separated peaks.
A density is defined only up to changes on sets of measure zero. Thus the distributional statement concerns existence of a representative with the required shape, rather than artificial isolated spikes added to one displayed formula. In higher dimensions, several notions of unimodality are used; one convex-level-set formulation requires convex superlevel sets, as in the quasiconcave function definition.[ref] The shape convention should be specified before extending a one-dimensional conclusion to another setting.
Weighted mean minimizes squared deviations 2026-10-14
For a finite list of real numbers and nonnegative weights of positive total , the weighted mean uniquely minimizes the weighted sum of squared deviationsExpanding around and using givesEvery displacement from the weighted mean increases the objective. This is the simplest one-parameter weighted least squares problem: represent all entries by one common value.
Weighted median Created 2026-10-20 Updated 2026-11-05
A median of a weighted probability distribution. With positive weights summing to one, at most half the weight lies strictly below a weighted median and at most half lies strictly above it. The lower weighted median chooses the first observed value whose cumulative weight reaches one half; the convention matters when the cumulative mass is exactly one half.
Wilcoxon signed-rank test 2026-10-14
For independent paired observations, form differences . Assume the differences are identically distributed, continuous and symmetric about a location . Under the zero-location null hypothesis, the signs are independent fair signs conditional on their absolute values. Rank the absolute differences and add ranks with positive signs. A negative-location alternative rejects for small . Its exact conditional distribution is that of a sum of an independently selected subset of ranks, each of the subsets equally likely. Thus its mean is and its variance is . Symmetry is essential: a zero median alone does not justify this signed-rank null law. Zero differences are omitted in the usual version, and tied absolute differences receive average ranks.

