1 Definition and basic concepts

A cumulative distribution describes how probability builds up over the values of a random variable. It summarizes the full distribution in a form that allows probabilities for thresholds and intervals to be read directly or computed with simple rules. The central object in this setting is the cumulative distribution function, usually abbreviated CDF.

1.1 Cumulative distribution function

The cumulative distribution function of a random variable assigns to each value x the probability that the variable is less than or equal to x. For a variable X, this is written as F(x) = P(X ≤ x). The function increases as x moves through the number line, reflecting the growing set of possible outcomes included in the probability statement.

1.2 Random variables

A random variable is a numerical representation of an uncertain outcome. It may describe counts, measurements, waiting times, or other quantities. Cumulative distributions apply to both discrete random variables, which take separate values, and continuous random variables, which vary over intervals.

1.3 Probability accumulation

The idea of accumulation refers to adding probability mass or probability density up to a specified point. In the discrete case, this means summing the probabilities of all values at or below that point. In the continuous case, it means integrating the density over all values up to that point.

1.4 Notation and terminology

The notation F(x) is standard for a cumulative distribution function. The term distribution function is also used in many texts. Related expressions include cumulative probability and accumulated probability, although the exact meaning depends on context. The CDF is often treated as the most compact summary of a probability distribution.

2 Properties

Cumulative distribution functions have a small set of general properties that hold for all random variables. These properties make CDFs especially useful for both theoretical work and practical computation.

2.1 Nondecreasing behavior

A cumulative distribution function never decreases. If x1 is less than x2, then the probability that the random variable is at most x1 cannot exceed the probability that it is at most x2. This follows from the fact that the second event contains the first.

2.2 Limits at infinity

As x becomes very small, the CDF approaches 0 for ordinary distributions on the real line. As x becomes very large, the CDF approaches 1. These limits express the idea that eventually all probability lies below a sufficiently large threshold.

2.3 Right-continuity

Cumulative distribution functions are right-continuous, meaning that the value at a point matches the limit from values just to the right. This property is important in probability theory and helps distinguish CDFs from arbitrary nondecreasing functions.

2.4 Range and interpretation

The values of a CDF always lie between 0 and 1. A value near 0 indicates that little probability lies at or below the point, while a value near 1 indicates that most probability has already accumulated. The function therefore provides a direct probabilistic interpretation of location on the number line.

3 Discrete cumulative distributions

For discrete random variables, the cumulative distribution function has a step-like shape. Each possible value contributes a jump whose size equals the probability assigned to that value.

3.1 Step functions

Discrete CDFs remain constant between possible outcome values and jump upward at each outcome. The size and placement of the steps reflect the support of the distribution. This makes the graph easy to interpret visually.

3.2 Probability mass functions

The probability mass function, or PMF, gives the probability of each discrete value individually. The CDF is obtained by summing PMF values up to a given point. In turn, the PMF can be recovered from the sizes of the jumps in the CDF.

3.3 Example computations

If a discrete random variable can take values 1, 2, and 3 with probabilities 0.2, 0.5, and 0.3, then F(2) equals 0.7 because it includes the first two outcomes. Likewise, F(3) equals 1 because all probability has been included. Such calculations are direct and often used in elementary probability.

3.4 Jump sizes and point probabilities

Each jump in a discrete CDF corresponds to the probability at that exact point. The difference between the CDF values just after and just before a point gives the point probability. This feature distinguishes discrete distributions from continuous ones, where single points usually have probability zero.

4 Continuous cumulative distributions

Continuous random variables are described by CDFs that typically change smoothly rather than in jumps. Their probabilities are spread across intervals, not concentrated at isolated points.

4.1 Relationship to probability density functions

For a continuous random variable, the probability density function, or PDF, describes relative concentration of probability. The CDF is the accumulated area under the density up to a point. Differentiation and integration connect the two whenever the density exists.

4.2 Integration of density

If a density function f(x) is available, then the CDF is obtained by integrating f from negative infinity to x. This integral represents the total probability contained to the left of x. In many cases, the CDF is easier to use for probability statements than the density itself.

4.3 Smoothness properties

Continuous CDFs are often continuous and may be differentiable where the density is well behaved. However, not every continuous distribution has a smooth CDF everywhere. Some distributions have corners or other irregular features while still remaining continuous.

4.4 Interval probabilities

For continuous variables, the probability that X lies between a and b is found by subtracting CDF values: P(a < X ≤ b) = F(b) − F(a). In these cases, probabilities of exact points are usually zero, so only interval calculations are meaningful.

5 Mixed distributions

Some distributions combine discrete and continuous features. In such cases, the cumulative distribution function reflects both jumps and smooth segments.

5.1 Combined discrete and continuous components

A mixed distribution may place positive probability on specific points while also spreading probability continuously over intervals. The CDF then has step increases at the point masses and gradual growth elsewhere. This type of distribution occurs in models that include thresholds, fixed outcomes, or measurement processes with special values.

5.2 Singular components

Certain distributions are neither purely discrete nor absolutely continuous. They can concentrate probability on sets that are highly irregular, producing a CDF that is continuous yet not generated by an ordinary density function. Such distributions are less common in elementary applications but important in advanced probability theory.

5.3 Piecewise forms

Many mixed CDFs are easiest to describe piece by piece. One formula may apply on a continuous interval, while another handles isolated jumps. Piecewise representation makes the structure of the distribution explicit and simplifies computation.

6 Quantiles and inverse functions

Quantiles identify values corresponding to specified cumulative probabilities. They are widely used to summarize position, compare distributions, and invert cumulative information.

6.1 Median and percentiles

The median is a value at which the CDF reaches 0.5, so half the probability lies at or below it. Percentiles generalize this idea: the 90th percentile, for example, is the value below which 90% of the distribution falls. These measures are robust summaries of distributional location.

6.2 Inverse cumulative distribution function

An inverse CDF maps a probability level back to a corresponding value of the variable. Because a CDF may have flat segments or jumps, the inverse is often defined carefully rather than as a literal algebraic inverse. It is commonly used in probability calculations and simulation.

6.3 Quantile function

The quantile function is another name for the inverse-type mapping from cumulative probability to value. It assigns each probability level p between 0 and 1 the smallest x for which F(x) is at least p. This convention works for both discrete and continuous distributions.

6.4 Percentile rank interpretation

Quantiles can be interpreted as percentile ranks. A value’s percentile rank tells how far into the distribution it lies. This interpretation is common in testing, ranking, and descriptive statistics, where results are often reported relative to a reference population.

7 Applications

Cumulative distributions are used throughout probability and statistics because they provide a direct way to express uncertainty and summarize observed data.

7.1 Probability calculations

CDFs simplify the computation of probabilities over intervals and threshold events. Instead of summing or integrating from scratch each time, one can evaluate the function at two points and subtract. This makes them practical for both hand calculations and software implementations.

7.2 Descriptive statistics

In descriptive analysis, a cumulative distribution shows the overall shape of a dataset or model. It helps identify medians, quartiles, tail behavior, and clustering. The CDF is often more informative than a single summary number because it captures the full distribution.

7.3 Simulation and random sampling

The inverse CDF method is a standard technique for generating random samples from a target distribution. A uniform random number is transformed through the quantile function to produce a sample with the desired distribution. This approach is widely used in computer simulation.

7.4 Statistical inference

Cumulative distributions appear in hypothesis testing, confidence intervals, and goodness-of-fit methods. They are used to compute p-values, critical values, and sampling probabilities. In many statistical procedures, the relevant distribution is expressed most clearly through its CDF.

8 Common examples

Several familiar distributions illustrate how cumulative distribution functions behave in practice. These examples show both discrete and continuous forms.

8.1 Bernoulli distribution

A Bernoulli distribution has two possible outcomes, usually 0 and 1. Its CDF has one jump at 0 and another at 1, reflecting the probability assigned to each outcome. This is the simplest discrete example of cumulative probability.

8.2 Binomial distribution

The binomial distribution counts the number of successes in a fixed number of trials. Its CDF is a step function that rises as the count increases from 0 up to the maximum number of trials. It is often used to model repeated yes-or-no events.

8.3 Normal distribution

The normal distribution has a smooth, symmetric CDF that does not admit a simple elementary formula. Its values are usually obtained numerically or from tables. The normal CDF is central in statistics because of the distribution’s many applications and its role in approximation theory.

8.4 Exponential distribution

The exponential distribution models waiting times between events. Its CDF rises quickly near zero and approaches 1 asymptotically. This makes it a standard example of a continuous distribution with a simple, closed-form cumulative function.

Several concepts are closely connected to the cumulative distribution function and are often used alongside it.

9.1 Survival function

The survival function gives the probability that a random variable exceeds a given value. It is the complement of the CDF in many continuous settings, since the two probabilities sum to 1. Survival functions are common in reliability and time-to-event analysis.

9.2 Complementary cumulative distribution

The complementary cumulative distribution function, or CCDF, records the probability of being above a threshold. It is especially useful for tail analysis, where large values are of interest. In practice, it provides a mirror image of the CDF.

9.3 Hazard function

The hazard function describes the instantaneous rate of occurrence of an event, conditional on survival up to a given time. It is derived from both the density and the survival function. Hazard models are important in reliability studies and event-time analysis.

9.4 Empirical cumulative distribution function

The empirical cumulative distribution function is built from observed data rather than a theoretical model. It assigns to each x the proportion of sample values at or below x. This makes it a fundamental tool for nonparametric statistics and data exploration.

</INTERNAL_LINK_CANDIDATES> Cumulative distribution function (function giving probability that a random variable is at or below a value) Random variable (numerical quantity representing an uncertain outcome) Probability mass function (function giving probabilities of discrete outcomes) Probability density function (function describing relative concentration for continuous outcomes) Survival function (probability that a variable exceeds a given value) Complementary cumulative distribution function (probability of being above a threshold) Hazard function (conditional rate of event occurrence over time) Empirical cumulative distribution function (sample-based cumulative distribution from observed data) Quantile function (mapping from cumulative probability to value) Inverse cumulative distribution function (inverse-type mapping from probability level to value) Median (value splitting probability into two equal halves) Percentile (value corresponding to a specified cumulative percentage) Bernoulli distribution (two-outcome discrete distribution) Binomial distribution (distribution of success counts in repeated trials) Normal distribution (symmetric continuous distribution with bell-shaped density) Exponential distribution (continuous distribution for waiting times) Discrete random variable (random variable taking separate values) Continuous random variable (random variable varying over intervals) Probability interval (range of values whose probability is found by subtracting CDF values) Simulation (computer generation of random samples from a distribution)