Understanding the two requirements for a discrete probability distribution is fundamental for anyone studying statistics, data science, or probability theory. A discrete probability distribution describes the likelihood of each possible outcome for a discrete random variable, but not every table of numbers qualifies as a valid distribution. To be mathematically sound, it must satisfy two non-negotiable conditions: every individual probability must fall between 0 and 1, and the sum of all probabilities across the entire sample space must equal exactly 1. This article explores these requirements in depth, explains why they exist, and shows you how to apply them with confidence.
What Is a Discrete Probability Distribution?
Before examining the requirements, it helps to define what we are working with. A discrete random variable is a variable that can take on a countable number of distinct values. Examples include the number of heads in three coin tosses, the number of customers arriving at a store in an hour, or the result of rolling a six-sided die Worth knowing..
A discrete probability distribution assigns a probability to each possible value of that random variable. Think about it: this assignment is often called a probability mass function, or PMF. Unlike a continuous distribution, which deals with intervals and density curves, a discrete distribution works with separate, individual outcomes And that's really what it comes down to..
For a distribution to be legitimate, it cannot simply list random numbers. It must obey the laws of probability, which brings us to the two core requirements.
The Two Essential Requirements for a Discrete Probability Distribution
Every valid discrete probability distribution must meet the following two conditions. If either condition fails, the distribution is not valid and cannot be used for reliable statistical inference.
Requirement 1: Each Probability Must Be Between 0 and 1
For every possible outcome x in the sample space, the probability assigned to that outcome must satisfy:
0 ≤ P(X = x) ≤ 1
In plain language, no probability can be negative, and no probability can exceed 1. A probability of 0 means the event is impossible, while a probability of 1 means the event is certain. Any value outside this range breaks the basic definition of probability.
As an example, if you are modeling the roll of a fair die, the probability of rolling a 3 is 1/6, which sits comfortably between 0 and 1. In real terms, if a table listed the probability of rolling a 3 as -0. 2 or 1.4, it would immediately violate this first requirement Most people skip this — try not to..
Easier said than done, but still worth knowing.
Requirement 2: The Sum of All Probabilities Must Equal 1
The second requirement states that if you add up the probabilities of every possible outcome in the sample space, the total must be exactly 1:
Σ P(X = x) = 1
This condition reflects the idea that something must happen. The sample space is exhaustive, meaning it includes every possible outcome the random variable can produce. Because of this, the total certainty across all outcomes must equal 100 percent, or 1 in probability notation.
Continuing with the die example, the probabilities for rolling 1, 2, 3, 4, 5, and 6 are each 1/6. When summed:
1/6 + 1/6 + 1/6 + 1/6 + 1/6 + 1/6 = 6/6 = 1
The requirement is satisfied.
The Scientific Explanation Behind the Requirements
These two rules are not arbitrary. They are rooted in the axioms of probability established by Andrey Kolmogorov,
Andrey Kolmogorov, whose 1933 monograph Foundations of the Theory of Probability formalized probability as a branch of measure theory. Kolmogorov’s three axioms provide the rigorous mathematical bedrock for the two requirements outlined above.
The First Axiom (Non-negativity) states that for any event $A$, $P(A) \geq 0$. In real terms, this maps directly to our first requirement: individual probabilities cannot be negative. Practically speaking, the Second Axiom (Normalization) states that the probability of the entire sample space $\Omega$ is 1, or $P(\Omega) = 1$. This is the theoretical parent of our second requirement; since the distinct values of a discrete random variable partition the sample space, the sum of their probabilities must equal the probability of the sample space itself. The Third Axiom (Additivity)—which dictates that the probability of a union of mutually exclusive events equals the sum of their individual probabilities—guarantees that we can simply sum the point probabilities to reach that total of 1.
In essence, the two requirements for a discrete distribution are not merely practical checks; they are the direct, applied translation of Kolmogorov’s axioms into the language of probability mass functions.
Validating a Distribution: A Step-by-Step Check
When presented with a table or formula claiming to represent a discrete probability distribution, apply a systematic verification process.
Step 1: Identify the Sample Space List every distinct value $x$ the random variable $X$ can take. Ensure no values are missing and no impossible values are included.
Step 2: Check Individual Bounds (Requirement 1) Scan the probability column $P(X=x)$. If any entry is ${content}lt; 0$ or ${content}gt; 1$, stop. The distribution is invalid.
Step 3: Check the Sum (Requirement 2) Calculate $\sum P(X=x)$. Due to floating-point arithmetic in software or rounding in textbooks, the sum may be $0.999$ or $1.001$. Generally, a sum within a small tolerance (e.g., $\pm 0.001$) of 1 is accepted, but theoretically, it must be exactly 1. If the sum differs significantly, the distribution is invalid.
Illustrative Examples
Example A: A Valid Distribution (Number of Defects) A quality control engineer records the number of defects $X$ per batch of 100 widgets Worth keeping that in mind. That alone is useful..
| $x$ | 0 | 1 | 2 | 3 |
|---|---|---|---|---|
| $P(X=x)$ | 0.80 | 0.15 | 0.04 | 0.01 |
- Check 1: All probabilities are between 0 and 1. Pass.
- Check 2: $0.80 + 0.15 + 0.04 + 0.01 = 1.00$. Pass.
- Verdict: Valid distribution.
Example B: Invalid – Negative Probability A model for customer complaints $Y$ per day.
| $y$ | 0 | 1 | 2 |
|---|---|---|---|
| $P(Y=y)$ | 0.90 | 0.15 | -0.05 |
- Check 1: $P(Y=2) = -0.05$. Fail. Probabilities cannot be negative.
Example C: Invalid – Sum $\neq$ 1 A proposed distribution for a loaded die roll $Z$.
| $z$ | 1 | 2 | 3 | 4 | 5 | 6 |
|---|---|---|---|---|---|---|
| $P(Z=z)$ | 0.1 | 0.1 | 0.1 | 0.1 | 0.1 | 0.1 |
- Check 1: All values in $[0, 1]$. Pass.
- Check 2: Sum $= 0.6$. Fail. The sample space is not exhausted; probability mass is missing.
Common Families of Discrete Distributions
While custom tables are useful for specific empirical data, statisticians rely on parametric families—mathematical formulas defined by a few parameters—to model broad classes of phenomena. Knowing the requirements allows you to verify that these formulas produce legitimate PMFs for all valid parameter values.
- Bernoulli Distribution: Models a single trial with two outcomes (Success/Failure). Parameter: $p$ (probability of success). PMF: $P(X=1)=p, P(X=0)=1-p$.
- Binomial Distribution: Models the number of successes in $n$ independent Bernoulli trials. Parameters: $n, p$.
- Poisson Distribution: Models
