Exact Binomial Probabilities in Context

What Does “Exactly” Mean?

A probability question may ask how likely it is that a process produces exactly a specified number of successes. For example, a quality engineer might ask how likely it is that exactly two of eight independently tested units fail a test. The requested event is not “at most two failures” or “two or fewer”; it includes only outcomes with two failures.

The binomial model describes a fixed number of trials, each with two possible outcomes, when the probability of success is the same on every trial and the trials are independent. Once those conditions hold, the probability of exactly $x$ successes can be calculated by combining the number of possible arrangements with the probability of each arrangement.

Definition: If $X$ is the number of successes in $n$ independent trials, each with success probability $p$, then $X$ has a binomial distribution. For an integer $x$ from 0 through $n$, the probability of exactly $x$ successes is $$ P(X=x)=\binom{n}{x}p^x(1-p)^{n-x}. $$ Here, $\binom{n}{x}$ counts the arrangements of $x$ successes among the $n$ trials, $p^x$ is the success-probability factor, and $(1-p)^{n-x}$ is the failure-probability factor.

Check the Model Before Calculating

The formula applies only when the situation fits the binomial structure. Before using it, state what counts as a success and check the conditions:

Fixed number of trials: The process has a specified number $n$ of trials.
Two outcomes per trial: Each trial is classified as success or failure for the question being asked.
Independence: The outcome of one trial does not change the probabilities for the other trials.
Constant success probability: Each trial has the same probability $p$ of success.

The word “success” is a label for the outcome being counted; it need not mean a desirable outcome. If a question is about failures, it is often convenient to define a failure as the success of the binomial count. The choice must be explicit, because $p$ refers to the probability of the outcome designated as success.

In a problem that explicitly states independent trials with a common probability, those statements establish the relevant assumptions. In a real setting, assumptions need justification. For example, repeated tests on separate units might reasonably be modeled as independent with a common failure probability if each unit is produced and tested under comparable conditions. If probabilities change from trial to trial or outcomes affect one another, the binomial formula is not justified without further modeling.

Why the Formula Has Three Factors

First consider one particular ordered sequence with $x$ successes and $n-x$ failures. Independence means that the probability of the whole sequence is the product of its trial probabilities. Since each success has probability $p$ and each failure has probability $1-p$, that sequence has probability $p^x(1-p)^{n-x}$.

But “exactly $x$ successes” does not specify their positions. The successes could occur in many different arrangements. The number of ways to choose which $x$ of the $n$ trial positions are successes is $\binom{n}{x}$. These arrangements are mutually exclusive: a particular trial sequence cannot have two different sets of success positions. Each arrangement has the same probability, so adding the probabilities of all arrangements gives the exact binomial formula.

\underbrace{\binom{n}{x}}_{\text{number of arrangements}} \;\times\; \underbrace{p^x}_{\text{success factor}} \;\times\; \underbrace{(1-p)^{n-x}}_{\text{failure factor}} = P(X=x)

This reasoning also explains why the exponent on the failure factor is $n-x$: if $x$ of the $n$ trials are successes, the remaining $n-x$ must be failures. The binomial coefficient counts arrangements; it is not itself a probability. The product of all three factors is the probability of the exact count.

Calculating an Exact Binomial Probability

A reliable calculation begins by translating the wording into the parameters. Write down $n$, $p$, and $x$, and check that $x$ is an integer between 0 and $n$. Then calculate the number of arrangements, the success factor, and the failure factor. Keeping the factors visible makes it easier to detect an incorrect exponent or an accidental use of the probability of failure as $p$.

Worked Example: Exactly Two Test Failures

Hypothetical situation. Suppose eight separate sensor units are tested. Assume each unit independently fails its test with probability 0.30, and the probability is the same for every unit. Find the probability that exactly two units fail.

Let a failure be a success for the count $X$. There are a fixed eight trials, two outcomes for each trial, independent outcomes, and a common failure probability $p=0.30$. Thus the binomial model is appropriate, with $n=8$ and $x=2$.

There are $\binom{8}{2}=28$ ways to choose which two units fail. The success factor is $0.30^2$, and the failure factor is $0.70^{8-2}=0.70^6$. Therefore,

\begin{aligned} P(X=2) &=\binom{8}{2}(0.30)^2(0.70)^6\\ &=28(0.09)(0.117649)\\ &=28(0.01058841)\\ &=0.29647548 \end{aligned}

The probability that exactly two of the eight units fail is about $0.2965$, or $29.65\%$. This is the probability of that exact count under the stated model; it does not include outcomes with zero, one, three, or any other number of failures.

Worked Example: Exactly Zero Delayed Deliveries

Hypothetical situation. Suppose five deliveries are tracked, and each is independently delayed with probability 0.08. Assume this probability is the same for every delivery. Find the probability that none is delayed.

Define success as “a delivery is delayed.” The fixed trial count is $n=5$, the probability of success is $p=0.08$, and “none is delayed” means $x=0$ successes. Each delivery has two outcomes for this classification; the stated independence and common probability complete the binomial conditions.

There is just one arrangement with zero delayed deliveries: all five are not delayed. The combination factor is $\binom{5}{0}=1$, the success factor is $0.08^0=1$, and the failure factor is $0.92^5$.

\begin{aligned} P(X=0) &=\binom{5}{0}(0.08)^0(0.92)^5\\ &=1(1)(0.6590815232)\\ &=0.6590815232 \end{aligned}

The probability that none of the five deliveries is delayed is approximately $0.6591$, or $65.91\%$. The zero in $x=0$ removes the success factor through $p^0=1$; it does not mean there are zero trials. All five trials contribute to the failure factor.

Worked Example: Exactly Seven Responses

Hypothetical situation. Suppose ten people are contacted separately, and each independently responds with probability 0.60. Assume the response probability is identical for all ten contacts. Find the probability that exactly seven people respond.

Define a response as success. There are ten fixed trials, each has response or no response as its two outcomes, and the stated assumptions give independence and a common $p=0.60$. Thus $n=10$ and $x=7$. The number of arrangements is $\binom{10}{7}=120$; the three remaining trials are nonresponses.

\begin{aligned} P(X=7) &=\binom{10}{7}(0.60)^7(0.40)^{10-7}\\ &=120(0.0279936)(0.064)\\ &=120(0.0017915904)\\ &=0.214990848 \end{aligned}

The probability that exactly seven of the ten people respond is approximately $0.2150$, or $21.50\%$, under this hypothetical model. The exponent 3 on $0.40$ accounts for the three nonresponses, not for the seven responses.

Interpreting and Reporting the Result

A complete answer has more than a decimal. It names the event, gives the probability, and connects the result to the context. For example, “Under the stated independent-trial model, the probability that exactly two of the eight units fail is about 0.2965.” A percentage may be added, but it should represent the same value: $0.2965$ is $29.65\%$, not $0.2965\%$.

The probability refers to the specified model and its assumptions. It describes how often the exact event would occur over repeated runs of the same process, in the long run. It is not a claim that the event will occur a particular number of times in one batch, nor is it a probability about whether the assumptions are true.

“Exactly” is essential. The value $P(X=x)$ concerns one value of the random variable, not a range of values. If a question asks for several possible counts, this exact-count calculation alone is not the requested final probability; each included count and the event wording must be handled explicitly.

Common Mistakes

Using $x$ for failures by accident. If success was defined as a response and exactly seven people respond out of ten, there are three failures. Use $p^7(1-p)^3$, not $p^3(1-p)^7$.
Forgetting arrangements. The probability of one particular sequence is only $p^x(1-p)^{n-x}$. For an exact count, multiply by $\binom{n}{x}$, because the successes can occupy different trial positions.
Treating $\binom{n}{x}$ as the answer. It counts arrangements and can exceed 1. It must be multiplied by the success and failure factors to obtain a probability.
Changing the event while interpreting the answer. “Exactly two” is not “at most two.” State the requested count precisely and do not describe a single-count probability as covering additional outcomes.
Using the model without checking its assumptions. A fixed trial count alone is not enough. The trials must also have two classified outcomes, be independent, and share the same success probability.

Key takeaway: For exactly $x$ successes in $n$ binomial trials, multiply the number of arrangements, $\binom{n}{x}$, by the probability of the successes, $p^x$, and the probability of the remaining failures, $(1-p)^{n-x}$. Then report what that exact probability means in the stated context.

Check Your Understanding

For each question, identify the event carefully before calculating or explaining.

A binomial model has $n=6$ and $p=0.25$. Write the formula for the probability of exactly two successes, including all three factors.
In a hypothetical process, four trials have success probability $0.10$, independently. What are the combination factor and the success and failure exponents for exactly zero successes?
Why must the probability of one particular sequence be multiplied by a combination count to obtain the probability of exactly $x$ successes?
A student calculates the probability of exactly three successes in eight trials using $p^3(1-p)^3$. Identify the error and write the correct failure exponent.
Explain in context what it means if a correctly calculated exact binomial probability is $0.12$. What event does that number describe?