Counting Success Arrangements With Binomial Coefficients

Why Count Arrangements?

In a sequence of trials, the same total number of successes can arise in different orders. For example, two successes in five trials might occur on trials 1 and 2, or on trials 1 and 3. If we want to reason about an outcome described only by its number of successes, we need to know how many distinct trial sequences produce that count.

The binomial coefficient answers this counting question. It counts the ways to choose which trial positions are successes, without treating different orders among those chosen positions as new choices. To connect this count to probability, we also need to know when the different sequences have equal probabilities. That is where the binomial model’s assumptions matter.

Definition: For integers \(n \geq 0\) and \(x\) with \(0 \leq x \leq n\), the binomial coefficient \({n \choose x}\), read “\(n\) choose \(x\),” is the number of ways to select \(x\) distinct positions from \(n\) labeled positions when the order of selection does not matter. Its formula is

{n \choose x}=\frac{n!}{x!(n-x)!}, \qquad 0!=1.

Here \(n!\), read “\(n\) factorial,” is the product of the positive integers from 1 through \(n\); for instance, \(5!=5\cdot4\cdot3\cdot2\cdot1=120\). The positions are labeled because the trials occur in a particular order: trial 1 is distinct from trial 2, even if the trials use the same procedure.

Deriving the Combination Count

Suppose we want to choose the \(x\) positions that will be successes among \(n\) trials. First count selections in which order matters: there are \(n\) choices for the first success position, \(n-1\) for the second, and so on, until \(n-x+1\) choices for the \(x\)th. The number of ordered selections is therefore \(n(n-1)\cdots(n-x+1)\).

But choosing position 2 and then position 5 identifies the same set of success positions as choosing position 5 and then position 2. Every set of \(x\) positions is counted \(x!\) times in the ordered count, once for each ordering of its members. Dividing by \(x!\) removes this overcount. The product can be rewritten using factorials:

{n \choose x} =\frac{n(n-1)\cdots(n-x+1)}{x!} =\frac{n!}{(n-x)!x!}.

This argument also covers the boundary cases. If \(x=0\), there is exactly one way to select no success positions: select none. The formula gives \({n \choose 0}=n!/(0!n!)=1\). If \(x=n\), there is exactly one way to select every position, and \({n \choose n}=1\). The factorial convention \(0!=1\) makes both results fit the formula.

Key interpretation: In a sequence of \(n\) trials, each particular choice of \(x\) success positions determines one success-failure sequence. Thus \({n \choose x}\) counts the distinct sequences with exactly \(x\) successes.

Worked Example: Two Successes in Five Trials

Suppose a hypothetical fair coin is tossed five times. Define success as observing heads. How many different sequences have exactly two successes?

Check the setting. There are five fixed trials. Each toss has two outcomes, the tosses are independent, and the probability of success is the same, \(p=0.5\), on every toss. These are the binomial-model conditions.

Choose which two of the five positions contain heads. The count is

{5 \choose 2} =\frac{5!}{2!3!} =\frac{120}{2\cdot6} =10.

For example, choosing positions 1 and 3 gives the sequence H-T-H-T-T. Choosing positions 1 and 4 gives H-T-T-H-T, a different sequence. There are 10 such position choices, so there are 10 sequences with exactly two heads.

Each particular sequence has two heads and three tails. By independence, its probability is

(0.5)(0.5)(0.5)(0.5)(0.5)=(0.5)^2(0.5)^3=\frac{1}{32}.

The tosses are fair, so every position contributes a factor of 0.5 regardless of whether it shows heads or tails. Consequently, all 10 sequences have the same probability, \(1/32\). The count tells us how many sequences there are; the equal-probability claim comes from the model assumptions.

Why the Arrangements Have Equal Probability

A binomial model describes a fixed number \(n\) of independent trials, each with the same probability \(p\) of success and probability \(1-p\) of failure. An arrangement is one specific ordered sequence of outcomes. For instance, S-F-S-F specifies success on trials 1 and 3 and failure on trials 2 and 4.

Binomial-model assumptions used here: The number of trials \(n\) is fixed; each trial has two possible outcomes classified as success or failure; trials are independent; and the success probability \(p\) is the same on every trial. Under these assumptions, any particular sequence with \(x\) successes and \(n-x\) failures has probability \(p^x(1-p)^{n-x}\).

The sequence-probability result follows directly from independence. Independence means that the probability of a specified combination of trial outcomes is the product of the individual probabilities. A sequence with \(x\) successes contributes \(p\) once for each success, and a sequence with \(n-x\) failures contributes \(1-p\) once for each failure. Multiplying those factors gives the stated expression. Since multiplication does not depend on the order of its factors, moving the success positions does not change the product.

\underbrace{p\cdot p\cdots p}_{x\text{ success factors}} \underbrace{(1-p)(1-p)\cdots(1-p)}_{n-x\text{ failure factors}} =p^x(1-p)^{n-x}.

The common probability depends on both assumptions: independence lets us multiply trial probabilities, and a common \(p\) makes each success factor the same across trial positions. When those conditions hold, the sequences with \(x\) successes are disjoint outcomes, each with the same probability. The combination count identifies how many such outcomes there are.

Worked Example: Counting Three Successes in Eight Trials

In a hypothetical quality-control exercise, eight items are checked in sequence. Suppose each check independently identifies an item requiring rework with probability \(p=0.2\), the same for every item. Let that finding be called a success for this example. How many sequences have exactly three successes, and what is the probability of one particular sequence?

Check the setting. There are eight fixed checks, each has two classifications for this question (success or failure), the checks are assumed independent, and each has the same success probability, 0.2. Thus the binomial assumptions apply.

The three success positions can be chosen in

{8 \choose 3} =\frac{8!}{3!5!} =\frac{8\cdot7\cdot6}{3\cdot2\cdot1} =\frac{336}{6} =56

ways. For example, the choice of positions 1, 4, and 8 gives one sequence; choosing positions 2, 4, and 8 gives another. These are distinct because the item positions differ.

For any one specified sequence with three successes and five failures, independence gives

(0.2)^3(0.8)^5 =0.008\cdot0.32768 =0.00262144.

The specific positions do not appear in this calculation. Every sequence with three successes therefore has probability 0.00262144 under the stated model. The number 56 is a count of possible sequences, not the probability of any one sequence.

When Equal Probability Fails

The combination count remains a valid count of position choices even if trial probabilities differ. But counting alone does not imply that the choices have equal probabilities. If different trials have different success probabilities, sequences with the same number of successes can have different probabilities. If outcomes are dependent, the probability of a sequence may also depend on the order or on earlier outcomes.

Worked Example: Different Trial Probabilities

Consider four hypothetical independent trials with success probabilities 0.2, 0.3, 0.4, and 0.5, respectively. How many sequences have exactly two successes, and do all of them have the same probability?

The number of ways to choose two success positions is still

{4 \choose 2} =\frac{4!}{2!2!} =6.

Now compare two of those arrangements. Let S denote success and F denote failure. The sequence S-S-F-F has probability

$$ (0.2)(0.3)(1-0.4)(1-0.5) =(0.2)(0.3)(0.6)(0.5) =0.018. $$

The sequence S-F-S-F has probability

$$ (0.2)(1-0.3)(0.4)(1-0.5) =(0.2)(0.7)(0.4)(0.5) =0.028. $$

Both sequences have two successes, but their probabilities differ. The trials are independent, yet the success probability is not constant across trials, so the binomial-model assumptions do not all hold. The value 6 still correctly counts the arrangements, but it cannot be used as a count of equally likely outcomes here.

Common Mistakes and Why This Count Matters

Counting different orders of chosen positions separately. Choosing positions 1 and 4 is the same selection whether position 1 or position 4 is named first. The factor \(x!\) in the denominator removes those repeated orderings.
Assuming a correct count proves equal probabilities. A binomial coefficient counts arrangements; it does not establish that the trials are independent or have the same success probability. Check the model before using equal-probability reasoning.
Confusing an arrangement with the total success count. “Exactly two successes” describes a collection of sequences. A particular arrangement, such as S-F-S-F-F, is just one member of that collection.
Forgetting that trial positions are distinct. In repeated trials, success on trial 1 and success on trial 2 are different positions, even when the trials follow the same procedure.

This counting idea matters because an event such as “exactly \(x\) successes” groups together many different ordered outcomes. Under the binomial assumptions, each of those outcomes has the same probability, so the number of arrangements is the relevant multiplicity when their probabilities are combined. For now, keep the two pieces of reasoning separate: \({n \choose x}\) answers “how many arrangements?”; the trial model answers “what is the probability of each particular arrangement?”

Key takeaway: Choosing \(x\) success positions among \(n\) trials gives \({n \choose x}=n!/[x!(n-x)!]\) arrangements. In a binomial model, independence and a common success probability make every particular arrangement with \(x\) successes have the same probability.

Check Your Understanding

Try these questions without looking back at the worked examples.

How many sequences of six trials have exactly four successes? Show how the factorial formula gives your answer.
In your own words, why does dividing by \(x!\) remove repeated counts when choosing \(x\) success positions?
For a sequence of seven independent trials with common success probability \(p\), write the probability of one specified sequence containing two successes and five failures.
Does independence alone guarantee that all arrangements with a fixed number of successes have equal probability? Explain.
How many arrangements have zero successes among \(n\) trials, and what does that arrangement represent?