From Counting to Probability
Many probability questions describe a finite set of possible outcomes and a random process that gives each outcome the same chance. Once the outcomes are counted, the calculation may be simple: count the outcomes that satisfy the event, then compare that count with the total. The important work is making sure the outcomes being counted are the right ones.
For example, a committee is an unordered group: its members do not have positions or an arrival order. A line-up, by contrast, is ordered: changing who stands first changes the outcome. Counting committees as though they were line-ups would count the wrong sample space, even if the arithmetic were flawless.
The numerator is the favorable count: it counts outcomes in the event, not necessarily outcomes that seem desirable. The denominator is the total count: it counts every possible outcome in the sample space defined by the random process. Both counts must use the same kind of outcome.
Why the Ratio Works
Suppose a sample space has \(N\) equally likely outcomes. Since their probabilities are equal and their probabilities sum to 1, each outcome has probability \(1/N\). If event \(A\) contains \(N(A)\) outcomes, those outcomes are disjoint individual possibilities, so their probabilities add. Thus the event probability is the number of its outcomes multiplied by the probability of each one.
This justification depends on equal likelihood. A list of possible outcomes alone does not make the ratio valid: if some listed outcomes are more likely than others, simply counting them gives each one the same weight when the random process does not.
Choosing the Outcome to Count
Before counting, describe one outcome in plain language. Ask what the random process regards as a distinct result. If a process selects a group, an outcome is a group. If it assigns people to ordered positions, an outcome is an arrangement. The counting method then follows from the structure of that outcome.
| What counts as one outcome? | Typical counting method | Example of a distinction |
|---|---|---|
| An unordered selection | Combination | Choosing A, B, and C is the same group as choosing C, B, and A |
| An ordered arrangement | Permutation | A, B, C in that order differs from C, B, A |
| One of several distinct types of favorable outcomes | Count disjoint cases and add | An even sum comes from two even numbers or two odd numbers |
The permutation and combination formulas count different things. For \(n\) distinct objects arranged in \(r\) ordered positions, the number of arrangements is \(n!/(n-r)!\). For an unordered selection of \(r\) objects from \(n\) distinct objects, the number of groups is \(\binom{n}{r}=n!/[r!(n-r)!]\). These formulas are useful for probability only when their counted outcomes match the sample space.
Worked Examples
Worked Example: A Randomly Selected Committee
A group has 4 trained volunteers and 5 other volunteers. One committee of 3 people is selected uniformly from all possible 3-person committees. What is the probability that the committee contains exactly 2 trained volunteers?
Model and assumptions. The sample space consists of all unordered groups of 3 people. Assume each such committee is equally likely to be selected. Because order within a committee does not matter, combinations count both the total outcomes and the favorable outcomes.
There are 9 people altogether, so the total number of committees is
For a favorable committee, choose 2 of the 4 trained volunteers and 1 of the 5 other volunteers. Each such choice gives one committee, and the trained and other volunteers are distinct groups, so the count is
Therefore,
The probability that a uniformly selected 3-person committee contains exactly 2 trained volunteers is \(5/14\), or about 35.71%.
Worked Example: Two People Adjacent in a Line-Up
Five distinct volunteers are placed in a uniformly random line-up. What is the probability that volunteers Mira and Noah stand next to each other?
Model and assumptions. Each outcome is a complete ordered line-up of the 5 volunteers, and assume all such line-ups are equally likely. The total count must therefore use permutations, not combinations.
There are \(5!\) possible line-ups:
To count line-ups in which Mira and Noah are adjacent, treat them temporarily as one block. The block and the other 3 volunteers make 4 objects, which can be arranged in \(4!\) ways. Within the block, Mira can stand before Noah or Noah can stand before Mira, giving \(2\) internal orders. Thus
The favorable count includes both possible orders inside the adjacent pair and every possible position for the block. The probability is
Under the stated uniform-line-up model, the probability that Mira and Noah are adjacent is \(2/5\), or 40.00%.
Worked Example: Counting Favorable Outcomes in Disjoint Cases
Eight distinct numbered tokens are labelled 1 through 8. A pair is selected uniformly from all unordered pairs of tokens. What is the probability that the sum of the two labels is even?
Model and assumptions. Each outcome is an unordered pair, and every pair is assumed equally likely. There are 8 tokens and 2 are in each pair, so the total count is
An even sum occurs in exactly two cases: both labels are even, or both labels are odd. These cases cannot overlap, since a pair cannot consist of both two even labels and two odd labels. There are 4 even labels and 4 odd labels, so the favorable count is
The probability is therefore
The probability that a uniformly selected unordered pair has an even sum is \(3/7\), or about 42.86%. Adding the counts works because the two cases are disjoint; no favorable pair is counted twice.
A Reliable Counting Process
A careful solution keeps the model, counting method, and probability calculation connected. Use the following sequence whenever a problem asks for a probability from finite counts.
State exactly what distinguishes one result from another, such as a committee or a full ordered line-up.
Confirm that each outcome in this sample space has the same probability under the stated random process. If this is not so, the simple count ratio is not justified.
Use combinations for unordered groups, permutations for ordered arrangements, or another count that matches the outcome definition.
Count outcomes satisfying the event. If there are disjoint cases, count each case and add their counts.
Divide the favorable count by the total count, simplify if useful, and describe the probability in the context of the random process.
Why the Counting Unit Matters
Consider the committee example. Each committee contains several possible orders in which its members could be listed, but those orders are not different committees. Counting ordered lists in the denominator while counting unordered groups in the numerator would compare unlike units. Even if the resulting fraction happened to fall between 0 and 1, it would not represent the intended probability.
The same principle applies when counting cases. Cases should be chosen so that every favorable outcome is included and no outcome is counted more than once. In the numbered-token example, “both even” and “both odd” are disjoint and together cover every pair with an even sum. If cases overlap, adding their counts directly would count the shared outcomes multiple times; the event must instead be counted without duplication.
A further check is to compare the favorable count with the total count. The favorable count cannot be negative or exceed the total. The resulting probability must be between 0 and 1. These checks can reveal arithmetic or logic errors, although passing them does not prove that the sample space was defined correctly.
Common Mistake
Wrong reasoning: “There are 48 favorable line-ups, and there are \(\binom{5}{3}\) ways to choose people, so the probability is \(48/\binom{5}{3}\).”
This mixes ordered line-ups in the numerator with unordered groups in the denominator. The two counts refer to different sample spaces. Correction: define each outcome as a complete line-up, then count both the total line-ups and the favorable line-ups using that same definition: \(48/5!=48/120=2/5\).
Check Your Understanding
For each question, identify the outcome being counted and explain how the counting supports the probability calculation.
- A uniformly selected 2-person group is chosen from 6 people. What is the total number of possible groups? Why is a combination appropriate?
- Four distinct books are placed in a uniformly random row. What is the total number of possible arrangements?
- A uniformly selected pair is chosen from 5 labelled objects, 2 red and 3 blue. How many pairs contain one red and one blue object? What is the total number of pairs?
- Explain why the equally likely outcome condition is needed for the favorable-count-to-total-count ratio.
- When adding counts for several favorable cases, what must be true of the cases to ensure no outcome is counted twice?