Counting: Permutations & Combinations
How to count arrangements and selections without ever writing them all out — and how to tell which one a problem is asking for.
Counting without listing everything
Suppose a lunch deal lets you pick one of 3 sandwiches, one of 4 sides, and one of 2 drinks. How many different lunches are possible? You could try to list them all — but there is a faster way. You have 3 choices, then for each of those 4 more, then for each of those 2 more, so the choices multiply: \(3\times4\times2=24\).
That simple habit — multiply the number of options at each step — is the engine behind almost every counting problem in probability. Everything else in this lesson is just careful bookkeeping on top of it.
Factorials: arranging a whole set
Now ask a sharper question: in how many orders can you line up all the items in a set? For 3 books on a shelf, there are 3 choices for the first slot, 2 left for the second, and 1 for the last: \(3\times2\times1=6\). This running product of an integer down to 1 is so common it gets its own symbol, the factorial, written with an exclamation mark.
So \(3!=6\), \(4!=24\), and factorials grow fast — \(10!\) is already over three million. By convention \(0!=1\) (there is exactly one way to arrange nothing).
Order matters — or it doesn't
Usually you do not arrange the whole set; you pick \(r\) items out of \(n\). Here everything hinges on a single question: does the order of the picks matter?
- If order matters — like awarding 1st, 2nd, and 3rd place — each ordering is a different outcome. These are permutations, counted by \({}^{n}P_{r}\).
- If order is ignored — like choosing a committee, where the same people are the same group no matter who was named first — each unordered group counts once. These are combinations, counted by \({}^{n}C_{r}\).
A permutation cares about arrangement; a combination cares only about which items got picked.
Which one should I use?
When a problem asks 'how many ways…', pause and ask whether swapping the order of the chosen items gives a genuinely different result:
- Different result → permutations. Rankings, finishing positions, passwords and PINs, seating people in a row, assigning distinct job titles.
- Same result → combinations. Choosing a committee, dealing a hand of cards, picking pizza toppings, selecting a subset with no roles or ranks.
A quick tell: if the picked items get labels, seats, or ranks, order matters. If they just go into an unlabeled group, it does not.
- Part (a) — the three roles are distinct, so order matters: this is a permutation. \({}^{10}P_{3} = 10\times9\times8 = 720\).
- Part (b) — a committee has no roles, so the same three people are the same committee in any order: this is a combination. \({}^{10}C_{3} = \frac{720}{3!} = \frac{720}{6} = 120\).
- Sanity check with the bridge: \({}^{n}P_{r} = {}^{n}C_{r}\times r!\) gives \(720 = 120\times 6\). Each committee of 3 can be arranged into \(3!=6\) different slates of officers.
Check your understanding
- The multiplication principle: multiply the number of options at each step to count the outcomes of a sequence of choices.
- A factorial n! counts the ways to arrange all n distinct items; factorials grow very fast and 0! = 1 by convention.
- Permutations (nPr = n!/(n-r)!) count ordered selections — use them when order matters, like rankings or seats.
- Combinations (nCr = n!/(r!(n-r)!)) count unordered selections — use them when order is ignored, like committees or hands of cards.
- The two are linked by nPr = nCr times r!: each selection of r items can be arranged in r! orders, so permutations overcount combinations by exactly r!.