When the order of a selection does not matter, how do you count the choices, and how do the restriction moves you learned for arrangements carry across to unordered selections?
Count unordered selections (combinations) using ^{n}C_{r} = ^{n}P_{r}/r! = n!/(r!(n-r)!), recognise that an n-element set has 2^n subsets and that the r-element subsets number ^{n}C_{r} with the symmetry ^{n}C_{r} = ^{n}C_{n-r} and row sum 2^n, and apply the at-least, complement, conditional and geometric counting techniques to selection problems
Counting unordered selections with combinations. Why nCr = nPr/r! = n!/(r!(n-r)!), why an n-set has 2^n subsets, the symmetry nCr = nC(n-r), and the row sum 2^n. Restriction technique with combinations: at least k by complement, must contain or exclude an item, committees with constraints, and geometry counts.
Reviewed by: AI editorial process; not yet individually human-reviewed
Have a quick question? Jump to the Q&A page
What this dot point is asking
The multiplication principle and ordered selections page counted selections where order matters: filling distinct positions, forming words, building numbers. The count of ordered selections of distinct objects from is the permutation . This page is about the other half of selection counting: when order does not matter. Choosing books to take on holiday, picking a committee of , drawing cards, joining points with a line, these are all unordered selections, and an unordered selection of distinct objects is just a subset of the set you are choosing from.
The central object is the combination , the number of -element subsets of an -element set, spoken " choose " and also written . There are three facts to lock in. First, the formula , where the division by is exactly the step that turns "ordered" into "unordered". Second, the total number of subsets of an -element set is , which gives the row sum . Third, the symmetry , because choosing the you keep is the same as choosing the you leave behind. With those three in hand, the rest of the dot point is restriction technique: "at least", the complement, "must contain or exclude" a particular item, committees with gender or role constraints, and geometric counts such as triangles from points on a circle.
The answer
From ordered to unordered: why you divide by
Start from something you can already count. The number of ordered selections of distinct letters from the five-letter set is
Now ask the unordered question: how many subsets of size are there? The subset is a single object, but it shows up among those ordered words as six different words,
because the three chosen letters can be ordered in ways. The same is true of every subset: each -element subset corresponds to exactly ordered words. So the ordered words count each subset times, and the number of subsets is
The picture is the whole proof in one frame: the six ordered words on the left all funnel into the single subset on the right, so to count subsets you take the ordered count and divide by the orderings that each subset is counted with. Doing this in general, there are ordered selections, each subset is ordered in ways, so the number of unordered selections is
That single division by is the only difference between a permutation and a combination. If you ever cannot remember whether to divide, ask "does reordering my chosen objects make a new outcome?" If yes, it is a permutation; if no, divide by and it is a combination.
Computing by hand, and the contrast
For actual numbers, do not expand both full factorials. Cancel the largest factorial, , against the top straight away, leaving falling factors over . For ,
Notice the shortcut: cancel the larger of and (here ), so only survives on top and on the bottom. That is the same as using the symmetry first, , which is why is easiest to compute when you make the bottom number .
The contrast with the permutation is worth seeing side by side. Choosing objects from in order gives
and dividing by recovers the combination, . The ordered count is times larger because each unordered choice of can be arranged in orders.
Every subset at once: an -set has subsets
Step back from a fixed size and count all subsets of an -element set together. Building a subset means going through the elements one at a time and deciding, independently, whether each is in or out, two choices per element. By the multiplication principle that is
subsets in total, where the all-out choice gives the empty set and the all-in choice gives the whole set. For the five-element set this is subsets.
The display sorts those subsets into columns by size. The column heights are , which read , and they add to . That gives a second way to count all subsets: count them size by size and add. Equating the two counts is the row sum
true because both sides count every subset of an -set exactly once, the left by sorting on size and the right all together. For this is , exactly the columns in the figure.
The symmetry
Look at the figure again and notice the columns read the same left to right as right to left: . That is the symmetry
The reason is a perfect pairing, not a coincidence. Every time you choose an -element subset to keep, you simultaneously leave behind an -element subset, its complement. Choosing the keepers and choosing the leavers are the same act described two ways, so the number of -subsets equals the number of -subsets. Choosing people from to make the tea is the same as choosing the who will not, so . In practice the symmetry is a labour-saver: compute using whichever of and is smaller.
Restriction technique with combinations
The reshaping moves from the grouping, complement and cases page carry straight across to selections, because a combination is itself a count you can subtract from or split. Four moves cover almost every restricted-selection question.
- Conditional: must contain a particular item
- If a chosen object is forced in, put it in first, then choose the remaining places from the remaining objects. A team of from that must include the captain has fixed, so the other come from the remaining : .
- Conditional: must exclude a particular item
- If an object is banned, simply remove it and choose from what is left. A team of from that must exclude an injured player chooses all from the remaining : . Combine the two: include and exclude leaves to choose from , namely .
- Split a constrained selection into independent groups
- When a committee must contain set numbers from disjoint pools, choose from each pool separately and multiply. A committee of from men and women with exactly men chooses of the men and of the women: .
- At least / at most: count the complement
- "At least " usually fragments into several size cases, so count the total and subtract the few unwanted cases. A committee of from women and men with at least women is the total minus the committees with or woman, , giving .
Geometry counts: lines and triangles from points
A clean family of combination questions hides inside geometry. Given points with no three on a straight line, every unordered pair of points determines exactly one line (chord), and every unordered triple determines exactly one triangle, so the counts are simply combinations.
For the points shown, the chords number , the triangles number , and the quadrilaterals number . The highlighted triangle is one of those : notice that choosing the three vertices in any order gives the same triangle, which is exactly why it is a combination and not a permutation. The "no three collinear" condition matters, because three collinear points would give a degenerate "triangle" (a straight line) and the count would overcount; on a circle no three points are ever collinear, so the count is exact.
Verifying a small case by listing
Whenever a count is small, the safest check is to list and tally. Take the triangles from points on a circle. The formula gives . Listing every -element subset in increasing order gives
which is triangles, matching the formula. Listing in a fixed order (here, smallest digit first) guarantees you write each subset once and never repeat, which is the discipline that makes a by-hand check trustworthy.
How exam questions ask about this dot point
The wording tells you which move to reach for. Map the phrase to the method:
- "...select / choose / pick a group / committee / team / hand", "in how many ways can be chosen from ". A plain combination: , because order does not matter.
- "...how many subsets", "how many possible selections of any size", "any number of toppings". All subsets: .
- "...show that ", "why are the answers the same". Symmetry by complement: choosing the kept is the same as choosing the left out.
- "...must include / contains a particular ...". Fix that item, choose the rest from the rest: .
- "...must not include / excludes / without a particular ...". Delete that item, choose normally: .
- "...exactly of one type and the rest another", "two men and two women". Multiply per-pool combinations: .
- "...at least / at most / a majority of ...". Complement: total minus the unwanted size cases.
- "...how many lines / chords / triangles / diagonals from these points", with no three collinear. A geometry combination: for lines, for triangles.
Practice questions
Original practice questions graded from foundation to exam level, each with a full worked solution. Try them before revealing the solution.
foundation3 marks(i) Evaluate . (ii) Evaluate . (iii) Hence state, in one sentence, why selecting people from to form a committee gives a smaller answer than choosing people from to fill four distinct offices (chair, secretary, treasurer, member).Show worked solution →
Part (i): use the formula for a combination. Cancel the larger factorial against the numerator:
Part (ii): use the formula for a permutation. Here order matters, so do not divide by :
Part (iii): explain the gap. Filling four distinct offices is an ordered selection, so it counts . A committee is unordered, so each committee of is counted times among those ordered lists, and dividing removes that overcount: . Order multiplies the answer by .
Answer. (i) ; (ii) ; (iii) the committee is unordered, so it equals , a factor of smaller.
foundation2 marksA pizza shop offers toppings. A customer may choose any number of toppings, from none at all up to all . How many different pizzas (counted by their set of toppings) are possible?Show worked solution →
Recognise this as counting subsets. A pizza is just a subset of the toppings, because the order you list the toppings in does not change the pizza.
Count every subset at once. Each topping is independently in or out, giving two choices per topping across toppings:
Confirm by adding the subset sizes. The pizzas with toppings number , the same total.
Answer. There are possible pizzas (including the plain pizza with no toppings).
core3 marksA committee of is chosen from women and men. In how many ways can the committee be chosen so that it contains at least women?Show worked solution →
Read 'at least' as a cue for the complement. Counting at least women directly means adding the cases of exactly , and women. The complement (fewer than women, that is or ) is shorter, so subtract it from the total.
Count all committees. Choose any of the people:
Count the unwanted committees ( or woman). With women all come from the men, and with woman the other come from the men:
Subtract.
Check directly. Summing the wanted cases, , which agrees.
Answer. There are committees with at least women.
core3 marksEight points lie on a circle, with no three of them on a straight line. (i) How many straight lines (chords) join pairs of these points? (ii) How many triangles have all three vertices among these points? (iii) How many of those triangles have a particular point, , as a vertex?Show worked solution →
Part (i): a line needs an unordered pair of points. Choosing the two endpoints in either order gives the same line, so count unordered pairs:
Part (ii): a triangle needs an unordered triple of points. No three points are collinear, so every choice of points really does form a triangle:
Part (iii): fix , then choose the rest. With already a vertex, the other vertices are any of the remaining points:
Answer. (i) lines; (ii) triangles; (iii) triangles through .
exam4 marksA netball squad has players, including the captain and two players and who are both injured. A team of is to be chosen. (i) In how many ways can a team of be chosen with no restriction? (ii) In how many ways can the team be chosen if the captain must be included? (iii) In how many ways can the team be chosen if both and must be excluded?Show worked solution →
Part (i): choose from , order irrelevant. A team is an unordered selection:
Part (ii): force in, then choose the rest. With already on the team, choose the other players from the remaining :
Part (iii): remove and , then choose freely. Excluding and leaves available players, from whom all are chosen:
Sanity check on (iii). By the symmetry , which agrees.
Answer. (i) ; (ii) ; (iii) .
exam4 marksA team of is chosen from men and women. (i) In how many ways can the team be chosen with exactly men? (ii) From a separate group of people, a working party of is chosen and then one of those is named as its leader. In how many ways can the working party and its leader be chosen?Show worked solution →
Part (i): choose the men and the women separately, then multiply. The two genders are chosen independently, so use the multiplication principle on two combinations. Choose of the men and of the women:
Part (ii): an unordered choice followed by a distinguished role. First choose the unordered working party of from , then pick its leader from those :
Confirm part (ii) another way. Pick the leader first ( ways), then choose the other members from the remaining : , the same value. The role is a separate, ordered choice on top of the unordered selection.
Answer. (i) ; (ii) .
Related dot points
- Use the multiplication principle to count ordered selections across stages, count ordered selections with repetition as n^r and without repetition as nPr = n!/(n-r)!, and apply restriction techniques: deal with the difficulty first, fix a position, keep items together as a block, and count 'at least one' by the complement
A first-contact answer to the HSC Maths Extension 1 dot point on counting ordered selections. The multiplication principle across stages, ordered selections with repetition (n^r) versus without (nPr), and the restriction moves: deal with the difficulty first, fix a position, keep items together as a block, and count at-least-one by the complement, using the filled-box slot method.
- Apply three reusable counting principles to ordered arrangements with restrictions: keep tied items together by grouping them into a block and ordering inside it, count an unwanted condition and subtract it from the total (the complement), and split an 'or' count into non-overlapping cases, subtracting the overlap by inclusion-exclusion when the cases meet
The three restriction moves the HSC Extension 1 counting dot point keeps testing. Group tied items into a block and order inside it, count the unwanted and subtract it (the complement), and split an 'or' count into cases, subtracting the overlap by inclusion-exclusion. Includes a symmetry shortcut, slot diagrams and a Venn picture.
- Define and evaluate factorials, use the recursive relation n! = n(n-1)!, and simplify factorial expressions and fractions by unrolling and cancelling
A first-contact answer to the HSC Maths Extension 1 dot point on factorial notation. What n! means, why 0! = 1, the recursive rule n! = n(n-1)!, and how unrolling cancels factorial fractions, combines them over a common factorial, and counts trailing zeroes, with the arranging-in-a-row motivation and worked examples.
- Count the distinct arrangements of n objects when some are identical: divide n! by the factorial of each repeat count to get n! over r1! r2! ... rk!, and handle the two-type special case where every object is one of two kinds (2^n arrangements in all, or choose the positions of one kind with a combination)
Why arranging objects with repeats needs n! over r1! r2! ... rk!: plain n! overcounts because swapping identical objects changes nothing, so you divide by the factorial of each repeat count. Covers word arrangements like PRESSES, the binary two-type case (2^n or choose positions with a combination), cases, and separating identical items.
- Use the combination formula to count unordered selections, including with restrictions and complementary counting
A focused answer to the HSC Maths Extension 1 dot point on combinations. The combination formula, why you divide by , key identities, applications including complementary counting, splitting into groups, and at-least/at-most counts, with a stepped diagram and worked examples.
- State and use the binomial theorem, identify general and specific terms, and relate it to Pascal's triangle
A focused answer to the HSC Maths Extension 1 dot point on the binomial theorem. The expansion of using binomial coefficients, the general term , applications to coefficient finding and approximation, and Pascal's triangle built row by row, with worked examples.