Study guide
10.2 Combinations, Permutations & Grouping
GMAT Quantitative Reasoning Exam
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Exam Structure, Adaptive Scoring & Problem Solving Fundamentals
1.1 GMAT Focus Quant Overview & Question Mechanics 1.2 Mental Math, Estimation & Non-Calculator Strategies 1.3 Problem Solving Strategies & Elimination Techniques
2 Chapter 2: Number Theory & Integer Properties
2.1 Divisibility Rules, Prime Factorization & Factors 2.2 Greatest Common Factor (GCF) & Least Common Multiple (LCM) 2.3 Division Algorithm & Remainder Arithmetic 2.4 Parity, Signs & Consecutive Integer Properties 2.5 Real Numbers, the Number Line & Properties of Operations
3 Chapter 3: Fractions, Decimals, Ratios & Percentages
3.1 Fraction Arithmetic & Operations 3.2 Decimals, Place Value & Terminating Decimals 3.3 Ratios, Proportions & Variation 3.4 Percentages, Percent Change, Markup & Profit
4 Chapter 4: Exponents, Roots & Radical Expressions
4.1 Laws of Exponents & Exponential Equations 4.2 Radicals & Operations with Roots 4.3 Exponent & Root Inequalities and Traps
5 Chapter 5: Algebraic Expressions, Equations & Inequalities
5.1 Algebraic Manipulation & Special Products 5.2 Linear Equations & Systems of Equations 5.3 Quadratic Equations, Roots & Factoring 5.4 Inequalities, Sign Charts & Absolute Value 5.5 Formulas, Units & Measurement Conversion
6 Chapter 6: Coordinate Geometry & Linear Functions
6.1 The Coordinate Plane, Distance & Midpoint 6.2 Slope, Equations of Lines & Intercepts 6.3 Functions, Notation & Transformations
7 Chapter 7: Rates, Work, Motion & Mixture Word Problems
7.1 Uniform Motion: Rate, Time & Distance 7.2 Relative Speed, Catch-Up & Current/Wind Problems 7.3 Work-Rate & Cooperative Labor 7.4 Mixture & Weighted Average Word Problems
8 Chapter 8: Applied Word Problems, Sets & Sequences
8.1 Overlapping Sets & Venn Diagrams 8.2 Arithmetic & Geometric Sequences 8.3 Simple & Compound Interest Calculations 8.4 Age Problems & Discrete Optimization
9 Chapter 9: Descriptive Statistics & Data Distributions
9.1 Mean, Median & Mode 9.2 Range, Variance & Standard Deviation 9.3 Frequency Distributions & Distribution Properties
10 Chapter 10: Permutations, Combinations & Probability
10.1 Fundamental Counting Principle & Factorials 10.2 Combinations, Permutations & Grouping 10.3 Probability Rules: Independent & Dependent Events 10.4 Complementary Probability & Multi-Stage Events
10.2 Combinations, Permutations & Grouping
Key Takeaways Explain This Topic AI
P(n, k) = n! / (n − k)! = n × (n − 1) × ... × (n − k + 1) when order matters: rankings, codes, assigned seats, titled offices.
C(n, k) = n! / (k! (n − k)!) = P(n, k) / k! when order does not matter: committees, hands, subsets. C(n, k) = C(n, n − k).
Groups of groups: multiply the combination (or permutation) count for each subgroup. If the 4 chosen people then elect a chair, multiply by 4.
Must include person A in a k-person committee from n: C(n − 1, k − 1). Cannot include B: C(n − 1, k). Exactly r from a special subgroup of size s: C(s, r) × C(n − s, k − r).
Circular arrangements of n distinct people are (n − 1)! only when seats are unlabeled around a table. GMAT rarely needs this; numbered seats are n!.
Last updated: August 2026
Order Versus Unordered Groups
Official Guide Math Review 3.4 asks you to count arrangements and selections . The entire skill is one question: does order matter?
If swapping two chosen people (or digits, or books) creates a different outcome — a ranking, a password, a labeled office — you are permuting. If swapping them creates the same committee, hand, or subset, you are combining.
- Situation Order? Tool Compact formula
- President, VP, treasurer from n people Yes Permutation P(n, 3) = n(n−1)(n−2)
A 3-person committee from n people No Combination C(n, 3) = n(n−1)(n−2) / 3!
- 4-digit codes with distinct digits, leading zero allowed as a code Yes Permutation P(10, 4)
- 4-digit integers with distinct digits Yes, plus a leading-zero ban Slot method 9 × 9 × 8 × 7
- Choose 2 starters from 5, order unused No Combination C(5, 2) = 10
On the 21-question, 45-minute, no-calculator section, write a short falling product (and divide by k! only when the group is unlabeled). You will not expand 12! in full.
- Permutations: P(n, k)
- P(n, k) = n! / (n − k)! = n × (n − 1) × ... × (n − k + 1)
You are lining up k distinct objects chosen from n distinct objects. The first slot has n options, the second n − 1, down to k slots.
Worked example. Eight candidates; choose a president, a vice president, and a treasurer, all different people.
P(8, 3) = 8 × 7 × 6 = 336.
Alice as president and Ben as VP is different from Ben as president and Alice as VP. Order matters, so do not divide by 3!. Dividing would produce C(8, 3) = 56, which counts unlabeled trios, not titled offices.
Worked example. How many 3-letter codes from 26 distinct letters, no repeats? P(26, 3) = 26 × 25 × 24 = 15,600. With repeats allowed it is not a permutation of distinct letters: 26³ = 17,576.
P(n, n) = n! / 0! = n!. Arranging all n objects is the factorial from 10.1. P(n, 1) = n. P(n, 0) = 1.
Cancel before you multiply: P(10, 4) = 10 × 9 × 8 × 7 = 5,040, which is 10! / 6!, not a request to write 3,628,800 / 720 by hand.
Combinations: C(n, k)
C(n, k) = n! / (k! (n − k)!) = P(n, k) / k!
The extra k! in the denominator erases order. Every unordered group of k people would have been counted k! times as a permutation — once for each way to line those k people up.
Worked example. A 4-person committee from 10 people.
C(10, 4) = (10 × 9 × 8 × 7) / (4 × 3 × 2 × 1) = 5,040 / 24 = 210.
If the stem had said "chair, secretary, treasurer, and member" as four distinct titles, you would have kept 5,040 = P(10, 4). Titles restore order.
Worked example. A 5-card hand from 12 distinct cards (a toy deck so the arithmetic stays small): C(12, 5) = (12 × 11 × 10 × 9 × 8) / 120 = 95,040 / 120 = 792.
Identities you actually use
C(n, k) = C(n, n − k). Choosing k to include is choosing n − k to leave out. C(10, 8) = C(10, 2) = 45, which is faster than building C(10, 8) as an 8-factor product.
C(n, 0) = C(n, n) = 1. One empty committee; one committee of everybody.
C(n, 1) = n.
C(n, 2) = n(n − 1) / 2. Pairs. Memorize this; it is the most common combination on Quant.
- n choose k Value Fast route
- C(7, 2) 21 7 × 6 / 2
- C(8, 3) 56 8 × 7 × 6 / 6
- C(9, 3) 84 9 × 8 × 7 / 6
- C(10, 3) 120 10 × 9 × 8 / 6
- C(10, 4) 210 10 × 9 × 8 × 7 / 24
- C(6, 2) 15 6 × 5 / 2
If k > n, C(n, k) = 0: you cannot choose 5 people from 4. That is an impossible case contributing 0, not a broken formula.
Groups of Groups
When a team has a stated makeup — 3 women and 2 men, 2 officers and 2 staff — multiply the combination counts for each subgroup. You are using FCP on the subgroups.
Worked example. 8 women and 6 men. A 5-person committee with exactly 3 women (hence exactly 2 men).
C(8, 3) × C(6, 2) = 56 × 15 = 840.
Not C(14, 5) = 2,002, which ignores the gender quota. Not C(8, 3) = 56 alone, which never picks the men.
Worked example. 5 men and 4 women; a 4-person team with 2 of each, and then the 4 elect a chair from among themselves.
First choose the people: C(5, 2) × C(4, 2) = 10 × 6 = 60.
Then choose the chair: 4 options.
Total: 60 × 4 = 240.
The chair step is a 1-slot permutation among the already-chosen 4. Mixing "choose the group, then assign a title inside it" is standard GMAT glue.
If the groups themselves are labeled (Team A and Team B, or a day shift and a night shift), two assignments of the same people to opposite labels are different. If the groups are unlabeled (just "two pairs"), divide by the number of ways to shuffle identical group labels. GMAT usually labels the groups, so you multiply and stop.
Worked example of unlabeled pairs, because it catches people. Six people split into 2 unlabeled teams of 3.
C(6, 3) = 20 chooses Team 1, the rest are Team 2 — but that double-counts every partition, because the other trio could have been chosen first. Unlabeled: 20 / 2 = 10.
If the teams are "Room 1" and "Room 2," keep 20.
Restrictions: Must Include / Cannot Include
Name the restricted people first, then choose the rest from whoever remains eligible.
Must include person A in a k-person committee from n people: A is seated. Choose k − 1 more from the other n − 1: C(n − 1, k − 1) .
Cannot include person B : drop B. Choose k from n − 1: C(n − 1, k) .
Must include A and cannot include B : seat A, drop B, choose k − 1 from n − 2: C(n − 2, k − 1) .
Exactly r of a special subgroup of size s : C(s, r) × C(n − s, k − r), provided 0 ≤ k − r ≤ n − s. If the second index is impossible, that case is 0.
Worked example. 12 people, 3 of whom are officers. A 4-person committee must include exactly 2 of the 3 officers.
C(3, 2) × C(9, 2) = 3 × 36 = 108.
If the stem said at least 2 officers, add the exactly-2 case to the exactly-3 case: 108 + C(3, 3) × C(9, 1) = 108 + 1 × 9 = 117.
If the stem said at most 1 officer: C(3, 0) × C(9, 4) + C(3, 1) × C(9, 3) = 126 + 3 × 84 = 126 + 252 = 378.
Worked example, must include / cannot include together. A 4-person committee from 10 people must include Ana and cannot include Ben. Seat Ana, drop Ben, choose 3 more from the remaining 8: C(8, 3) = 56.
A short note on circular arrangements
If n distinct people sit around a round table and only relative order matters (rotating everyone one seat is the same seating), the count is (n − 1)! . Fix one person as a reference, then arrange the other n − 1: 6 people around a table → 5! = 120.
GMAT Quantitative Reasoning rarely needs this. If seats are numbered, or there is a distinguished head of the table, it is an ordinary line: n!. If the problem never mentions a circle, do not divide by n. One sentence of awareness is enough; do not build a circular-permutation toolkit, and do not flip a necklace (that extra ÷ 2 is not a GMAT move).
Traps on Quant
Using C(n, k) for titled roles (president / VP).
Using P(n, k) for an unlabeled committee.
Forgetting to divide P(n, k) by k! when converting to C(n, k).
Computing C(n, k) when a quota requires a product of two combinations.
"At least two" computed as a single C(s, 2) term instead of summing feasible r.
Circular division on a row of seats.
Treating C(n, k) with k > n as undefined rather than 0 for that case.