Symmetry and Exchangeability

Use symmetry among exchangeable random variables to shortcut counting arguments.

1. The idea

When random variables are exchangeable — their joint distribution is unchanged by permuting their labels — every ordering of their values is equally likely. This turns a question about a specific ordering into a simple counting problem over permutations.

A classic example: if you deal cards from a shuffled deck, the position of any particular card (say, the ace of spades) is uniform over all positions, by symmetry, without needing to track the shuffling mechanics.

Symmetry arguments are fast and rigorous once exchangeability is established, but that assumption must actually hold — the technique fails if there's dependence that breaks the symmetry (e.g. sampling without replacement across dependent draws where the marginal distributions differ).

2. Why it matters

Symmetry is the fastest tool available in an interview: it converts a seemingly complex ordering or ranking question into a ratio, one divided by the number of equally likely orderings, often solvable in seconds.

3. Core mathematics

Uniform rank among exchangeable variables

P (X_{i} has rank k) = \frac{1}{n}

For $n$ exchangeable, continuous (a.s. distinct) random variables.

Specific ordering of exchangeable variables

P (X_{1} < X_{2} < \dots < X_{n}) = \frac{1}{n !}

4. How to recognize it

Think of this when…

The question asks for the probability of a specific relative order among iid or exchangeable variables.
"Which card/ball/customer comes first/last" type questions with symmetric setups.
Random shuffles or random arrival orders with no distinguishing information among items.
You are tempted to compute a complicated conditional probability, but the variables are actually exchangeable.

5. Worked example

Worked example

Three people A, B, C arrive at random, uniformly over all orderings. What is $P (A arrives before B before C)$ ?

1
There are $3! = 6$ equally likely orderings of A, B, C by symmetry (no ordering is favored).
2
Exactly one of these orderings is A before B before C.
3
Apply the uniform-ordering formula.
$P (A < B < C) = \frac{1}{3 !} = \frac{1}{6}$
Answer: $P (A before B before C) = 1/6$

6. Common mistakes

Applying symmetry when the variables are not actually exchangeable (e.g. they have different distributions).
Forgetting ties: symmetry arguments for strict orderings require continuous (a.s. distinct) variables, or explicit tie-breaking.
Overcounting orderings by treating some permutations as indistinguishable when they are not.
Using symmetry to compute a marginal probability when the question actually requires a conditional one.

7. Quick check

Four distinct cards are shuffled uniformly at random. What is $P (card 1 is on top)$ ?

Five iid continuous random variables are drawn. What is $P (X_{3} is the maximum)$ ?

8. Practice

30 QuantSolve questions use this concept.

Practice this concept

9. Related methods

Symmetry

Create a free account to save lesson progress.