Self-Assessment Quiz: Discrete Probability
Twenty questions to check your understanding. Answer before opening the key. Aim for 16+.
Question 1
The sample space of an experiment is:
A) the most likely outcome B) the set of all mutually-exclusive, exhaustive outcomes C) any subset of outcomes you care about D) the probability assigned to each outcome
Question 2
An event is formally defined as:
A) a single outcome B) a probability between 0 and 1 C) a subset of the sample space D) a random variable
Question 3
The equally-likely (Laplace) formula $P(E) = \lvert E \rvert / \lvert S \rvert$ is valid:
A) for every sample space B) only when the outcomes of $S$ are equally likely C) only when $E$ is a single outcome D) only for infinite sample spaces
Question 4
By the complement rule, $P(\overline{E})$ equals:
A) $P(E)$ B) $1 - P(E)$ C) $1/P(E)$ D) $P(E) - 1$
Question 5
For events $E$ and $F$, the general formula $P(E \cup F) = P(E) + P(F) - P(E \cap F)$ is needed instead of $P(E) + P(F)$ precisely when:
A) $E$ and $F$ are disjoint B) $E$ and $F$ overlap (are not disjoint) C) $E \subseteq F$ D) the outcomes are not equally likely
Question 6
A random variable $X$ on a sample space $S$ is:
A) a randomly chosen number B) a subset of $S$ C) a function $X \colon S \to \mathbb{R}$ D) the probability of an event
Question 7
For an indicator random variable $\mathbf{1}_A$, the value $E[\mathbf{1}_A]$ equals:
A) $1$ B) $0$ C) $P(A)$ D) $\lvert A \rvert$
Question 8
Linearity of expectation, $E[X + Y] = E[X] + E[Y]$, requires:
A) $X$ and $Y$ to be independent B) $X$ and $Y$ to be indicators C) no special hypothesis — it always holds D) $X$ and $Y$ to have the same distribution
Question 9
In the hat-check problem ($n$ people, random hat returns), the expected number of people who get their own hat back is:
A) $0$ B) $1$ C) $n/2$ D) $n$
Question 10
The expected number of empty buckets when $n$ keys are hashed uniformly into $m$ buckets is:
A) $m/n$ B) $m\left(1 - \tfrac{1}{m}\right)^n$ C) $n\left(1 - \tfrac{1}{n}\right)^m$ D) $m - n$
Question 11
The computational formula for variance is:
A) $\operatorname{Var}(X) = E[X] - (E[X])^2$ B) $\operatorname{Var}(X) = (E[X])^2 - E[X^2]$ C) $\operatorname{Var}(X) = E[X^2] - (E[X])^2$ D) $\operatorname{Var}(X) = E[(X - \mu)]^2$
Question 12
Two random variables both have $E[X] = 100$, but one is constant and one swings between 0 and 200. What distinguishes them?
A) their expectation B) their variance / standard deviation C) their sample space size D) nothing — they are identical
Question 13
Variance can never be negative because:
A) probabilities are at most 1 B) it is the expectation of a squared (hence non-negative) quantity C) the mean is always positive D) standard deviation is its square root
Question 14
The probabilistic method concludes that an object with property $Q$ exists by showing:
A) $P(\text{random object has } Q) = 1$ B) $P(\text{random object has } Q) > 0$ C) every random object has $Q$ D) $E[Q] = 0$
Question 15
The probabilistic method is, at its logical core, an instance of:
A) proof by induction B) proof by contradiction C) direct proof D) a counterexample
Question 16 (True/False, justify)
True or false: If you run a Monte-Carlo simulation 10 million times and the empirical probability comes out to $0.167$, you have proved that the true probability is exactly $1/6$. Justify in one sentence.
Question 17 (True/False, justify)
True or false: $E[X^2] = (E[X])^2$ for every random variable $X$. Justify in one sentence (and name what the difference between the two sides measures).
Question 18 (True/False, justify)
True or false: For the experiment of flipping two fair coins, the correct equally-likely sample space has three outcomes: zero heads, one head, two heads. Justify.
Question 19 (Short answer)
State the three probability axioms (non-negativity, normalization, additivity) in your own words.
Question 20 (Short answer)
In two or three sentences, explain the general strategy "write it as a sum of indicators, then add up the probabilities," and why it is so powerful for computing expected values.
Answer Key
| Q | Ans | Note |
|---|---|---|
| 1 | B | The sample space lists all mutually-exclusive, exhaustive outcomes. |
| 2 | C | An event is a subset $E \subseteq S$; it occurs when the outcome lies in $E$. |
| 3 | B | $P(E) = \lvert E\rvert/\lvert S\rvert$ holds only under the equally-likely assumption. |
| 4 | B | $E$ and $\overline{E}$ are disjoint with union $S$, so $P(E) + P(\overline{E}) = 1$. |
| 5 | B | When $E, F$ overlap, naive addition double-counts $E \cap F$; subtract it once. |
| 6 | C | A random variable is a (deterministic) function from outcomes to numbers. |
| 7 | C | $E[\mathbf{1}_A] = 1\cdot P(A) + 0\cdot(1-P(A)) = P(A)$ — the indicator lemma. |
| 8 | C | Linearity needs no independence; it follows from splitting a sum (Ch. 11). |
| 9 | B | $E[X] = \sum_{i=1}^n \tfrac1n = 1$, independent of $n$ (the hat-check result). |
| 10 | B | Each bucket is empty with prob. $(1-\tfrac1m)^n$; sum $m$ indicators. |
| 11 | C | $\operatorname{Var}(X) = E[X^2] - (E[X])^2$ — one pass, then subtract. |
| 12 | B | Same mean, different spread: variance/standard deviation separates them. |
| 13 | B | $\operatorname{Var}(X) = E[(X-\mu)^2]$ averages non-negative terms. |
| 14 | B | Positive probability forces existence (a probability of 0 would mean none exist). |
| 15 | B | It is proof by contradiction: "no such object" would force the probability to 0. |
| 16 | False | Simulation gives evidence, not certainty; a finite sample never pins the exact value (you'd need the axioms/counting). |
| 17 | False | In general $E[X^2] \ne (E[X])^2$; the gap $E[X^2] - (E[X])^2$ is exactly the variance. |
| 18 | False | Those three are not equally likely; the equally-likely space is $\{HH, HT, TH, TT\}$ (size 4), in which "one head" has probability $2/4$. |
| 19 | — | Non-negativity: $P(E) \ge 0$. Normalization: $P(S) = 1$. Additivity: disjoint events' probabilities add, $P(E \cup F) = P(E) + P(F)$. |
| 20 | — | Express the quantity as $X = \sum_i \mathbf{1}_{A_i}$; then $E[X] = \sum_i P(A_i)$ by linearity. It is powerful because linearity needs no independence, so you never analyze how the (often tangled) pieces interact. |
Topics to review by question
| Questions | Topic | Section |
|---|---|---|
| 1–2 | Sample spaces and events | §20.1 |
| 3 | The equally-likely (Laplace) model | §20.2 |
| 4–5, 19 | Probability axioms and their consequences | §20.2 |
| 6, 18 | Random variables and choosing the sample space | §20.3 |
| 7, 8, 9, 10, 20 | Expected value and linearity | §20.4 |
| 11, 12, 13, 17 | Variance and standard deviation | §20.5 |
| 14, 15 | The probabilistic method | §20.6 |
| 16 | Simulation vs. proof (theme four) | §20.6 |