Chapter 28: Quiz

Instructions

Select the best answer for each question. Some questions may have multiple correct-seeming answers; choose the one that is MOST correct or MOST complete based on the chapter material.

Question 1

What is the primary purpose of a prediction market?

A) To provide entertainment through wagering B) To produce well-calibrated probability estimates by aggregating beliefs C) To generate trading fees for the platform operator D) To identify which traders are most knowledgeable

Answer: B

Explanation: While prediction markets may serve other purposes (entertainment, fee generation, identification of superforecasters), their primary purpose is to aggregate dispersed beliefs into price signals that serve as probability estimates. All design decisions should be evaluated primarily against this criterion.

Question 2

In the SMART framework for market design, what does the "A" stand for?

A) Actionable B) Achievable / Assessable (Resolvable) C) Accurate D) Automated

Answer: B

Explanation: In the SMART framework adapted for prediction markets, A stands for Achievable or Assessable, meaning the question must be resolvable --- the answer can actually be determined by reference to observable data. A question that is specific and measurable but whose answer can never actually be determined (like "What is China's actual GDP?") fails this criterion.

Question 3

Which of the following is the BEST resolution source for a market on US inflation?

A) A major newspaper's reporting on inflation B) The Bureau of Labor Statistics CPI release C) A Federal Reserve governor's speech mentioning inflation D) A poll of economists' inflation estimates

Answer: B

Explanation: The BLS CPI release is a Tier 1 source (official government statistics). It is the primary, authoritative data source for US inflation. Newspapers (Tier 3), Fed speeches (commentary, not data), and polls (opinions) are all inferior sources for resolution.

Question 4

A market asks: "Will Apple or Google release a VR headset in 2026?" What is the primary wording pitfall?

A) Scope creep B) Moving goalposts C) Ambiguous "or" (inclusive vs. exclusive) D) Temporal ambiguity

Answer: C

Explanation: The word "or" is ambiguous. It could mean: (a) at least one of them releases one (inclusive or), or (b) exactly one of them releases one (exclusive or). The fix is to specify "at least one of Apple or Google" for inclusive, or create separate binary markets.

Question 5

What is a self-referential market?

A) A market that references another market on the same platform B) A market where the market's own price or existence influences the outcome C) A market where the creator is also a trader D) A market that asks about the prediction market industry itself

Answer: B

Explanation: A self-referential market is one where the market's own price influences the outcome, such as "Will this market's price be above 50% at close?" This creates logical paradoxes and indeterminate prices, violating the fundamental purpose of prediction markets to aggregate beliefs about external events.

Question 6

When should a market resolve as N/A (void)?

A) When the outcome is not what the market creator expected B) When the question becomes meaningless, resolution sources are unavailable, or a fundamental error is found C) When fewer than 10 traders participate D) When the market price never moves from its initial value

Answer: B

Explanation: N/A resolution is appropriate when the question becomes meaningless (e.g., subject dies), resolution sources are permanently unavailable, or the question contained a fundamental error. It is not appropriate based on the creator's expectations or participation levels.

Question 7

For a multi-outcome market, which two properties must the outcome set satisfy?

A) Relevance and time-boundedness B) Mutual exclusivity and exhaustiveness C) Specificity and measurability D) Liquidity and participation

Answer: B

Explanation: For multi-outcome markets, outcomes must be mutually exclusive (no overlap, so each possible world-state maps to exactly one outcome) and exhaustive (all possible world-states are covered). This ensures prices sum to 1 and there is no ambiguity about which outcome applies.

Question 8

What is the Herfindahl Index used to measure in prediction markets?

A) Market calibration B) Bid-ask spread C) Trader diversity / concentration D) Resolution dispute frequency

Answer: C

Explanation: The Herfindahl Index (HHI = sum of squared market shares) measures trader concentration. Lower values indicate more diverse participation. A market dominated by one trader has HHI close to 1, while a market with many equal-sized traders has HHI close to 1/n.

Question 9

Which subsidy decay function is described by $s(t) = s_0 \cdot e^{-\lambda t}$?

A) Linear decay B) Exponential decay C) Step function decay D) Logarithmic decay

Answer: B

Explanation: This is exponential decay. The subsidy decreases by a constant percentage per unit time, never quite reaching zero but becoming negligibly small. This provides strong initial liquidity that gracefully transitions to organic market-making.

Question 10

What is the Brier score for a prediction of 0.80 on an event that actually occurred (outcome = 1)?

A) 0.04 B) 0.16 C) 0.20 D) 0.80

Answer: A

Explanation: Brier score = $(f - o)^2 = (0.80 - 1)^2 = (-0.20)^2 = 0.04$. Lower Brier scores indicate better predictions. A perfect prediction of 1.0 would yield a Brier score of 0.

Question 11

What is the "cold start problem" in prediction markets?

A) Markets perform poorly in cold weather B) New traders lose money on their first trades C) New markets have no liquidity, but traders avoid illiquid markets D) Market prices start at 50% regardless of the true probability

Answer: C

Explanation: The cold start problem is the chicken-and-egg problem where new markets have no liquidity and traders avoid illiquid markets, but markets need traders to become liquid. Subsidization and AMM seeding are the primary solutions.

Question 12

Which of the following is NOT a recommended practice for bracket market design?

A) Making brackets mutually exclusive B) Using equal-probability brackets based on priors C) Having brackets overlap at boundaries for safety D) Matching bracket precision to the resolution source's precision

Answer: C

Explanation: Brackets must NOT overlap. Overlapping boundaries create ambiguity about which outcome a data point falls into. Boundaries should use half-open intervals (e.g., [3.0%, 3.5%)) to ensure every possible value maps to exactly one bracket.

Question 13

According to the resolution source hierarchy in this chapter, which tier do major media organizations (Reuters, AP) fall into?

A) Tier 1 B) Tier 2 C) Tier 3 D) Tier 4

Answer: C

Explanation: Major media organizations are Tier 3. They are useful for event-based resolution but are subject to correction and retraction, making them less reliable than official government statistics (Tier 1) or major institutional data (Tier 2).

Question 14

A market on "Will the CEO of TechCorp announce a new product in Q2 2026?" is primarily vulnerable to which design problem?

A) Scope creep B) Self-referential paradox C) Insider-triggerable outcome D) Temporal ambiguity

Answer: C

Explanation: This market is insider-triggerable because TechCorp employees (especially the CEO and their direct reports) know the announcement schedule and can trade with near-certainty. The outcome is directly controlled by a small group who could profit from their foreknowledge.

Question 15

What is the recommended way to handle negative framing in market questions?

A) Add "NOT" in capital letters for emphasis B) Use double negation for clarity C) Reframe using positive framing and let the price reflect the negative probability D) Create two separate markets: one positive and one negative

Answer: C

Explanation: Negative framing ("Will X NOT happen?") is confusing because YES means nothing happened. The recommended approach is positive framing ("Will X happen?") and letting the price naturally reflect the negative probability (a price of 0.20 means 80% probability it will not happen).

Question 16

In the market lifecycle, what should happen during the "dispute window"?

A) All trading is frozen and new positions cannot be opened B) Traders can contest the preliminary resolution with evidence C) The market creator changes the resolution criteria D) The AMM automatically adjusts prices to match the resolution

Answer: B

Explanation: The dispute window is a period (typically 24-72 hours) after preliminary resolution is posted during which traders can contest the resolution with evidence. If no valid disputes are raised, resolution is finalized and settlement occurs.

Question 17

What composite market quality score formula does the chapter recommend?

A) $Q = \text{Volume} \times \text{Accuracy}$ B) $Q = \sum w_j \cdot \tilde{m}_j$ (weighted sum of normalized metrics) C) $Q = \frac{\text{Traders}}{\text{Disputes}}$ D) $Q = 1 - \text{Brier Score}$

Answer: B

Explanation: The chapter recommends a composite score that is a weighted sum of normalized metric values across multiple dimensions: calibration (30%), participation (20%), liquidity (20%), low dispute rate (15%), and low N/A rate (15%).

Question 18

For the bid-ask spread formula $\text{Spread} = \frac{p_{\text{ask}} - p_{\text{bid}}}{p_{\text{mid}}} \times 100\%$, what does a spread of 2% indicate compared to a spread of 10%?

A) The 2% market is less liquid B) The 2% market is more liquid with lower transaction costs C) The 10% market has more traders D) The spreads are not comparable without knowing the price level

Answer: B

Explanation: A tighter (lower) bid-ask spread indicates better liquidity and lower transaction costs for traders. A 2% spread means a trader loses approximately 2% in transaction costs on a round trip (buy then sell), compared to 10% in the wider-spread market.

Question 19

What is the main advantage of template-based automated market creation?

A) It eliminates the need for human review entirely B) It generates hundreds of consistent-quality markets from a single template C) It always produces better questions than human designers D) It reduces the need for resolution criteria

Answer: B

Explanation: Template-based generation can produce hundreds of markets from a single well-designed template, maintaining consistent quality across all generated markets. This is its primary advantage for scaling. It does not eliminate the need for human review or resolution criteria.

Question 20

Why might a market on "Will humans visit Mars?" be poorly designed even though it seems clear?

A) It is not relevant to anyone B) "Visit" is ambiguous (flyby? orbit? landing?), "Mars" could include moons, and there is no time bound C) The resolution source is not specified D) Both B and C

Answer: D

Explanation: This question has multiple problems: "visit" is ambiguous (does a flyby count? orbital insertion? landing?), "Mars" could be interpreted to include Phobos and Deimos, there is no time bound (the market could remain open indefinitely), and no resolution source is specified. Both B and C are correct, making D the best answer.

Question 21

According to the chapter, what is the optimal initial AMM price for a binary market where the base rate for the event is approximately 5%?

A) Always 50% (0.50) to be neutral B) Close to 5% (0.05) to reflect the prior C) 25% (0.25) as a compromise D) It does not matter as long as there is enough liquidity

Answer: B

Explanation: The chapter specifically notes that setting a 50/50 price is not always appropriate. If the base rate is 5%, starting at 50% invites easy arbitrage that depletes the subsidy. The initial price should reflect a reasonable prior estimate, in this case close to 5%.

Question 22

What is the purpose of including an "Other" or "None of the above" outcome in a multi-outcome market?

A) To provide a low-probability option for risk-seeking traders B) To ensure the outcome space is exhaustive (covers all possibilities) C) To increase the number of tradeable contracts D) To make the market more interesting

Answer: B

Explanation: The "Other" outcome serves as a catch-all to ensure exhaustiveness. Without it, an unlisted outcome (e.g., a surprise third-party candidate in an election market) would have no corresponding contract, making the outcome space incomplete and prices unable to sum to 1.

Question 23

The chapter describes the tension between precision and participation. What is the recommended strategy to balance this?

A) Always prioritize precision over participation B) Always prioritize participation over precision C) Use a clear, simple title with detailed resolution criteria in the description D) Create two versions of each market: one simple, one detailed

Answer: C

Explanation: The recommended approach is a layered design: a clear, simple title that attracts participants and conveys the essence of the question, paired with detailed resolution criteria in the market description that provide the precision needed for unambiguous resolution.

Question 24

Which of the following is a valid reason to use scalar markets instead of bracket markets?

A) Scalar markets are always more liquid B) Scalar markets capture full distributional information without requiring the designer to choose thresholds C) Scalar markets are simpler for traders to understand D) Scalar markets do not require specifying a resolution source

Answer: B

Explanation: Scalar markets capture the full distribution of beliefs about a continuous outcome without requiring the designer to pre-specify bracket boundaries. This is their main advantage. However, they are more complex to understand and implement, not simpler.

Question 25

According to the case studies, what is the main lesson from the "Brexit" market design problems?

A) Political markets should be avoided entirely B) Even seemingly simple questions can have multiple reasonable interpretations of the triggering event C) Markets should only use government data sources D) Long-running markets always produce better forecasts

Answer: B

Explanation: The Brexit case study illustrates that even a seemingly straightforward question ("Will Brexit happen?") can be interpreted in multiple ways: referendum passing, Article 50 triggered, UK leaving, transition period ending, etc. Markets with different implicit definitions showed different prices, causing confusion. The lesson is that the triggering event must be precisely defined.