Quiz: Game Simulation

Question 1

Monte Carlo simulation gets its name from:

A) A famous mathematician B) A casino in Monaco C) A football stadium D) A computer algorithm

Question 2

The primary advantage of simulation over point predictions is:

A) Faster calculation B) Providing full probability distributions C) Guaranteed accuracy D) Simpler interpretation

Question 3

To achieve a standard error of approximately ±0.5% for a probability estimate, you need approximately:

A) 100 simulations B) 1,000 simulations C) 10,000 simulations D) 1,000,000 simulations

Question 4

NFL team scores are correlated primarily because:

A) Teams copy each other's plays B) Game pace affects both teams similarly C) The football is the same size D) Weather only affects one team

Question 5

In a score-based simulation, the standard deviation of team scores is typically around:

A) 3 points B) 7 points C) 10 points D) 20 points

Question 6

Drive-by-drive simulation differs from score-based simulation by:

A) Being faster to compute B) Modeling the game process, not just outcomes C) Ignoring touchdowns D) Only simulating one quarter

Question 7

The most common drive outcome in the NFL is:

A) Touchdown B) Field goal C) Punt D) Turnover

Question 8

In season simulation, playoff probabilities are calculated by:

A) A single simulation B) Mathematical formula only C) Aggregating many season simulations D) Using last year's results

Question 9

Live win probability during a game depends on:

A) Only the current score B) Score, time, and possession C) Only the teams playing D) The original prediction

Question 10

A game simulation showing 55% home win probability means:

A) Home team will definitely win B) In 1000 simulations, about 550 were home wins C) The spread should be -5.5 D) The simulation is unreliable

Question 11

Score correlation of 0.15 between teams means:

A) Teams will score exactly 15% more B) Scores are slightly positively related C) Home team has 15% advantage D) 15% of games are ties

Question 12

Why might simulated score distributions differ from real NFL scores?

A) Real scores cluster at certain values (7, 10, 14...) B) Simulations are always wrong C) Real games are shorter D) NFL uses different point values

Question 13

Overtime simulation typically shows the coin toss winner has:

A) 30% chance to win B) About 50% chance to win C) 55-60% chance to win D) 90% chance to win

Question 14

To validate a simulation model, you should compare:

A) Only mean scores B) Full distributions against historical data C) Just win percentages D) Number of simulations run

Question 15

Which statistical test compares two continuous distributions?

A) Chi-squared test B) T-test C) Kolmogorov-Smirnov test D) ANOVA

Question 16

Simulation provides "false precision" when:

A) Running too many simulations B) The underlying model assumptions are wrong C) Using a computer D) Results are too consistent

Question 17

In fantasy football, simulation helps by:

A) Guaranteeing player performance B) Providing floor/ceiling projections C) Predicting exact scores D) Eliminating variance

Question 18

A team leading by 14 at halftime typically has win probability of approximately:

A) 55% B) 70% C) 85% D) 95%

Question 19

Sensitivity analysis in simulation involves:

A) Making the simulation more sensitive B) Testing how results change with different inputs C) Removing outliers D) Running fewer simulations

Question 20

The best use of simulation for betting is:

A) Finding guaranteed winners B) Understanding the distribution of possible outcomes C) Predicting exact final scores D) Replacing all other analysis

Answer Key

  1. B - Monte Carlo simulation is named after the famous casino in Monaco, referencing the random nature of the method.

  2. B - Simulation's key advantage is generating full probability distributions, not just single predictions.

  3. C - For a 50% probability, SE = 0.5/√N. For SE ≈ 0.005, need N ≈ 10,000.

  4. B - Game pace affects both teams: fast-paced games tend to be higher scoring for both.

  5. C - Individual team scores typically have standard deviation around 10 points.

  6. B - Drive-by-drive simulation models the actual game process, capturing how scores accumulate.

  7. C - Punts are the most common drive outcome at about 38% of drives.

  8. C - Playoff probabilities come from simulating many seasons and counting how often teams make playoffs.

  9. B - Live win probability depends on score, time remaining, and who has possession.

  10. B - 55% means in repeated simulations, about 55% result in home wins.

  11. B - Correlation of 0.15 means scores are slightly positively related (when one is high, other tends to be slightly higher).

  12. A - Real NFL scores cluster at specific values due to scoring combinations (TD+XP=7, FG=3, etc.).

  13. C - The coin toss winner in overtime wins about 55-60% of the time due to the possession advantage.

  14. B - Proper validation compares full distributions, not just summary statistics.

  15. C - The Kolmogorov-Smirnov test compares two continuous distributions.

  16. B - False precision occurs when more simulations create unwarranted confidence in a flawed model.

  17. B - Simulation helps provide floor (bad game) and ceiling (great game) projections.

  18. C - A 14-point halftime lead typically corresponds to ~85% win probability.

  19. B - Sensitivity analysis tests how changing input parameters affects outputs.

  20. B - Simulation helps understand the range of outcomes and their probabilities, not predict specific results.

Scoring Guide

  • 18-20: Excellent - Ready for advanced simulation projects
  • 15-17: Good - Solid understanding of simulation concepts
  • 12-14: Satisfactory - Review validation and distributions
  • Below 12: Needs Review - Revisit Monte Carlo fundamentals