Prediction Models with Sports Data Quizzes & Answers

Dive into the exciting world of sports analytics with engaging quizzes and expert answers on Prediction Models with Sports Data. Explore the intersection of data science and sport, where predictive modelling is changing the way we understand and analyse sports performance. These quizzes serve as a gateway to unravelling the complexities of building accurate predictive models with sports data, from player performance to game outcomes.

Whether you are a sports enthusiast fascinated by the data behind the game, or a data science enthusiast looking to apply predictive analytics in a sporting context, this collection offers valuable insights into the power of data-driven decision-making in sport. Join us on a journey of statistical discovery as we explore the realm of predictive models using sports data and open up opportunities for data-driven insights and strategic advantages in the world of sports analytics.

Quiz 1: Week 1 – Quiz 1

Q1. From the LPM, what was the regression coefficient for the Pythagorean winning percentage in explaining the binary winning variable, named “WIN”?

Pythagorean Win%: 6.2946
Pythagorean Win%: 2.6461
Pythagorean Win%: 5.542
Pythagorean Win%: 16.393

Q2. From the LPM, what was the R-squared value?

0.102
10.2
1.02
0.0102

Q3. From the Logistic Regression, what was the regression coefficient for the Pythagorean winning %?

30.5739
26.550
34.598
1.027

Q4. From the Logistic Regression, what is the standard error for the Pythagorean winning %?

2.053
20.53
1.027
1.023

Q5. From the Logistic Regression, what fraction of results were correctly predicted %?

Q6. From the Multiple Logistic Regression which incorporated the home team advantage, what fraction of results were correctly predicted %?

Quiz 2: Week 1 – Quiz 2

Q1. After splitting the regular season dataset using the game id (i.e., GAME_ID), how many games were Atlanta Hawks and Chicago Bulls played previously (refer to the “NBA17_pre_team” dataset)?

Atlanta Hawks: 40
Chicago Bulls: 42
Atlanta Hawks: 41
Chicago Bulls: 41
Atlanta Hawks: 41
Chicago Bulls: 42
Atlanta Hawks: 44
Chicago Bulls: 40

Q2. What was the correlation coefficient between the Pythagorean winning % and Winning % in the 1st half of the data set (refer to “NBA17_pre_team” dataset)?

0.78
0.89
0.45
0.91

Q3. What was the winning % of Chicago Bulls in the 2nd half of the dataset (refer to the “NBA_17_post_team” dataset)

Q4. From the forecasting model, what were the regression coefficients for each independent variable (i.e., “wpc_pre” and “pyth_pre”

Pythagorean Win %: 3.75
Win %: 0.825
Pythagorean Win %: 6.25
Win %: 0.756
Pythagorean Win %: 4.55
Win %: 0.567
Pythagorean Win %: 7.55
Win %: 0.625

Week 2: Prediction Models with Sports Data Coursera Quiz Answers

Quiz 1: Week 2 Quiz

Q1. What is the correlation between the home team win probability and home team wins across the entire 2018/19 season?

+0.413
+0.576
+0.397
-0.198

Q2. What is the correlation between the home team win probability and home team wins for games where the points difference was less than 9?

+0.414
-0.198
+0.455
+0.198

Q3. What is the correlation between the home team win probability and home team wins for games where the points difference was greater than 9?

-0.319
+0.321
+0.198
+0.576

Q4. Considering the answers to the last two questions, what do you think is the most likely explanation of these results

Bookmakers make the odds more random to attract bets on close games
The observed differences in the correlations are just random
Bookmakers are not good forecasters
Uncertain games are ones where the bookmakers odds are most likely to be wrong and the scores are likely to be closest

Q5. What is the correlation between the home team win probability and home team wins for games where the game went to overtime?

+0.414
-0.397
+0.319
+0.032

Q6. What is the correlation between the home team win probability and home team wins for games where the game was finished in regular time?

+0.576
+0.414
-0.413
+0.503

Q7. What is the correlation between the home team win probability and home team wins for games where the game was played in calendar year 2018?

+0.338
+0.319
+0.322
-0.322

Q8. What is the correlation between the home team win probability and home team wins for games where the game was played in calendar year 2019?

+0.455
+0.379
-0.438
+0.438

Q9. In which month was the correlation coefficient between the home team win probability and home team wins greatest?

January
October
December
April

Q10. In which month was the correlation coefficient between the home team win probability and home team wins lowest?

March
February
November
October

Week 3: Prediction Models with Sports Data Coursera Quiz Answers

Week 3 Quiz

Q1. Based on the crosstab, what percentage of games did the bookmaker predict correctly over the entire season

Q2. From the ordered logit model, what is the coefficient of the TM ratio variable?

0.1129
-0.6734
0.3356
0.5981

Q3. In the ordered logit model, what is the best we can say about the statistical significance of the TM ratio variable?

It is statistically significant at the 10% level (p-value)
It is statistically significant at the 1% level (p-value)
It is not statistically significant
It is statistically significant at the 5% level (p-value)

Q4. In the logistic regression model, if the ratio of the TM values equaled one, then

The value of the constants alone would determine the probability of a win, draw or loss for the home team
Each team would have an equal chance of winning
The result would be completely random
Each team would be equally good

Q5. Based on the bookmaker odds, what fraction of results were correctly predicted from game 224 onwards?

Q6. Based on the ordered logit model, what fraction of results were correctly predicted

Q7. What was the Brier score derived from the bookmaker odds?

0.562
0.692
0.477
0.587

Q8. What was the Brier score derived from the logistic model?

0.393
0.747
0.399
0.594

Q9. A lower Brier score implies

The match results are more random
The probabilities were closer to the actual the outcomes
The probabilities were further away from the actual the actual outcomes
The match results are less random

Q10. Suppose that the ordered logit model were updated after every game in the season, which of the following is most likely to be true:

The ordered logit model would be more accurate as the season progressed
The ordered logit model would produce more reliable forecasts
The ordered logit model would still perform less well than the bookmaker odds

Week 4: Prediction Models with Sports Data Coursera Quiz Answers

Quiz 1: Week 4 Quiz

Q1. How many games were played in calendar year 2018

1230
542
543
540

Q2. From the logistic model, what is the coefficient of the salary ratio variable?

1.1216
0.4452
5.026
2.482

Q3. In the logistic model, what can we say about the statistical significance on the variables?

Both are statistically significant at the 5% level (p-value)
Both are statistically significant
Both are statistically significant at the 1% level (p-value)
Only the constant is statistically significant at the 5% level (p-value)

Q4. In the logistic regression model, what is the interpretation of the constant (intercept)

It reflects the value of home advantage
It is a random parameter
It is the predicted probability of a home win
It has no natural interpretation

Q5. Based on the bookmaker odds, what fraction of results were correctly predicted

Q6. Based on the logistic model, what fraction of results were correctly predicted

Q7. What was the Brier score derived from the bookmaker odds?

0.394
0.692
0.587
0.477

Q8. What was the Brier score derived from the logistic model?

0.393
0.747
0.399
0.477

Q9. A lower Brier score implies

The match results are more random
The probabilities were further away from the actual the actual outcomes
The match results are less random
The probabilities were closer to the actual the actual outcomes

Q10. Suppose that the logistic model were updated after every game in the season, which of the following is most likely to be true:

The logistic model would produce more reliable forecasts
The logistic model would still perform less well than the bookmaker odds
The logistic model would be more accurate as the season progressed