Wednesday, 7 August 2019

Reading questions: Struck by Lightning


The book Struck by Lightning, by Jeffery Rosenthal hits that balance of scientific correctness and approachability just right for a general audience book on probability. It’s been in print for 14 years now, and was a Canadian bestseller, so there’s nothing new that would come from a traditional review and therefore I won’t write one.

Instead, two points:

1.      It should be required K-12 or 100-level math/stats reading.
Most of this demographic would get 10 times as much long-term benefit from Struck by Lightning than they would from an equal amount of time spent on probability exercises. Those that are math and math-adjacent fields would still benefit, and will probably get around to those exercises anyways.

2.      I wish I’d written this book first.

It’s easy enough to say this book should be classrooms, so let’s back that up with some teacher support in the form of 30-odd reading and math questions that require the book. (Pages may vary by printing)

15-20 of these will appear with answers in my own book, Writing for Statistics and Data Science.

(p.13) How might Milgram's experiment have gone differently if the packages were originally sent all over the world? What if the experiment was restricted to just one city?

(p.16) How can the average number of matches go higher than 1 if the probability isn't going past 100%

(p.16) If there were 500 days in a standard year instead of 365, would the probability of a match be higher for 4,10,20, ... , 40,45,50 people?

(p.19) By the Poisson distribution, the chance of X murders is exp(-lambda)*lambda^X / X!  where lambda is equal to the average. What is the probability of 0 murders in Toronto? Of 1? 2? 5? 10 murders?

(p.24) What is the other name for average value given in the "Laying Down the Law" chapter?

(p.28) In European style roulette, there are 18 red, 18 black, and only 1 green "zero" space. What are the probabilities and expected losses for the bets in Table 3.1, but for European style roulette?

(p.46) Why can't players blame their bad performance in duplicate bridge on bad cards?

(p.49) Why is "drawing to an outside straight" better than "drawing to an inside straight"?

(p.51) In poker, bets of good players tend to get larger as the money at stake, the "pot" gets larger. Why?

(p.56) Why do blackjack tables in casinos use 6 or more decks of cards at the same time?

(p.58) In blackjack, is it better for the player if the dealer starts with a 5, or with a 9? Why?

(p.70) How can we conclude that the rate of murder is going down in the Canada if the number of murders in total is staying the same?

(p.84) Why is it rational to wear a seatbelt even if car accidents are rare?

(p.91) Most people will choose to win a guaranteed $500,000 over a 50% chance of winning $1 million. Using utility theory, explain why this makes sense.

(p.101) Why was the 5% standard threshold for p-values chosen?

(p.103) What is the name of the problem that comes from choosing people to observe in a non-random fashion?

(p.109) What is the name of the problem that comes from selecting one of many studies to reveal, even if all the studies are conducted properly.

(p.112) Give another example of two (likely or known) correlated phenomena and a possible common cause.

(p.119) Do you expect Barbados to stay near the top of countries with a rate of lightning fatalities or to move around a lot? What about South Africa? Explain the difference or similarity.

(p.150) Why do some voters declare themselves undecided in a poll, but vote differently in an election or referendum.

(p.164) What is the primary difference between probability theory and statistical inference?

(p.165) In the 1/20 times that the error in a poll is larger than the stated margin of error, is the error more likely to be far beyond the margin or a little beyond the margin?

(p.178) How is haphazardly (using only your imagination) writing down a sequence of numbers different from rolling 6-sided dice and writing down the sequence?

(p.186) What was the first use of Monte Carlo simulation?

(p.188) How is Markov chain Monte Carlo different from ordinary Monte Carlo simulation?

(p.191) In what situation could one anticipate the start of a race even though the start is a random time between 2 and 6 seconds after a beep.

(p.191, advanced) What feature of the exponential distribution makes it "impossible to anticipate"?

(p.201) What is the name for disease protection granted from people around being immunized?

(p.219) What are the two groups of statisticians that you can easily start a fight between?

(p.228) What is the main feature of emails that filter programs use to identify spam?

(p.240) What is a fair way to judge a source of predictions of binary events like "will it rain tomorrow or not?" or "which team will win the sportsball?”


Want more? Dr. Rosenthal has written his own classroom discussion questions for the book.



No comments:

Post a Comment