The title of this book may sound oxymoronic. To many, probability and statistics are not on their radar for fun. Mathematicians, engineers, data researches and the like don't need to be convinced. But as the author explains, everyday life is full of uncertainty. So marketers, salespeople, decision makers - just about everyone - can benefit from thinking about problems in a Bayesian way.
Most are familiar with frequentist statistics which is based on the idea that probability is based on the frequency that something happens. Like tossing a coin. If you keep tossing the coin, the frequency of heads will be about half the time.
Bayesian stats is concerned with "how probabilities represent how uncertain we are about a piece of information." Bayesian says that we are equally unsure whether tossing a coin will yield a heads or tails. Its value comes from making a reasonable decision from available information. So for sports results, elections, traffic patterns and the like the Bayesian model is a solid approach.
The book from no starch press starts with a simple introduction to probability and the connection between Bayesian thinking and everyday reasoning. The author uses an example of seeing a bright light in the night sky and reasoning whether it is a UFO. He leads you through the process of recognizing your existing experience/knowledge, observing data, creating hypotheses and repeating. He stresses that data should inform beliefs as opposed to your beliefs informing data.
This leads pretty quickly to some more advanced tools using logic such as true or false values between 0 and 1 and adding AND, NOT and OR from classical logic. The examples are from the real world such as how to act when you are stopped for speeding. Complicated topics such as Binomial and Beta distributions are explained using the popular Japanese Gacha Games.
Part II introduces conditional probability as opposed to the earlier independent probability using a flu shot example. The real fun comes with using Legos to visualize the space of possible events. It gets better. He then uses an example from Star Wars involving C-3PO and Hans Solo. It keeps the advanced math and logic engaging.
Part III introduces Parameter Estimation whether it is counting jelly beans in a jar at a county fair or estimating snowfall. Normal Distributions are described with the scenario of a cartoon villain setting off a bomb in a bank vault.
Part IV deals with a hypothesis test, an A/B test. It uses the example of testing whether an image in email helps or hurts the conversion rate. The Bayesian Reasoning section uses a classic Twilight Zone episode. That is followed by an example of selecting rubber ducks at a carnival.
Let me be clear that this is not a frivolous book. All the math and logic is here and well presented. It's just that a lot of fun scenarios are used to present the ideas. Each chapter has exercises to try and the answers are online.
The Appendix to Bayesian Statistics the Fun Way: Understanding Statistics and Probability with Star Wars, LEGO, and Rubber Ducks has a nice introduction to the R programming language, a great language for stats and data science. The 2nd appendix is a primer on Calculus which helps with some of the book's examples.
Great Lakes Geek Rating:4.5 out of 5 pocket protectors.
Reviewed by Entreprenerd Dan Hanson, the Great Lakes Geek
What are you reading? Let us know at firstname.lastname@example.org
Top of Page
Back to Great Lakes Geek Book Reviews