Monday, April 20, 2015
Interesting and Informative, or, I Finally Read Some Non-Fiction
Other than the intro/conclusion chapters, each chapter looks at a specific topic often subjected to statistical analysis, or more specifically, to "forecasting" (predicting the future based on present / past data). So each chapter is an informative read on how forecasting works in a certain field, ranging from weather, to sports, to climate change. I don't think there was one topic I had in-depth knowledge of, so each chapter was an education in itself. Some hold more interest to me than others (climate change more than baseball :)) and some more day-to-day relevance (weather) but all interesting and good reading. It really gave me a new way to think about each of these fields. On climate change specifically, I really appreciate Silver's temperate approach toward the science and have more-or-less adopted it as my own (for the record, the models are by no means perfect and certainly have something to answer for, but they do all suggest rises in carbon dioxide are going to be problematic if the models are accurate).
And besides for educating me in a few specific topics, the book also introduced me to the notion of Bayesian forecasting. I know about Bayes but certainly hadn't thought to apply it as widely as Silver does. It really is a clarifying way to approach predictions of both the future and general unknowns. It's very helpful to think in terms of current assumptions and how they affect the assumptions about new propositions (not that I mean I use this in any real sense, just in my rich intellectual life :p).
As someone with no statistical background, no background in any of the topics Silver discusses and not really a critical thinker, it's not like I'm a in a place to really evaluate Silver's book on its academic merits. Suffice it to say, it was not only an interesting read, but has contributed to the way I think about the climate change debate, weather forecasts, mortgage securities, earthquake predictions, and Bayesian reasoning. Not bad :)
Verdict: 4/5 (based purely on how much I enjoyed this book, not on its technical merits or lasting impact)