Tuesday, November 19, 2013

The Sports Gene, by David Epstein

This book was right up my alley. Vetted science & peer-reviewed data? Check. Jaw-dropping feats of strength, skill, & endurance? Check. Melodramatic scenery chewing? Er....okay, I could have done without some of that. But that's sports, I suppose. Still, with the exception of a few passages, the writing is reasonably good. I appreciate that Epstein opens the book with the tired old question of "OMG nature or nurture????" and then immediately answers it: "Duh, both, *obviously*, and everyone hates that answer. So let's talk about interesting stuff." Over sixteen chapters, Epstein delves into a wide range of issues related to athletic performance and genetics, including the role of practice, sex & hormone differences, "trainability", body type, geography, ancestry, injury & disease, pain tolerance, and will power. (The chapter on breeding champion sled dogs for work ethic instead of speed was particularly interesting.) The upshot: there is no simple answer, but the topic is utterly fascination.

Friday, November 8, 2013

The Predictioneer's Game, by Bruce Bueno De Mesquita

I've been meaning to read this book since it came out in 2009 (I think?) & couldn't have enjoyed it more. In it, political scientist Bruce Bueno De Mesquita discusses how he's spent the last thirty-some-odd years harnessing the power of math & logic (in the form of game theory) to make spookily accurate predictions about business, politics, legal battles, & all kinds of other situations that involve human beings negotiating and scheming for the best outcome possible. Even more impressive, these aren't fields in which he has any particular expertise or access to privileged information--as he reminds us over and over and over again, humans who want things are predictable, which means their behavior can be modeled mathematically. Once you know just a few things about all the major players (level of influence, desired outcome, how much credit they want, how committed to/flexible they are regarding their desired outcome, & whether they have veto power in the situation), it's just a matter of quantifying those things, feeding the values into the model, & interpreting the results.

Impossible, you say? That must be why Bueno De Mesquita's predictions have had something like a 90% success rate over the course of his career.

I also enjoyed the part where BDM engaged in a bit of retroactive predictioneering--ie, using only his prediction model & information widely available at the time to show that both the Cold War & its resolution were basically preordained at the end of World War II, that Britain likely could've prevented WWI, and Sparta might have survived but for its addiction to horse racing. It was also neat to see that predictions he & his students made in the last chapter using the model (the failure of the Copenhagen Conference to do anything about global warming, & the US upping aid to Pakistan to $1.5 billion being enough to get them to stop making side deals with the Taliban Al Qaeda, but not to squash them completely) have come to pass since the book was written in '09.

Brilliant. If you want a friendly, interesting, not-too-technical introduction to game theory (and also how to save hundreds or thousands when buying a car), this book is for you.