Top critical review
Reviewed in the United States on February 14, 2019
Mr. Crosby has given us five pieces of advice in this book. They are:
• Systems trump discretion.
• Diversification and conviction can co-exist.
• Prepare for bursting bubbles without being too fine-tuned to them.
• Less is more when it comes to information.
• Look for evidence, theory, and roots in behavior.
While I generally agreed with Mr. Crosby’s conclusions, I think there were ways he could have improved his presentation, partially by reversing his fourth piece of advice, at least as it applies to writing the book. I found the book to largely be a well-organized assembly of other people’s work, with brief breaks to discuss them. This is where I had a problem. I don’t think he looked into some of his research examples deeply enough, violating his fifth principle. I will give three examples, all of which I have encountered before reading his book.
In the first example, page 16, cites a Harvard study in which students were paired and given $100 to divide. Person 1 decided how to split it and person 2 decided whether or not to accept it. If person 2 rejected, neither got anything. Harvard found that fifty percent of the low offers were rejected. Mr. Crosby concluded that even a 99 to 1 offer should be accepted because one dollar was better than none. I disagree totally. Person 2 would either consciously or unconsciously be valuing their self in future negotiations, to their own disadvantage. What would happen if person 2 said no and the exercise was repeated a week later with the same players? Would person 1 ask for $99 dollars again? Certainly not. He might be more inclined to ask for $50 while he might have gotten $60 or more each time. Person 2 would be $48 ahead. Harvard could have run this test better, but Crosby failed to see and discuss the possibilities. What if person 1 had been given a specific amount to ask for and each participant had gone through the exercise a number of times at different amounts so that a statistical analysis could have offered information on optimal behavior in this situation? Would this not be good discussion for the investor, especially when evaluating one’s actions when the other person has the hammer? There are more possibilities, but I will move on.
Mr. Crosby cites two pieces of supposed research that claim to prove that people over estimate their own abilities or characteristics. One of his examples involves asking people if they were more athletic than the average person. Eighty percent said they were above average. I have seen this before as it applies to a woman’s looks and to good drivers. None of them define terms clearly, define average, or give the person a chance to evaluate themselves in a meaningful way. As far as I know, none of them account for age, change in physical condition over time, and the effects of opinions offered by others? Even if thirty percent (do the math) of the people overestimate their place on the scale, is it because of their ego, or input from other people that helped them form their opinions? How many people tell a woman that her looks are below average, whatever that is, and how much does that impact her view of herself? Mr. Crosby and the researchers have drawn conclusions based on data, not cause and effect. They haven’t even done that. Supposedly 70% put themselves on the right side of the line. So, more than two against one made correct evaluations about themselves. That's pretty good.
The third example is on page 154. Mr. Crosby does a wonderful job of telling us how to apply statistical analysis to understanding situations. However, his example is horrible and thus his lesson is limited if not almost useless. He tells us that 1 in 1,000 drivers is intoxicated and that breathalyzers are 95% correct, but that in a random check most of the positives are incorrect and gives us the analysis to back this up. As I said, the math is correct. The problem is that the one in 1,000 average is meaningless in real life. Why would I pay the police my tax dollars to catch one in 1,000 people driving drunk? The distribution of drunk drivers is not spatially or temporally even. The police need to give random checks on Las Vegas Blvd. at 1:00 A.M., not at 10:00 A.M. in your neighborhood. Mr. Crosby failed to realize that broad averages are just that and you need specific, applicable information to do good investment analysis. My suggestion for Mr. Crosby is less examples and more in-depth analysis. Do I recommend this book? Sure, just read it carefully.