First, a disclaimer. I am, professionally, a statistician. I do not have a Ph.D. in my field because I feel that statisticians with Ph.D.’s are devoid of practicality and usefulness to the real world. I work at a factory where I assist engineers in better understanding how processes work and making things better. I generally feel that I make a worthwhile contribution to the world. I bought and read this book because it was critical of statisticians. I do not believe in surrounding myself with ‘y
First, a disclaimer. I am, professionally, a statistician. I do not have a Ph.D. in my field because I feel that statisticians with Ph.D.’s are devoid of practicality and usefulness to the real world. I work at a factory where I assist engineers in better understanding how processes work and making things better. I generally feel that I make a worthwhile contribution to the world. I bought and read this book because it was critical of statisticians. I do not believe in surrounding myself with ‘yes men’ in the form of books and actively seek to challenge my personal beliefs through the things I read and study. Also, the only fields of statistics that I have ever avoided are time series (forecasting) and actuarial science (incredibly boring).
NNT (as he loves to refer to himself in the book) is an idiot. Actually, he’s worse than an idiot, he’s a charlatan of the worst order. If I were NNT I wouldn’t have to defend that statement at all, pretentious phonies reading this to feel intelligent about themselves would nod in agreement at the ‘wisdom’ I’ve laid at their feet. But I’m not NNT and I believe in substantiating otherwise baseless claims.
One of the first things criticized in this book is the narrative for conveying information. Yet that is all NNT does in this book is lay out narrative. No philosopher is quoted, no idea co-opted without some flourishing tale of how they were never appreciated despite their obvious intelligence or of how they were recognized for their genius but the cold, unending march of human forgetfulness relegated them to the annals of history until someone else rediscovered the idea and NNT bought the original book at a used bookstore in some non-American city that has an air of academia to it.
Later in the book, NNT makes one of his few cogent points (I’ll chalk it up to luck on his part). Silent evidence is a major problem everywhere we look and in every field (sadly). The negative studies are almost never published, the failures are not chronicled, etc… The hard thing about silent evidence is that it’s almost never available at all and we rarely recognize that we’re not seeing it. Yet NNT frequently ignores silent evidence. He discusses casinos and all the money they put into preventing cheating (something that, apparently, comes from mediocristan and is easily predictable) but mocks them for doing so because the biggest losses they’d suffered in recent history had nothing to do with cheating. Apparently NNT failed to recognize that perhaps the systems in place so effectively prevented cheating that it was no longer a potential source of lost income. Perhaps if he had looked further back in time he would have seen the financial cost of cheating. It would be like criticizing a store for employing anti-shoplifting techniques when their biggest loses came from a lost shipment, a dishonest accountant and some other unpredictable and essentially unavoidable problem. The suggestion from NNT must be that dealing with the things we can is stupid and we should focus on the things that we cannot predict (and therefore cannot prevent).
NNT spends a whole chapter discussing luck and how every successful economist, banker, investor or other scalable professional is successful not due to skill, but to luck. I’m not going to debate that as the entire premise renders coherent arguments null and void (you cannot disprove the assertion that someone is chronically lucky, well played NNT). However, this luck doesn’t apply to his favorite philosophers. NNT spends a chapter lauding Poincare for being a ‘thinking mathematician’ because he didn’t rely on rigor, but rather intuition. NNT lambastes other mathematicians for criticizing Poincare by calling his techniques ‘hand waving’ which he decides is due to childishness on the part of the other ‘nerd’ mathematicians. But success due to intuition is not success due to skill and is therefore not success to be recognized or rewarded (at least, that’s the case with bankers and investors). NNT doesn’t understand the reason why mathematicians and other ‘hard’ scientists don’t like hand waving is because there’s no way to know if it’s success or luck, it isn’t repeatable and it isn’t verifiable. Also, NNT ignores the silent evidence of intuition. He looks to Poincare as a savior and steward of his profession while ignoring the unmarked graves of all the other ‘thinking mathematicians’ who failed miserably in their intuitive hand waving.
For all of his experimental ‘proof’ offered in defense of claims about how we understand, learn and process things, NNT never gives more than one study as evidence (although he will claim, without a footnote or other reference) that many other studies have verified that particular claim. He accepts these theories as facts and bases large portions of his argument upon them, yet he criticizes doctors, biologists and other scientists for using experimental evidence to make theories on why things work instead of simply accepting that they do work. How many times does he bring up ‘anchoring’ as a theory for why things happen, yet he cannot accept the fact that perhaps birds and humans use different brain regions to perform similar tasks?
NNT at one point is criticizing the models that trader and economists have made to predict the stock market and quotes a study that compared the model’s performance to that of the naive model (today’s value is my guess for tomorrow) which concluded that, “statistically sophisticated or complex models DO NOT NECESSARILY provide more accurate forecasts than simpler ones” (page 154, emphasis added). The study didn’t prove that complex models are no different from simple ones, just that not all of them were better (but no claim that they were ever worse). So why would I get rid of something that doesn’t do worse but could do better? If I buy a lottery ticket that is guaranteed to make me my money back and could make me more than what I paid, why wouldn’t I buy it? Now, I understand that predicting the future is foolhardy, and I’m not saying that it’s something we should put a lot of stock into (pun intended) but past information can give us a general idea about the future, even if it doesn’t give us a great one.
NNT passes himself off as some cool headed, rational thinker who sees beyond the noise and chaos of the world and invites the world to join him on the greener side of the pasture. But nearly all of his arguments are based on contradictions with other arguments that he has made. Further, the remaining arguments that are defensible are impossible to disprove because they impossible to prove. Much like a believing person who argues that without empirical proof of man evolving from lower life or of the big bang, the scientist cannot be right, NNT argues that because models are not perfect, no one can use them to any benefit.
The book is altogether too long given the core point of the book which is this: the most important things that happen (or that don’t happen) are unknowable. Because we cannot predict the future with certainty or even near certainty, we should not even try but rather just do whatever the heck we want because sometimes it’s just as good.