Thoughts on The Drunkards Walk

Recently, I finished the Drunkard’s Walk by Leonard Mlodinow. Thought it was a good read, especially for someone who isn’t very familiar with statistics or with Kahneman’s work.

A big point of the book is to hammer the idea that a lot of the world is up to chance, despite the way that we often interpret events. So Mlodinow looks at things that happen to him and others through a statistical lense. As an example, there was a baseball player named Roger Maris that broke Babe Ruth’s home run record in 1961 (of 60 homers), despite the fact that Maris had never hit more than 39 home runs in any previous year. The way Mlodinow describes it is Maris was an above average home run hitter who hit a home run in around one out of every 15 at bats normally. A novice statistician might look at this case and say that the odds that Maris were to hit as many home runs as he did were 1 out of 32 based on his previous skillset if it were purely by luck. The conclusion would therefore be that there must be another explanation for his success that year. That makes a lot of sense, but Mlodinow instead looks at it like this: between when Ruth set the record in 1927 and 1961, there was a player in one of every three years that was as strong of a home-run hitter as Maris. That means that the probability, by purely chance, that one of these players were to break Ruth’s record was a bit over 50%. The fact that it happened to be this guy was likely completely coincidental.

The book dove into some of the biases that we have, many examples of statistical errors that have been made by people (I’ll give a few below), and gave a basic foundation of some statistical concepts that I have learned in my stat classes.

All of that was interesting, but much was familiar. (if you read it and have that knowledge, you could probably skip around).

My biggest takeaways were the following:

Unlikely scenarios occur more often than you might imagine

This seems like a low hanging fruit, but there are these crazy things that have extremely low odds of happening (like someone breaking Babe Ruth’s record). The nature of society, though, is that there are a ton of people that are in these unlikely situations, so even if the chances that something happen are one in one thousand, if there are a million people that have that chance, then 1000 people will still have it happen to them. It basically just makes me realize that uncommon events occur more frequently than we’d think.

Thinking about life in terms of expected payoff can be kind of funny (and dangerous)

An example that Mlodinow gave was that in all of the time that he has had his car, he has dinged his mirrors backing out of his driveway three times. If you count the cost of those incidents and all of the times he backed out of his driveway, the expected value for him to back up was something like $2. The equivalent of there being a toll at his house to enter his driveway each time he drove up it.

We could think about this in terms of parking tickets or other risks we take where there is a potential for loss. Doing so could end up saving you money, but it’s often not worth the effort or energy.

There are some really big mistakes that are made because people don’t understand statistics

Researchers asked physicians to estimate the probability that an asymptomatic woman between the ages of 40 and 50 who has a positive mammogram actually has breast cancer if 7 percent of mammograms show cancer when there is none. In addition, the doctors were told that the actual incidence was about 0.8 percent and that the false-negative rate about 10 percent. Putting that all together, one determine that a positive mammogram is due to cancer in only about 9 percent of the cases. In the German group, however, one-third of the physicians concluded that the probability was about 90 percent, and the median estimate was 70 percent. In the American group, 95 out of 100 physicians estimated the probability to be around 75 percent.

That’s really bad.

It doesn’t just happen in hospitals either. There are wrongly punished people in court, wrong decisions from life insurance companies and so on, because people miss this relatively simple idea of base rates.

It just makes you realize that people who share stats with you might not be right, and being skeptical is never a bad idea.

The future is hard to predict, but the past is easy to understand

There are all of these things that happen in our lives and on a more macro scale that are very bad. Pearl Harbor, 9/11, car accidents, being left at the altar by your loved one, etc. Whenever something like this happens, it is extremely easy to look back and see all of the signs. For Pearl Harbor, there were tons. A few months before, there was a letter intercepted saying the Japanese were trying to find the locations of US battleships on the island. Then the Japanese attack ships had been MIA for a while. There was even a note intercepted a day before the attack that was a bit foreboding. Yet, nobody caught the attacks beforehand and afterwards people looked back and pointed fingers at all of the missed signs.

That’s one way to look at things. The other perspective is that there was no way to know that those events were going to lead to the future that they did. There were many times when ships went MIA, letters were intercepted and ignored, etc. The takeaway is that it’s often not fair to look back on the past and blame ourselves for missing things that in hindsight seem clear when in the moment, there were tons of potential future outcomes.

Overall

The foundation that the Drunkard’s Walk paints is great. It’s a book I’d recommend to most people to at least skim through, especially if you have an interest in probabilities. I haven’t read the Signal and the Noise by Adam Silver yet (only started it), so I’m looking forward to seeing how it compares.

Thoughts on this review/the book in general? Comment or send me a note :)

Full reading list here