It is oft-stated that most people really do not get statistics. Just say that word, “statistics,” and most people’s eyes glaze over. Confession from this engineer and MBA: I did horrifically in my undergrad stats course. (Fortunately, I did better in my MBA course, thanks to a big dose of good teaching from Bob Winkler, and a small dose of being a decade more mature).
What does all of this have to do with Coca-Cola?
It turns out Warren Buffett, the 84-year-old über-investor (as in the German word “über”, not the ride-sharing company Uber), has 5 cans of Coke per day! His first is at breakfast, and the rest goes on from there.
Why, exactly, would he do that, other than the fact that he may really like (or be addicted to) the soft drink?
According to the article, Buffett said:
“I checked the actuarial tables, and the lowest death rate is among six-year-olds. So I decided to eat like a six-year-old. It’s the safest course I can take.”
When I read this, I actually laughed out loud. Buffett must either have been joking, or having fun at the reporter’s expense.
The mistake here, of course, is confusing correlation with causation. Just because two things happen together with each other (they correlate) does not necessarily imply that one causes the other. There are 4 possibilities:
- The first may cause the second;
- The second may cause the first;
- Both may be caused by some external third factor;
- The two may be completely unrelated and happen to occur together.
But seeing the two together is never enough to assume that one causes the other. Buffett is more than smart enough to know this (he is a lot smarter than me), which leads me to believe he was having fun at the reporter’s expense.
What is wrong with what Buffett said? Well, the lowest death rate is among six-year-olds, and six-year-olds consume a lot of Coke, to quote Monty Python, “therefore, logically“, Coke causes six-year-olds to have low death rates, and if only 80-year-olds drank Coke, they would live longer.
The scene from Monty Python’s “Arthur and the Holy Grail” is a classic and very funny example of confusing cause and effect, and is a worthwhile watch.
By that logic, six year olds also have no facial hair, so 80-year-olds should shave their beards to live longer. For that matter, six-year-olds wear size 1 shoes, so all 80-year-olds should wear too-small shoes to ward off death.
Of course, it is not eating like a 6-year-old that causes 6-year-olds to have a low death rate; it is being a six-year-old in a six-year-old’s body that causes them to have a low death rate. There probably is a combination, as well, of low stress (no need to earn a living), minimal dangerous situations, along with a host of other factors that come from being, well, 6.
Years ago, a very prominent management consulting firm recruited me. About half the people I met during the process were brilliant; I learned an enormous amount from them. The other half were fools who knew how to look smart. Unfortunately, the senior partner who interviewed me was the second type.
We were reviewing financial statements from a case study. He was looking for me to connect one number as an indicator of another. I believe that the loss of a company was, say, $14.2MM, and there was an issue with some other part of the company that had $14.2MM in some asset.
Now, those two could be related – just like six-year-olds and Coke could be related – but they are not necessarily related.
I did what any good businessperson (and scientist) should do: I looked for the connection. There was none; it was a coincidence. With enough numbers, coincidences are bound to happen.
It is the equivalent of seeing that a division has a $10,500 loss, and revenue from all the customers in New Zealand is equal to $10,500. The numbers indicate that it is worth a look, but chances are pretty good that there is no connection.
The partner absolutely insisted that the number connection – $14.2MM here and $14.2MM there – was crucial. I knew it was a red herring… and that this partner not only did not know the difference between correlation and causation, but was not mature enough to recognize his own mistakes and learn from them.
Needless to say, I didn’t want to work for them. After I didn’t just let him have his incorrect win at the expense of the client, I doubt he wanted me to work for them either.
Does Buffett drink lots of Coke? If he said so, I am sure he does. Does he do it because six-year-olds do? I suspect not. Does he know the difference between correlation and causation, and how to read numbers? I am quite certain he does.
Was he having fun at the reporter’s and readers’ expense? That would be a fair guess.
Knowing how to read your numbers, when two issues are connected or are causal and when they just look similar, is crucial to knowing what your business is doing and what the future has in store. Even the smartest can get led astray by seeming connections, especially when we are emotionally connected to those numbers.
Do you want to know what they portend? Ask us to help.