Benford’s law

I’m going to start tonight’s post with a simple game. If you can guess how many Romanian lei are in my bank account, I’ll give you all of it. (In reality I won’t. I can’t. But just imagine for a second.) Actually, to make it easier, you don’t have to guess the full amount, but just the first digit. Given that there is some money in there, i.e. it’s not zero, what digit (from 1 to 9) do you choose? Maybe you’ll choose 5, because it’s in the middle. Perhaps you’ll pick 7, your favourite number. It really doesn’t matter, because they’re all just as likely as each other, right?

Wrong. With absolutely no other knowledge, you should absolutely pick 1. Always. I can guarantee that around the world, there are currently more bank balances with 1 as their first digit than any other. There are quite a few 2s too, but certainly not as many as 1s. A little further behind are 3s, and so on. Comparatively few begin with an 8, and even fewer with a 9.

You see this pattern over and over again. Population sizes (or areas) of towns, cities, or whole countries, lengths of rivers, volumes of lakes, market caps of companies, speakers of languages, vote counts, and so on. Anything that can exhibit a large range of values will show a very similar distribution of leading digits, skewed heavily towards the lower digits. (Not everything works this way, for instance the ages or heights of the people you work with.)

There’s a name for this phenomenon—it’s called Benford’s law—and it has even been used in forensic accounting. People have completely fabricated figures and come unstuck because what they thought were random-looking numbers started with unnaturally many 7s, 8s and 9s. The coronavirus pandemic—where cases increase exponentially—is a perfect example of Benford’s law in action. Look at the case figures for countries or regions and you’ll see a definite bias towards lower initial digits. Heck, you can even see it on my chart below.

As I was writing all of that, I got a phone call from the elderly lady who lives with her husband on the sixth floor. It was a pleasure to talk to her. She was replying to the note I’d sent her, asking if they needed any help with food or other essentials. She said they have people staying from outside the city (breaking the lockdown rules, but hey). Her prediction for the end of lockdown—1st May—is unfortunately wishful thinking. She asked me what my religion was, because it affects what version of Easter I celebrate. I get that question quite often and I always say that I’m a Catholic.

My aunt called me earlier today. She seemed extremely concerned, as if Romania had 4300 coronavirus deaths and the UK 150, instead of the other way round. It’s never that easy talking to her, because she doesn’t really listen. Still, she was thinking about me, and I appreciated that. She also lives on her own, so I’d better call her next weekend.

Here is the latest chart:

Coronavirus Romania 4-4-20

Leave a Reply

Your email address will not be published. Required fields are marked *