Picking at random
On 24 December 2016 Jigsaw Man (@Jigsawman2014) organized an elephant race on Twitter. The concept was pretty simple: all you had to do was to pick one in four absolutely identical elephants. With such a setting one would assume that each elephant — let’s call them A, B, C and D — would have gathered about one fourth of the 8,702 votes. Well no, they didn’t. Here are the results:
A few days later, on 27 December, Jörg Lohrer (@empeiria) made his own race. A snail race. No less than 13 457 people took the survey. Once again, you would expect the votes to be uniformly distributed between snail A, B, C and D but, once gain, this is not what happened:
Here are the stylized facts: A and D typically get 10–20% of votes each, 25–30% of people would pick B and the overwhelming majority of people (40–45%) choose C.
So there’s obviously something happening here. As a matter of fact, people are not picking these options at random. At first, it is safe to say they dismiss extremes options then, it looks like they favor odd numbers relative to even numbers — maybe an even number sounds like a “too obvious” choice, I don’t know. The fact is, in such a four-choices experiment, a significant majority of people would pick the third option.
Interestingly, FXL (@Effixel) sent me a video made by Brady Haran (@numberphile / @BradyHaran) where a guy named James explains that, when asked to choose a number between 1 and 10 at random, most people would pick 7. Unfortunately, I can’t test it on Twitter but I made a small Google form. We’ll see (help me to get more answers by circulating the link!).
Oh and by the way, I also made a unicorn race but don’t answer since you know the trick.