Between a BBC experiment and and anecdote from Francis Galton from a 1906 county fair, there is definitely a possibility that the wisdom of the crowd works well when it comes to making guesses.
First, the county fair. An ox was on display and was the subject of a guessing contest. Whoever got closest to guessing the weight of the ox would get the slaughtered meat, a tremendous prize worth trying for today, let alone 100 years ago. While nobody got the weight right, when the guesses were averaged, they nailed it.
When I watched the video below, my first thought was that had the silly person who guessed 50,000 jelly beans (after correcting from her original guess of 80,000) not been so far off, that the wisdom of the crowd experiment would not have been so close. Then I started wondering about the ways of the universe, quantum physics, spiritual components, and human nature, and I wondered, “do really bad guesses make it possible for the wisdom of the crowds to work?”
In other words, are there stupid people cancelling each other out on such things? What does this say about the balance of things in the universe?
Before you get annoyed with me calling others stupid, keep in mind that my guess when seeing the jelly bean jar was 1100, way off from the 4,510, so I qualify to be in the realm of the dumb. Here’s the video. Do you think that stupid people cancel each other out in such cases?
advocatus diaboli says
Of course the accuracy of this cancellation effect is dependent of the probabilities and occurrences of under- and over-estimating being very close to equal—in other words symmetrically distributed about the mean. I believe this is not the case very often due to human emotional biases. I also wonder why a guy “techie” enough to have an iphone and use the calculator, doesn’t have a tablet or laptop with a spreadsheet to enter the data in on the fly, easily total up the entries without error, and allow him to sort and remove outliers to determine their effect on the total. Was the data in the form of a normal curve or a more skewed distribution (I think most cases are the latter when flawed, biased human decision-making is involved). From one nerd to another Sal 😉