The Wisdom of Crowds – a 1-week long experiment done by me

“By three methods we may learn wisdom: First, by reflection, which is noblest; Second, by imitation, which is easiest; and third by experience, which is the bitterest.” -Confucius

Ah, people. Billions of us have existed and communicated with each other for thousands and thousands of years. And arguably the most important parts of our survival as a species: knowledge, wisdom, and communication. Without those, we would be nowhere.

These ideas are the basis of the famous social concept, The Wisdom of Crowds. Developed by James Surowiecki, it states that if given a large quantity for people to guess the exact amount of, the average of all their guesses will be more accurate than any one individual guess.

It sounds crazy, but it actually works. A BBC documentary as well as a few TEDx videos have been done about this theory. The documentary even demonstrates the theory, when a professor tells students to guess the number of jellybeans in a jar. The average of all their guesses was less than 1% off from the actual number of jellybeans. Additionally, this was done in the early 20th century, where Sir Francis Galton was able to accurately guess the weight of a bull by averaging the guesses from 800 villagers.

Seeing this concept for the first time, it really intrigued me, and naturally, it made me want to try it myself. So that’s exactly what I did. Here is the whole entire process of me replicating this experiment, and analyzing the results!

(bottle of skittles)

Step 1: Give them something to guess

The obvious first step in this experiment would be to get something for them to guess, something with a large quantity of a familiar small item in a certain container. Will it be a jar of jellybeans? A bucket of sunflower seeds? This is what you have to choose during this step.

My choice was a bottle of skittles, as you can see above. Small, compact, and something that many people see regularly, and are familiar with the size of. This allows for more educated guesses, which is good for the experiment, really. And since I planned to give the entire bottle to the winner, I thought it to be a good idea to include a candy, to emphasize the reward aspect of it.

Step 2: Count your objects

This step is very crucial. It’s very important for the sake of this experiment that you be as accurate as you possibly can. This is to ensure that the experiment is consistent and orderly, and to maximize the accuracy of the guesses.

(skittles counting paper)

This is the paper I used to count the skittles, and as you can see I had a pretty good system going. I counted 10 skittles at a time, and then after I put them in the bottle, I put down 1 dot to represent it. After that I counted 10 more, put them in the bottle, marked it down, and then kept doing that until I counted all of the skittles in the bag. If I got to the end of the bag and there were less than 10, then I’d put a dash for each 1 skittle instead. I repeated that process for each of the 5 big bags of skittles, and separated each bag with a line, for the sake of simplicity.

(Skittles 3)


1) The bottle in which the skittles would be placed in. I removed the label so as to make it easier to see the inside.

2) The entrance to the bottle. My original plan was to drop them in the funnel which would then go into the bottle, but the opening was too small, so I just washed my hands and counted them out by hand, very carefully.

3) The bowl of skittles that I was counting. As it turns out, it was just big enough to hold one of the big bags, so I knew that when the bowl was empty that I had to start a new row.

4) The big bags of skittles that I used for the experiment. Much more practical than using the small pocket-sized bags.

5) I put 10 skittles at a time on this napkin. I also double checked to make sure that there were exactly 10 as I put them inside the bottle. It was very methodical, really.

6) The paper where I put all of the counted skittles. It was full by the end of the experiment.


The whole process took me about 3 hours to complete. Ten by ten, counting out skittles and putting them in the bottle. And after all of that work and precision, it finally brought me to my grand total of 1,868 Skittles.

By then, I had already done so much work on it that there was no way I was going to mess it up. I proceeded to go and collect as many different guesses as possible, trying to get a large and diverse sample size. For the 1 week duration of the experiment, I learned so many interesting things, and I really enjoyed the whole process.

For the sake of brevity, I’m keeping this post under 1,000 words. Look out for the post detailing the next phases of the experiment in the coming days, though; I’m looking forward to showing it to you guys. A lot of pictures were taken as well, so I will make a photo gallery that has all of the photos I took. Keep checking back to see when the next post has been made!

Leave a Reply