veritasium
patreon
Patreon Exclusive: Markov Chains - Footage from the public that didn't make the cut
🕑 Added 2025-08-19 14:56:58 +0000 UTCFor our recent video on the topic of Markov Chains we wrote and filmed a segment where people had to guess how many jellybeans were in a jar. This involved producer Casper counting an amount of beans into the jar and then taking them out on the street to gather a set of both independent and dependent guesses. The point being that 'wisdom of the crowd' works best when people guess independently of each other.
This went as we expected, however, when we coded a website and put this question to the folks at home, we got a very different set of results...
Thanks to all those who took part in this and for more about Markov Chains, check out the full video on our YouTube channel - https://youtu.be/KZeIEiBrT_w
Comments
Karina C
The next question is: how many people do you have to ask to get the precise answer? Practically (which was the girl who guessed) and theoretically (is there some way to predict this)
Mohammad Nazzal
Taking into account that people might guess repetitively.
Mohammad Nazzal
Veritasium is on dropline. Real must be real :) Being said on average that the jar would have between 500-3000 and you want to make people to guess it. The chances of people would guess it correct has a very low percentage around 1.67% with 10,000 surveys. Assuming you only surveyed around 100 in-person that chance would drop to 0.0067%. and it's nearly impossible unless she's an Einstein but i doubt that lol.
