How does the birthday paradox work?β
|
Each coin flip has a 1/2 chance of landing on heads. Since each flip is independent, you multiply the probability of heads (1/2) by itself 10 times. This gives you:
Therefore, the chances of flipping heads 10 times in a row is about 0.1%.
Now let's flip things around: If the chance of 10 heads is 0.1%, the probability of landing on tails at least once in those 10 flips is 99.9%. Keep this in mind because the same principle applies to the birthday paradox.
Returning to the birthday paradox, it states that there's an over 50% chance of a group of 23 people having at least two people sharing the same birthday.
First we need to find the probability of no one in that group of 23 people sharing a birthday. The formula for that is:
To complete the formula and calculate the chances of there being no birthday matches, let's determine these two missing variables.
What's the probability of it not being someone's birthday?
To keep things simple, we'll assume:
That leads to a 364/365 chance of a given day not being someone's birthday. So now we've filled in part of the formula:
How many total pairs are there in the group?
If there's a group of 23 people, how many possible pairings can you make between everyone in the group? You could count this all up manually, but let's make our life easier with some math.
First, take the total number of options you have to make a pair in this group. Since you can choose anyone in the group, you have 23 options. Next, determine the total number of options you have for the second person in this pairing. Since you can't select the same person again, you now have 22 options.
So to find the total number of pairings, you take (23 x 22) and divide by 2. The division is important so you don't double count the pairings. For our purposes, Doug paired with Patti is the same as Patti paired with Doug. (π Hi, Mom and Dad!)
This shows us that there are 253 possible pairings in a group of 23 people. Let's bring it all together now and fill out these missing variables we had in the beginning:
This formula can now be rewritten (and calculated to be):
But remember that this formula is calculating the chances of no one in that group sharing a birthday. Subtract 0.4995 from 1 to find the probability of at least one birthday match. That results in a 50.05% chance of a birthday match being found in a group of 23 people.
The reason why it seems like 23 feels like a shockingly small group to find a birthday match is because it's hard to comprehend how many possible pairings there are in a group (253). Combination calculations to find pairings are simply tough to intuit, so thank goodness we have calculators.
|
--- Thanks for reading this week's newsletter! If you have any thoughts, questions, or favorite GIFs, my inbox is always open. Just hit reply to send me a note! :) All my best, |
βSources for this week's newsletterβ
|
I happened upon this book as I was perusing my local library, and I'm glad I picked it up. It's a delightful book that whisks you around the world to teach you about the origin stories of 80+ words that have made their way into the English language. If you know me well, you know that I love exploring etymology and knowing the backstories of words. It's a great little book for all my word nerds out there.
β
βCheck out the full list of books I've recommended here.
P.S. π€ Do you have a question for the newsletter? Click here to submit your question (or reply to this email!)
|
Learn something new every Friday. Join 2,000+ nerds and satisfy your curiosity one question at a time with topics including science, technology, history, and more.
Longtime readers of Today You Should Know may recognize this question β it's a repeat! I'm working on a newsletter that is causing me to relearn genetics topics I haven't studied since I was 14, so it's taking a bit of time to put together. Side note: It also will include many cat pictures, so I hope it'll be worth the wait. π All my best, π Caitlin Why do paper cuts hurt so much? This question came from a reader submission (thanks Rodrigo!). Has a curious question stumped you lately? Feel...
Is the βfive-second ruleβ legitimate? π The short answer While time is a factor for how many bacteria contaminate dropped food, the five-second rule is not a legitimate food safety guideline. Bacteria transfer almost immediately on contact, and dropping food on wet or flat surfaces increases the chances of contamination. π The long answer You may have heard that if you pick up dropped food within five seconds after it hits the floor, itβs still safe to eat. But is there any truth to this rule...
Why do we drive on parkways and park on driveways? This question came from a reader submission (thanks Aunt Erin!). Has a curious question stumped you lately? Feel free to submit your own question here: Submit your question π The short answer Parkways were named for the parks they passed through, not for parking. Driveways were named for the driving people did on these private roads to reach a barn or a house, before many driveways became too short for driving. π The long answer The English...