Topics Numbers and logic

What is the birthday paradox?

The birthday paradox is the surprising result that in a group of only 23 people, the chance that at least two share a birthday is a little over 50 percent. With 70 people, it is almost certain. It is called a paradox because it feels wrong, not because it contradicts anything. The math is solid, and it assumes birthdays are spread evenly across 365 days and ignores leap days.

The trick is in what you count. Most of us picture finding someone who shares our birthday, which is rare. But the question is whether any two people match, and a group of 23 contains 253 different pairs. Each pair is a small chance, and hundreds of small chances add up fast. Real birthdays are not perfectly even, which nudges the odds up slightly.

An episode on this would walk through the pairs, show why it is easier to calculate the chance of no match and subtract it from one, and then follow the same idea into places where it matters, like hash collisions in computing. bre is a podcast app whose hosts are AI, so they can make mistakes, and you can press Talk to ask them to slow down or redo a step.

What a bre episode would cover

An outline of the episode bre would make for this question. Every episode is written fresh when you ask, so yours will differ.

  1. The claim that feels wrongThe hosts state the result plainly: 23 people, about even odds. Then they ask why almost everyone's gut says it should take far more.
  2. Pairs, not peopleCounting every possible pair in a room shows why the number of chances grows much faster than the head count.
  3. Doing the math backwardIt is simpler to find the chance that nobody matches, then subtract from one. The episode walks through that step by step with small numbers.
  4. You versus anyoneWhy the chance that someone matches your birthday is a different, much smaller question, and how mixing them up creates the illusion.
  5. Real birthdays are not evenBirth dates cluster a bit through the year, and leap days exist. The hosts explain how that changes the answer only slightly.
  6. Collisions in the wildThe same logic shows up when computers assign short codes or fingerprints and two different things end up with the same one.

How the episode might open

A sample exchange between two of bre’s AI hosts, bre and Arlo. Both are AI; this is written by AI, as every bre episode is.

  1. breAI host

    Okay, picture a room with 23 strangers. No one planned anything. I say there's about a fifty-fifty chance two of them share a birthday. Does that sound right to you?

  2. ArloAI host

    No. Rubbish. Twenty-three out of 365? Tiny.

  3. breAI host

    That's what everyone says, and it's the honest reaction. Your brain is checking whether anyone matches you. That's one person against twenty-two others.

  4. ArloAI host

    Yeah. So?

  5. breAI host

    So the question isn't about you. It's whether any two people in the room match. Every person gets compared with every other person.

  6. ArloAI host

    How many comparisons is that?

  7. breAI host

    Two hundred and fifty-three pairs. Not 23. Each pair has a small chance, but there are a lot of them.

  8. ArloAI host

    Right. Lots of small shots. Fair enough, show me the sums.

Questions people also ask

Is the birthday paradox really a paradox?
Not in the logical sense. Nothing contradicts itself. It is called a paradox because the answer clashes with intuition. The math is straightforward once you count pairs of people instead of single people, and it is easy to check with a simple calculation.
Why does it take only 23 people?
With 23 people there are 253 possible pairs, and each pair has about a 1 in 365 chance of matching. Those chances combine. Working out the chance that nobody matches and subtracting from one gives just over 50 percent.
Does the birthday paradox work with real birthdays?
Roughly, yes. Real birthdays are not perfectly evenly spread, and leap days add a wrinkle. Uneven birthdays make a match slightly more likely, not less, so the 23-person figure is a good approximation, not an exact rule for every group.
Where is the birthday paradox used?
It appears in computer science and security, where it explains why two different inputs can produce the same hash value sooner than expected. This is called a collision, and designers use birthday-style math to choose how long codes need to be.

Related topics

More: all 300 topics, numbers and logic, or the longer reads on /learn.

bre’s hosts are AI, and every episode is generated, so they can be wrong: check anything that matters. This page outlines what an episode would cover. It is for interest and learning, not medical, financial or legal advice.