How Many People Until Two Share a Birthday?
50.7%
chance that at least two people share a birthday
Possible pairs to compare: 253
More likely than not — the odds favor a match!
🎯 A Simple Example
Before reading further — guess: how many people does it take for a 50% chance of a shared birthday?
1️⃣ Set the slider to your guess and check the probability. Were you close?
2️⃣ Now set it to 23. That's the magic number — just 23 people for a coin-flip chance!
3️⃣ Hit Simulate a few times at 23 to see it happen (or not). About half the time, you'll get a match.
4️⃣ Try 50 people — the probability jumps to 97%. Run simulations and watch matches appear almost every time.
Classroom Tip: In a class of 30 students, there's a 70.6% chance two share a birthday. Ask your students to test it — it's a fantastic live demonstration of probability!
Data Source: Von Mises Birthday Problem • Public domain • Solo-developed with AI
The Surprise: Ask most people: "How many people do you need in a room before there's a 50% chance two share a birthday?" Most guess around 183—half of 365. The real answer? Just 23. This is the Birthday Paradox, and it reveals a deep flaw in how our brains estimate probability.
Why We Get It Wrong: The trick is that the question isn't "will someone share YOUR birthday"—it's "will ANY two people match." With 23 people, there are 253 possible pairs to compare (23 × 22 ÷ 2). Each pair has a small chance of matching, but 253 small chances add up fast. By 70 people, you have 2,415 pairs and the probability exceeds 99.9%.
Real-World Impact: This isn't just a party trick. Cryptographers use the birthday paradox to estimate how quickly a hashing algorithm can produce collisions. The "birthday attack" is a real technique in computer security—it's why modern hash functions need such large output sizes. A 128-bit hash can be broken in roughly 264 operations, not 2128, because of this exact mathematical principle.
DNA Databases and False Matches: The birthday paradox has serious implications for forensic science. A national DNA database with 100,000 profiles contains nearly 5 billion possible pairs. Even if each pair has a 1-in-a-billion chance of a coincidental match, the expected number of false matches in the database is around 5. This is why forensic statisticians never say a DNA match "proves" identity — they report the probability of a match. The 2004 Brandon Mayfield case, where the FBI wrongly identified a lawyer as a Madrid bombing suspect based on a "fingerprint match," became a landmark warning about coincidental matches in large databases.
Why Hash Functions Need to Be Twice as Long: Cryptographers use the birthday paradox to size secure hash functions. If you want a collision to require 2256 operations, you need a 512-bit hash — not 256-bit — because a birthday attack only needs √(2512) = 2256 attempts. SHA-256 is safe against collision attacks because finding two inputs that produce the same hash requires roughly 2128 work. Same math, vastly different stakes.
The History: The problem was first formalized by mathematician Richard von Mises in 1939. But it gained fame through the work of Harold Davenport, who used it as a teaching tool to show how combinatorics defies gut instinct. Today it remains one of the most popular "aha!" moments in introductory probability courses.
🔬 Explore the Probability Lab
🐾 From the Lab Cat's Combinatorial Nap Research Unit:
I have observed this "paradox" in action. My conclusions:
Conclusion: The true paradox is why humans only celebrate their existence once a year. Very inefficient. 🐈