Why Does a Group of 23 Make a Shared Birthday More Likely Than Not?

In a room of 23 people, the standard probability model gives about a 50.7 per cent chance that at least two share a birthday. This feels surprising because we often substitute a different question: will anyone share my birthday? The actual problem singles out nobody. It compares every possible pair in the room, and 23 people create 253 pairs, so opportunities accumulate quickly.

Calculating “at least one match” directly is awkward, so it is easier to calculate the complement: every birthday is different. The first person may have any birthday. The second avoids it with probability 364/365; the third must avoid two occupied dates, giving 363/365; the product continues down to 343/365 for the twenty-third person. The result is about 0.493. Subtracting this from 1 gives about 0.507 for at least one shared birthday.

This is not a logical paradox but a finite probability calculation under stated assumptions. The standard model uses 365 days, ignores 29 February, and treats birthdays as independent and uniformly distributed. Real birthdays are not perfectly uniform, so empirical values may shift slightly. The mechanism remains: as the number of people grows, the number of possible pairs grows roughly with the square of the group size. Intuition counts people; the calculation must count relationships between them.

https://www.chiefscientist.gov.au/2011/05/the-birthday-problem
https://www.amstat.org/asa/files/pdfs/stew/HappyBirthdaytoTwo.pdf