The birthday paradox is the surprising result that a group of only 23 people has about a 50% chance of containing at least one shared birthday. It is called a paradox because most people expect the required group size to be much larger. The key idea is that we are not comparing everyone to one chosen person, but comparing every pair of people in the room.
As the number of people grows, the number of pairs grows quickly, making matches much more likely.
Understanding Statistics: The Birthday Paradox
A useful way to calculate this kind of chance is to focus on the opposite event first. Imagine people entering a room one at a time. The first person cannot clash with anyone already there.
Each later person must avoid every birthday already seen if the room is to remain free of matches. That condition becomes harder at each step. The chance of no match is found by multiplying the chances of successfully avoiding the earlier birthdays.
After finding that value, subtract it from one. This complement method is common in statistics because a complicated event can have a much simpler opposite.
The calculation uses a model, not a perfect description of real birth records. It treats each day of a non leap year as equally likely and assumes one person's birthday does not affect another's. Real birthdays are unevenly spread through the calendar.
Some months have more births than others. Twins and siblings can create connections between birthdays, so people in the same family are not fully independent.
These details usually make a match slightly more likely than the simple model predicts. The model is still valuable because it captures the main reason matches become common in ordinary groups.
It helps to separate two different probability questions. One question asks whether anyone in a group matches one particular chosen person. That chance grows fairly slowly because every comparison uses the chosen birthday.
Another question allows any two people to match. There are many more opportunities for this to happen. This distinction appears outside birthday problems.
Computer systems can have two files with the same short identifier. Scientists can find matching genetic markers.
A shop can notice two customers sharing a rare trait. In each case, the result depends on whether one fixed target is being checked or whether every possible pair can produce a match.
Students should pay attention to what counts as a match. The event is satisfied as soon as one pair shares a birthday. It does not require exactly one pair, and it does not require all birthdays to be different except for that pair.
A larger group may contain several separate matching pairs or three or more people born on the same day. These outcomes are included in the overall chance. A simulation can make this clear.
Generate many random groups, record whether each group has any repeated day, then compare the fraction of successful groups with the theoretical result. Simulations vary from run to run, but with many trials they tend to settle near the predicted probability. This is a good reminder that probability describes long run patterns, not a guarantee for one particular classroom.
Key Facts
- Number of pairs in a group of n people: C(n, 2) = n(n - 1)/2
- Probability of at least one shared birthday = 1 - P(no shared birthdays)
- P(no shared birthdays among n people) = (365/365)(364/365)(363/365)...((365 - n + 1)/365)
- For n = 23, P(no shared birthdays) is about 0.493, so P(at least one match) is about 0.507
- The 50% threshold occurs at 23 people, assuming 365 equally likely birthdays and ignoring leap years.
- The chance of at least one match rises rapidly because the number of possible pairs grows quadratically with group size.
Vocabulary
- Birthday paradox
- The surprising probability result that a relatively small group can have a high chance of at least two people sharing a birthday.
- Complement
- The complement of an event is the outcome that the event does not happen, and its probability is 1 minus the event's probability.
- Pair
- A pair is a group of two people whose birthdays can be compared for a match.
- Combination
- A combination counts selections where order does not matter, such as choosing 2 people from a group.
- Independent assumption
- The independent assumption treats each person's birthday as not affecting anyone else's birthday.
Common Mistakes to Avoid
- Comparing everyone only to one specific person, which is wrong because the birthday paradox counts matches between any two people in the group.
- Adding probabilities as if each comparison were independent, which is wrong because birthday comparisons in the same group overlap and affect each other.
- Forgetting to use the complement, which makes the calculation harder because it is simpler to find the chance that all birthdays are different first.
- Assuming 23 people guarantees a shared birthday, which is wrong because 23 people only gives about a 50% chance, not certainty.
Practice Questions
- 1 How many unique pairs of people are there in a group of 23 people? Use C(n, 2) = n(n - 1)/2.
- 2 For a group of 10 people, write the exact expression for the probability that all birthdays are different, assuming 365 equally likely birthdays and no leap years.
- 3 Explain why the birthday paradox becomes likely much faster than many people expect, even though there are 365 possible birthdays.