Bacle
Jul21-11, 07:51 AM
Hi, All:
The standard way of approaching the birthday problem, i.e., the problem of
determining the number of people needed to have a certain probability that
two of them have the same birthday, is based on the assumption that birthdays
are uniformly-distributed, i.e., that the probability of someone having a birthday
on a given day is 1/365 for non-leap, or 1/366 for leap.
But there is data to suggest that this assumption does not hold; specifically,
this assumption failed a chi-square at the 95% for expected-actual, for n=480,040
data points.
Does anyone know of a solution that uses a more realistic distribution of birthdates?
The standard way of approaching the birthday problem, i.e., the problem of
determining the number of people needed to have a certain probability that
two of them have the same birthday, is based on the assumption that birthdays
are uniformly-distributed, i.e., that the probability of someone having a birthday
on a given day is 1/365 for non-leap, or 1/366 for leap.
But there is data to suggest that this assumption does not hold; specifically,
this assumption failed a chi-square at the 95% for expected-actual, for n=480,040
data points.
Does anyone know of a solution that uses a more realistic distribution of birthdates?