>>>>>>>That's right. What the question actually asks is for the probability of any two people in the group sharing their birthdate on any day of the year. That's the reason the probability is higher than most people expect.
>>>>>>
>>>>>>I'm somewhat probability-challanged so what would it be if you said, "What is the probability that out of the 24 people in this room that one of them has my birthday?"
>>>>>
>>>>>The probability that at least one will have the same birthday is around 6.5% (using same basis, calculate probabilty that none have the same birthday & subtract from 1)
>>>>>
>>>>>1 - (364/365) ^ 24
>>>>>
>>>>>if you're one of the 24, sum is 1 - (364/365) ^ 23 & result is about 6.2%
>>>>
>>>>Probability of only 1 person having the same birthday is around 0.016% calculated as (364 ^ 23) / (365 ^ 24)
>>>
>>>Where do you get these numbers from?
>>>
>>>Adapting the previous formula, you get:
>>>
>>>1 - (364/365) ^ 1 = 0,27%.
>>>
>>>Which is exactly the same as 1/365.
>>>
>>>Hilmar.
>>
>>Combination of the probability of 1 person having the same birthday (1/365) and the probability of the remaining 23 having a different birthday ((364/365)^23).
>>
>>(Actually, I've got the feeling I'm not totally right, but I can't work out why, but I think it's in the right area).
>
>Len;
>
>You might approach a correct answer but it will never be absolute. The numbers used indicate a smoothed average and not the number of births, which differ each day, throughout the year and over any number of years. That just makes things more fun! :)


I know, I also haven't taken into account leap years which makes the calculations somewhat harder too.

>
>I guess we have to qualify the answer to some extent and then do not forget to include a factor of “uncertainty”, as we term it in Metrology.
>
>To obtain a more accurate answer simply ask each of the 24 people in the room for the day and year they were born. Then go to your database (surely we must have a database) which will compare date entries and lookup the results for the number of births on a specific day, plug in the proper algorithm and then define accurately (within limits) the answer. Simple – however seems like it might take away some of the fun! :)
>
>By the way I had to measure things down to a micron when I was a Metrologist. Statistical Analysis is a way of life for Metrology.

Sounds fun, I've always enjoyed playing with numbers.

BTW, see a later reply to Hilmar, I worked out where I was going wrong in the calculation, I made the mistake of initially testing a 'trivial' case (2 people) & applied the result to the more complex case without thought (it's been a long day, at least thats my excuse).