Timeline for birthday problem - expected number of collisions
Current License: CC BY-SA 3.0
11 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Sep 12, 2023 at 10:19 | comment | added | Olivier Lalonde | Anyone is able to solve for N? e.g. given n people and an expected number of people who share the same birthday, what is N (number of days in year). It could be useful to calculate the expected number of random numbers generated per collision. | |
Aug 17, 2021 at 10:32 | comment | added | Henry | @JordanBrough The $23$ in the birthday problem (with $N=365$) is the median number of people when the first duplicate occurs. This rather different question is about the expected number of people who share a birthday with someone else, so not about the median and not about the first duplicate. The distribution of the number of people with duplicate birthdays is right skewed (you could have a large number with low probability - note it cannot it $1$ since at least $2$ must share), so you should not think there will be a simple relationship between the answers to different questions. | |
Aug 16, 2021 at 16:40 | comment | added | Jordan Brough | Shouldn't expected number of people be less than 1 when n=22 and N=366, since the probability of a birthday collision only exceeds 50% when n=23 for the birthday problem? Yet I get ~1.23 for n=22 and N=366 when using this formula — wolframalpha.com/input/… | |
Mar 16, 2020 at 3:27 | comment | added | zyxue | @Henry, do you know if there is a derivation for the variance of the number of collisions in a hashing context (not the kind of collision as described in this birthday context) | |
Mar 11, 2020 at 22:35 | comment | added | Henry | @zyxue Another question asked that | |
Mar 11, 2020 at 21:23 | comment | added | zyxue | what would be the variance of the number of people that share brithdays with somebody? | |
May 6, 2011 at 11:38 | comment | added | brannerchinese | I would like to be able to cite your help in the paper I am writing (about philology, not birthdays). Would you mind to look me up at brannerchinese.com and contact me off-list? There is no regular private-messaging function on the SE site (meta.math.stackexchange.com/q/632/9263) and I can see no other non-public means to ask you for a name by which I can acknowledge your help. I understand if you prefer to remain anonymous or "Henry". | |
May 5, 2011 at 18:05 | comment | added | brannerchinese | I wrote a simulation and ran several million trials using various N and n; the results are within .001n of what your formula predicts. Thanks again. | |
May 5, 2011 at 18:03 | vote | accept | brannerchinese | ||
Apr 29, 2011 at 13:15 | comment | added | brannerchinese | Beautiful in its clarity. Thank you. | |
Apr 29, 2011 at 9:09 | history | answered | Henry | CC BY-SA 3.0 |