Given a system where users register quit often. When a user registers he gets an ID. Two users that register one after another will have sequential IDs. If I'll divide them into 2 groups, one for even IDs and the other for odd IDs, can I assume that there is no statistical connection between users in the same groups? In other words, is dividing in that way is equal to truly random classification?
Can you count on a oddness of a ID number to be random?
1
$\begingroup$
probability
statistics
-
0It might be helpful if you provide more context. How many users? How many nearly simultaneous assignments? How long do IDs last? What's bad if users have sequential IDs? How bad is it? There are surely better schemes - how much programming are you willing to do? This may be a better question for stackoverflow. – 2017-01-31
-
0It's not a practical question. A good solution would be to do hash function and then divide into even and odd groups.I'm asking if there is a case when it won't be good classification. – 2017-01-31