rand() problem

[xenoviaquarta]

int main() 
{
srand(time(NULL));
int bunny_name_index = rand() % 100 + 0;
std::cout << bunny_name_index;
}

first run:
84
second run:
87
third run:
90

why is this happening?
the random numbers should be far from the previous one

[Moschops]

Why should random numbers be far from the previous one? If a random value was affected by a previous random value, it wouldn't be random.

[xenoviaquarta]

so how do i do it?

[Ganado]

No it shouldn't. Having to be far from the previous number would be a less-random requirement. Your sample size is 3, certainly not enough to judge the quality of the randomness.

(rand() isn't perfectly uniform though, but that is not the issue here)

[Moschops]

How do you do what? Generate random numbers? Or do you want to generate numbers that are far away from previous numbers (i.e. not random)?

[xenoviaquarta]

i recorded 10 outputs:

first run:
92
second run:
93
third:
93
fourth:
96
fifth:
98:
sixth:
100:
seventh:
2
eigth:
5
nineth:
9
tenth:
10

[Moschops]

That's nice. Is there a question?

[xenoviaquarta]

the random number i was expecting is like

23
94
57
1
76
42

[Moschops]

If you expected them, they aren't random. You don't understand what random means.

You think it means that each random number will be far away from the previous one. That's not true. That's not what random means. That's not how random numbers work.

It sounds to me like you don't want random numbers, but you want some numbers that are nicely spread out. Is that right?

[Chervil]

If the program is run several times in succession, the seed value won't have changed by much, and (depending on the implementation) the first few random numbers may not vary by much either.

If that's a problem, it's worth adding an extra call (or several) to rand()to generate and discard the first number after a call to srand().

Here, srand() is called with values differing by just 1, the first randoms are not very random. An extra call to rand() seems to help.

#include <iostream>
#include <cstdlib>
#include <ctime>
#include <iomanip>


int main() 
{
    time_t seed = time(NULL);
    
    for (int i=0; i<10; i++)
    {
        srand (seed++);
        std::cout << std::setw(3) << rand() % 100;
    }    
    
    std::cout << "\n\n";
        
    for (int i=0; i<10; i++)
    {
        srand (seed++);
        rand();  // extra call to rand()
        std::cout << std::setw(3) << rand() % 100;
    }
    
}

My output:

 41 44 47 50 54 57 60 63 67 70

 61 42 90 38 19 67 16 96 44 93

Notice the first line of numbers just increase by a small amount in a fairly predictable fashion. The second set of numbers is less predictable.

edit: Note, this is for demonstration purposes only. Normally you would call srand() just once at the start. My program is an artificial example meant to simulate the effect of frequent re-running of a program.

[Moschops]

For completeness, no thread on C++ rand would be complete without a word from the STL himself:

https://channel9.msdn.com/Events/GoingNative/2013/rand-Considered-Harmful