Random #

[L E G I O N]

hey guys i need to make a random # generator... and im having a problem... it keeps telling me that i need to return something.... but i've tried a bunch of dif. things... and im comin up short....

#include <iostream>
#include <string>
#include <cstdlib>
using namespace std;
 
int random (int min, int max) 
	{
    
 cout <<  rand() % 10 + 1 << endl;

		return;
}
int min = 1;
int max = 10;

void main() 
{ 

	cout << random (1,10) << endl;

	
}

this is the error its givin me

Error 1 error C2561: 'random' : function must return a value c:\users\www\documents\visual studio 2008\projects\rand0mz\rand0mz\randy numberz.cpp 11 Rand0mz

[Disch]

1) There's no point to those min / max global variable on line 13, 14

2) You don't srand at the start of your program.

3) Your 'random' function ignores the passed min/max parameters. It will always return [1..10] no matter what min/max you specify.

4) random should not be outputting to cout. Instead it should return a random number.

5) main should return an int, it should not be void.

[Bazzy]

int random (int min, int max)  // notice: not void
	{
    
 cout <<  rand() % 10 + 1 << endl;

		return; // what are you returning here?
}

[L E G I O N]

idk... i got it working like this....

#include <iostream>
#include <string>
#include <cstdlib>
#include <ctime>
using namespace std;
 
int random (int min, int max) 
	{
	    srand(time(NULL));
			return  rand() % max + min;
}
		int min = 1;
			int max = 10;

void main() 
{ 

			cout << random (1,10) << endl;

			return;
}

[jsmith]

Bzzzt.

Call srand() exactly once in your program. You are calling it every time random() is called.

Lines 12 and 13 are pointless.

Line 10 is wrong. Consider if min = 5 and max = 10. rand() % 10 yields [ 0... 9 ], plus 5
yields 5 ... 14.

[Fsmv]

It should be changed to:

#include <iostream>
#include <string>
#include <cstdlib>
#include <ctime>

using namespace std;

srand(time(NULL)); 

int random (int min, int max) {
	return  rand() % (max + min);
}

void main() { 
	cout << random (1,10) << endl;
}

[R0mai]

@Fsmv : still bad

//Returns a random number in the range [min, max)
int random (int min, int max) { 
    return rand()%(max - min) + min;
}

[Disch]

+ you can't have srand outside of a function like that

+ main should not be void

[tummychow]

Just call srand as the first statement in main(). (which should be int not void)
I'm looking forward to the random generation changes in 0x, I'll say that much.

[imi]

Dunno whether you care that much (many people don't care), but rand() % max is not really random for almost all values of max. To be precise: for all max except when they are a power of two (max = 1,2,4,8,16...)

This is because rand returns a random from 0 to RAND_MAX (which is usually 32767), so first of all it will never return anything bigger than RAND_MAX. But even for smaller values, you get some numbers much more often then other numbers. This problem is worse, the bigger "max" is.

Easiest to get a grasp of "why???" is: What if RAND_MAX would be 2 and you do rand()%2, then you get:

rand() returns 0 -> 0
rand() returns 1 -> 1
rand() returns 2 -> 0

so you get 0 twice as much as 1. Not particular random...


This a nice piece of code to try it out for yourself:

int main()
{
	int count = 0;
        int max = RAND_MAX/2;
        int number_of_tests = 10000000;

	for (int i = 0; i < number_of_tests; ++i)
		if ((rand() % max) == 0)
			count++;
	cout << "Actual: " << count << endl;
	cout << "Expected: " << (number_of_tests/max) << endl;
}

You count the numbers of zeroes. Should be one out of "max" times, right? Wrong! You'll be surprised.. ;-).


What is better? Well, I use this:

int better_rand(int max)
{
	int r,x = 1;
	while (x < max) x <<= 1;
	while ((r = rand() % x) > max);
	return r;
}

not the fastest, but doesn't have the issues with "rand() % max". If you want industrial strength random stuff in your game, forget about std::rand() at all...

And upcoming standard library will fix these problems anyway, by providing a fully fledged random number generator system.. ;-)

Ciao, Imi.

Edit: Just remembered: Beware, my better_rand returns values from 0 to max includive, whereas rand() % max does never result in "max". To get the same as rand()%max, change the second while to "...>= max..."

[jsmith]

Output from my machine:

Actual: 0
Expected: 0

because RAND_MAX = 2^32 - 1 nowadays. 32767 comes from the 16-bit DOS days.

[jsmith]

Nonetheless there is a slight bias involved in the formula:

rand() % ( max - min ) + min

which depends upon the values of min and max (actually
upon max - min).

To understand, let's say rand() only returned numbers between 0 and 9 inclusive.
Let's say max = 7 and min = 3.

rand()          result
   0               3
   1               4
   2               5
   3               6
   4               3
   5               4
   6               5
   7               6
   8               3
   9               4

P(3) = 3/10
P(4) = 3/10
P(5) = 2/10
P(6) = 2/10

[imi]

because RAND_MAX = 2^32 - 1 nowadays. 32767 comes from the 16-bit DOS days.

*sniff*.. I am still living in good old 16-bit times.. ;-)

No, fun aside. RAND_MAX is 32767 on my latest VC.

So I conclude that my worrying is only really interesting if you are either using Windows and VC or really big max-values...

Ciao, Imi.

[jsmith]

Sorry, typo on my part.

RAND_MAX = 2^31 - 1 (it is signed).