I am guessing that you are working some kind of dice game. The "built-in" random number generators in Perl, like in C and other languages aren't really that random. They are fast and good enough for lots and lots of things. There are other algorithms that are slower, but better - better meaning more like true randomness. You will have to do some investigation about this if you are really going to get serious about modeling some betting strategy or whatever.

So you have actually asked what can be a very complicated question. Here is one link for you: randomness

I made a little program to satisfy your requirement and then added a little histogram and a way to loop multiple times (now 10 times). Program rolls dice until it sees a 6, then stops. Along the way to the 6, it keeps track of other numbers rolled. The int(rand(max+1)) gives out numbers from 0..max.

#!/usr/bin/perl -w
use strict;

my %histo;
for (1..10)
{  
   my $count=0;      
   while ( (my $r = int(rand(6)) +1 ) != 6) 
   {                #int(rand(6) generates nums 0-5
     $count++;
     $histo{$r}++;
   }
 #update  ####  $count++;  # there is an off by one error below
   print "took $count tries to get a 6\n";
   $histo{6}++;
}
foreach my $side (sort keys %histo)
{
   print "number $side rolled $histo{$side} times\n";
}

__END__
took 18 tries to get a 6
took 5 tries to get a 6
took 9 tries to get a 6
took 1 tries to get a 6
took 3 tries to get a 6
took 1 tries to get a 6
took 1 tries to get a 6
took 0 tries to get a 6
took 15 tries to get a 6
took 13 tries to get a 6

number 1 rolled 15 times
number 2 rolled 17 times
number 3 rolled 11 times
number 4 rolled 15 times
number 5 rolled 8 times
number 6 rolled 10 times
[download]

Only this part is needed for the original question:

my $count=0;      
while ( (my $r = int(rand(6)) +1 ) != 6) 
{                #int(rand(6) generates nums 0-5
     $count++;
}
print "took ++$count tries to get a 6\n";
[download]

Update: of course there is an off by one error in the loop above, $count doesn't get incremented when loop condition is satisfied, so all try counts should be +1. I took that into account for the histo, but not for the simple count.