There was another thread at Biased random number selection which has information on how to optimize the selection process for speed.

But here's the most straightforward (but not necessarily fastest... and not tested) solution:

# A list of weights; initialized for your problem
my @weights = (8, 1, 1);

# Count the total of all the counts
my $sum;
map { $sum += $_; } @weights;

# Choose a random number, then find a weighted random result
my $target = rand($sum);
for (1..@weights) {
   $target -= $weights[$_];
   last if $target <= 0;
}

print "The number is $_\n";
[download]

If you aren't worried about space, then that approach is fine enough, but perhaps:

@a = split /\s+/, join " ",
  ("1 " x 8, "2 " x 1, "3 " x 1);
print $a[int(rand 10)];
[download]

However, I would not approach it this way. Instead, I would do something like:

my %wheel = {"One" => 8, "Two" => 1, "Three" => 1};
my $sum;
my $item;
foreach $item (keys %wheel) {
  $sum+=$wheel{$item};
}
my $spin = int(rand($sum));
foreach $item (keys %wheel) {
  last if $sum < $wheel{$item};
  $sum-=$wheel{$item};
}
print "$item was picked\n";
[download]

G'Luck!
Gryn

This idea rocks. But I think this is correct

$random = 2 + ( int( rand( 10 ) ) <=> 8 );
[download]


this will make
$random = 1 ...... 80% of the time
$random = 2 ...... 10% of the time
$random = 3 ...... 10% of the time

mmmm...ternary

  ($x) = map { $_<1 ? 2 : $_<2 ? 3 : 1}  rand 9;
[download]

my $seed=rand;
my $result=($seed>=0) + ($seed>=.8) + ($seed>=.9);
print "$result\n";
[download]

use strict;

my %odds = ( 1 => 80, 2 => 10, 3 => 10 );

for my $i (1..25) {
    print "$i: ", &go, "\n";
}

sub go {
    my $i = int(rand(100) + 1);

    foreach my $k (keys %odds) {
        $i -= $odds{$k};
        return $k if $i <= 0;
    }

    return undef;
}
[download]

$randnum = int(rand(100));

if ( $randnum > 20 ) #80%
{
    #stuff for 1
}
if ( $randnum > 90 ) #10%
{
    #stuff for 2 and 3
}
[download]

I don't want to be picky, but that would do #stuff for 1 70% of the time.

It would do #stuff for 1 _and_ #stuff for 2 and 3 10% of the time.

And 20% of the time it would do nothing...

Remember to think before you code! :-P

cmburns

You probably missed that your question was put at How do I get random numbers that follow standard distribution? in Categorized Questions and Answers.

There are several answers there, including mine at Re: How do I get random numbers that follow standard distribution?.

UPDATE
After talking with you I found that you have two rather irritating conditions. You cannot easily install modules, and you want to skew the output. Here is some sample code that gives you an idea how to do that, but I am warning you that it will take a lot of playing around to get this to produce what you want...

my @weights = qw(0.5 1 2);
my @random = sort {$a <=> $b} map{rand()} 0..$#weights;
my $tot;
foreach (0..$#weights) {
    $tot += $random[$_] * $weights[$_];

}
print $tot;
[download]

Update: My apologies. I remember having that exact question in a first year cs class. I don't want to encourage CS students to post their home work and submit our answers.

I'm done with school, for now.

The question is for redirection of traffic at my company's website. There are two homepage test pages, and my boss wants 10% of our traffic directed to each.

Even if it was homework, isn't it a cool problem?