Picking random item from PHP array based on probability

Question

I have a map of users and a count of entries/tickets they have purchased for a raffle (or lottery or any other similar event).

The map of users to entries is in the structure as seen below:

// Map the person to the amount of entries they have purchased
$entries = [
    'Adam' => 5,
    'James' => 3,
    'Oliver' => 4,
    'Holly' => 8
];

I want to choose a random user but take into consideration their chance based on ticket count. The probability of winning must be:

(User Ticket Amount / Total Ticket Count) * 100 = Probability percentage

For example, from the test array the expected results is Holly will win the raffle 8 times out of 20 times.

Kindly post some of your code/efforts then only we can help you. — Prabhjot Singh Kainth
– Prabhjot Singh Kainth, Commented Dec 10, 2019 at 13:42
Well, in term of probability can you just build an array with all the person (5 adam, 3 james, etc.) and then just use array_rand ? — Mickaël Leger
– Mickaël Leger, Commented Dec 10, 2019 at 13:44
@MickaëlLeger how would i do that? go over array and then add the user to a new array as many times as they got entries? surely if there are thousands of entries for one user and they have a longer name this can cause parformance issues? — Adam Zorchubiez
– Adam Zorchubiez, Commented Dec 10, 2019 at 13:50

Josh Wood · Accepted Answer · 2019-12-10 14:35:21Z

8

It's a coincidence you are asking this as I created a method to do just this the other day when writing a script for a betting site.

Please see the comments for explanation of what the code consists of:

// The array passed to the function should be your $entries array
function randProb(array $items) {
  $totalProbability = 0; // This is defined to keep track of the total amount of entries

  foreach ($items as $item => $probability) {
      $totalProbability += $probability;     
  }

  $stopAt = rand(0, $totalProbability); // This picks a random entry to select
  $currentProbability = 0; // The current entry count, when this reaches $stopAt the winner is chosen

  foreach ($items as $item => $probability) { // Go through each possible item
      $currentProbability += $probability; // Add the probability to our $currentProbability tracker
      if ($currentProbability >= $stopAt) { // When we reach the $stopAt variable, we have found our winner
          return $item;
      }
  }

  return null;
}

Any questions feel free to ask below, if you would like to output the value (probability) of the winning item, return $probability at the breakpoint instead of $item

edited Dec 10, 2019 at 14:35

answered Dec 10, 2019 at 13:46

Josh Wood

4724 silver badges14 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Adam Zorchubiez Over a year ago

this is perfect I think this will work... I'm unsure where to put this code though at top of my file?

Josh Wood Over a year ago

@AdamZorchubiez I'm glad I could help

Iman Emadi Over a year ago

@JoshWood this can't be completely random , as if $stopAt is a little number , like 4 and the user with low probability like 7 , ( assume others above 7 ) can win this if hes first in the list . sort of the items matter , so it's not random

Luca C. Over a year ago

does this work both with sorted and unsorted arrays?

GrumpyCrouton · Accepted Answer · 2019-12-10 13:50:40Z

1

This should do what you are trying to do, and is pretty simple to understand.

function rand_with_entries($entries) {

    //create a temporary array
    $tmp = [];

    //loop through all names
    foreach($entries as $name => $count) {

        //for each entry for a specific name, add name to `$tmp` array
        for ($x = 1; $x <= $count; $x++) {
            $tmp[] = $name;
        }

    }

    //return random name from `$tmp` array
    return $tmp[array_rand($tmp)];
}

answered Dec 10, 2019 at 13:50

GrumpyCrouton

8,6367 gold badges37 silver badges81 bronze badges

5 Comments

Cid Over a year ago

Let's hope there won't be too much people buying too much tickets. RAM will suffer a lot

GrumpyCrouton Over a year ago

@Cid I think you would sooner hit your max allowed memory size in many environments before this causes any sort of issues. I tried it on multiple online PHP testers with 300k+ array keys and it spit out the name instantly. On my own personal server it also handled 300k with ease. I would do more, but generating the huge $entries array starts to take some time above 300k

Cid Over a year ago

300k numeric or string keys ?

Cid Over a year ago

Check this fiddle

GrumpyCrouton Over a year ago

@Cid Strings, I was using md5(rand()) for the keys.

Josh Wood · Accepted Answer · 2019-12-13 13:01:58Z

0

Ok so you have this data :

$entries = [
    'Adam' => 5,
    'James' => 3,
    'Oliver' => 4,
    'Holly' => 8
];

One way can be to fill an array according to this data and use array_rand. Here is an example :

function pickRand($array, $nb) {
    $newArray = [];
    foreach ($array as $name => $probability) {
        $num = $probability;
        while ($num > 0) {
            $newArray[] = $name;
            $num--;
        }
    }
    return $newArray[array_rand($newArray, $nb)];
}

And to use it :

$randomName = pickRand($entries, 1);

edited Dec 13, 2019 at 13:01

Josh Wood

4724 silver badges14 bronze badges

answered Dec 10, 2019 at 13:51

Mickaël Leger

3,4482 gold badges21 silver badges37 bronze badges

2 Comments

GrumpyCrouton Over a year ago

What is your $nb variable? Amount of returns?

Mickaël Leger Over a year ago

@GrumpysaysReinstateMonica yes, since it's a param in array_rand, better to use it in a function I think !

Collectives™ on Stack Overflow

Picking random item from PHP array based on probability

3 Answers 3

4 Comments

5 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

5 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related