How do I make random number generation more random?

[civic lynx]

Hello! So I’ve been trying to write a program for coupon collector program in case of six faced die.
I have attached pic of my code + pasted my code below. Lemme know if you have any idea!

import java.util.Random;
import java.util.Arrays;

class die{

public static void main(String[] arg){

Random rand = new Random();

int[] T = new int[2];
T[0] = 1+rand.nextInt(6);
int i = 1;
int j = 1;

        do{
             int h = 0;
             T[i] = 1+rand.nextInt(6);
             
             for(int k=0; k<i; k++) { if(T[i] != T[i-k]){h++;} }
             if(h==i){j++;}  
            
              
              i++;

              int[] temp = new int[i+1];
              temp=T;

              T = new int[i+1];
              T=temp;                      }while(j!=6);

System.out.print(i);


}}```

[spark pendant]

Detected code, here are some useful tools:

Formatted code

import java.util.Random;
import java.util.Arrays;

class die {
  public static void main(String[] arg) {
    Random rand = new Random();
    int [] T = new int [2] ;
    T[0]  = 1 + rand.nextInt(6);
    int i = 1;
    int j = 1;
    do {
      int h = 0;
      T[i]  = 1 + rand.nextInt(6);
      for (int k = 0; k < i; k++) {
        if (T[i]  != T[i - k] ) {
          h++;
        }
      }
      if (h == i) {
        j++;
      }
      i++;
      int [] temp = new int [i + 1] ;
      temp = T;
      T = new int [i + 1] ;
      T = temp;
    }
    while (j != 6);
    System.out.print(i);
  }
}

<@&987246883653156906> please have a look, thanks.

[civic lynx]

im kinda more interested in why this code doesn't work rather than looking at the optimized one

actually screw this

im using lists

so much better

:nozoomi:

[mortal girder]

What are you trying to do

Wait nvm

Nvm what are you trying to do

[civic lynx]

Trying to simulate coupon collector problem for a die

I got it btw

import java.util.Random;
import java.util.function.Supplier;

class CC{
public static void main(String[] arg){
  
  int k = Integer.parseInt(arg[0]); //Number of coupons to be collected
 
  Random rand = new Random();
  List<Integer> T = new ArrayList<Integer>(); //Coupon Collector
  
  int i = 1;
  int j = 1;
  
  T.add(1+rand.nextInt(k));

                                do{  int x = 1+rand.nextInt(k);
                                    
                                     if(T.contains(x)){ }
                                     else{ T.add(x); j++; }

                                     i++;                       }
                                
                                while(j!=k);

  System.out.print(i);

}}

[spark pendant]

Detected code, here are some useful tools:

Formatted code

import java.util. * ;
import java.util.Random;
import java.util.function.Supplier;

class CC {
  public static void main(String[] arg) {
    int k = Integer.parseInt(arg[0] );
    //Number of coupons to be collected
    Random rand = new Random();
    List<Integer> T = new ArrayList<Integer>();
    //Coupon Collector
    int i = 1;
    int j = 1;
    T.add(1 + rand.nextInt(k));
    do {
      int x = 1 + rand.nextInt(k);
      if (T.contains(x)) {
      }
      else {
        T.add(x);
        j++;
      }
      i++;
    }
    while (j != k);
    System.out.print(i);
  }
}

[civic lynx]

This simulates CC for k coupons

[civic lynx]

Okay this is weird

I checked my program pretty thoroughly so I don't think there are any mistakes

But I have problem with the degree of randomness Random() offers

This simulation does give the correct expected value of Time(random variable) for coupon collector problem(CC)

But the averages are too small compared to real life

What can I do to improve the degree of randomness ?

How do I make random number generation more random?

[civic lynx]

Here is the code if anyone whats to check for mistakes

[hidden halo]

The randomness of the default isn't terrible. SecureRandom can be used for a cryptographically strong source of randomness - but it's comparatively expensive.

You can simplify the code to Find the first k unique values.

1. use a Set.
2. loop while the Set size is less than the target k adding a random value to the set (duplicates will be ignored for you).
3. Use ThreadLocalRandom.current() so you don't have to create a random to use.

If you care about the order they were encountered you can add successful checks to a list (or use LinkedHashMap to collect the matches) - or do it in a method that takes a consumer so you can decide later.

[civic lynx]

👍 will try that

Will have to look all those up

[civic lynx]

I was thinking of I could using smthg like this

https://www.random.org/

And integrate it in my code

Tho I dunno much about interface and stuff :P

[uneven salmon]

Built in random is random enough for what you are doing.

Are you just trying to make a histogram of random numbers?

[worldly bone]

also use ThreadLocalRandom instead

[uneven salmon]

meh... leave an option for that by making the random a parameter. That way when you do testing you can chose a fixed seed.

[worldly bone]

ok yeah if thats needed

[civic lynx]

No, so I use my code above as a method and use it to find the averages say : sum(1<=i<=N) C(6) /N where N varies

And I noticed these averages are smaller compared to when I actually try to do this irl

Which is weird

[molten pebble]

@civic lynx

you don't need any of that

just use what you had

Random is good enough so it's impossible for human to see any pattern

if you see a pattern, either your perception is biased, or the problem is in your code

[civic lynx]

.

[molten pebble]

here it should average to 3

in a long run

[civic lynx]

for 6 sided die ?

for 6 sided die the expected stopping time is 6H_6

About 14.7

[molten pebble]

wdym

[civic lynx]

The average

By law of large numbers the avg in our case converges to expected value

when i say expected value I’m talking about expectation of our underlying time random variable

but we don’t need to get into detail

So i wrote a code to find this avg

import java.util.Random;
import java.util.function.Supplier;

class avgCC {

 public static int CC(int y) {
  
  Random rand = new Random();
  List<Integer> T = new ArrayList<Integer>(); //Coupon Collector
  
  int i = 1;
  int j = 1;
  
  T.add(1+rand.nextInt(y));

                                do{  int x = 1+rand.nextInt(y);
                                    
                                     if(T.contains(x)){ }
                                     else{ T.add(x); j++; }

                                     i++;                       }
                                
                                while(j!=y);

  return i;         }

 public static void main(String[] arg) {
              
           int k = Integer.parseInt(arg[0]);
           int N = Integer.parseInt(arg[1]);
           float avg=0;
     for(int m=1; m<N; m++){ avg = avg + (float)CC(k)/(float)N; };
    
     System.out.print(avg);
} }

[spark pendant]

Detected code, here are some useful tools:

Formatted code

import java.util. * ;
import java.util.Random;
import java.util.function.Supplier;

class avgCC {
  public static int CC(int y) {
    Random rand = new Random();
    List<Integer> T = new ArrayList<Integer>();
    //Coupon Collector
    int i = 1;
    int j = 1;
    T.add(1 + rand.nextInt(y));
    do {
      int x = 1 + rand.nextInt(y);
      if (T.contains(x)) {
      }
      else {
        T.add(x);
        j++;
      }
      i++;
    }
    while (j != y);
    return i;
  }
  public static void main(String[] arg) {
    int k = Integer.parseInt(arg[0] );
    int N = Integer.parseInt(arg[1] );
    float avg = 0;
    for (int m = 1; m < N; m++) {
      avg = avg + (float ) CC(k) / (float ) N;
    };
    System.out.print(avg);
  }
}

[civic lynx]

This does converge for k=6 and large N to 14.7 (on average)

but for lower values of N the avg tend to be much smaller than what is observed irl

[molten pebble]

see, it works

[civic lynx]

Thats not the same thing

My random variable is different

You are avging the x themselves

Which will tend to 3.5

bcuz x is uniformly distributed

[civic lynx]

I’m talking about these avgs

[molten pebble]

what does it give ?

[civic lynx]

So basically

[molten pebble]

i am saying, that random works perfectly

[civic lynx]

Sure in limit it works perfect

that is if you sample it large number of times

[civic lynx]

This code basically does the coupon collector problem N times

And takes their mean

so for k=6 this is asks how many times you have to roll a die so that you see all the faces

and i do this process N times

Then I take the mean

That’s basically what the avgCC program is doing

[civic lynx]

But when I perform this irl

even for N=10

the avg given by program is a bit small

[civic lynx]

I’ve done it multiple times irl now

so this makes me wonder if it’s my code or the random()

[molten pebble]

that's what you need to do yes

[molten pebble]

like ?

and did you calculate the probablility of N ?

[civic lynx]

So irl the avg is almost always between 17-25

But the program gives me 11-14

Always

[civic lynx]

N is a integer

[civic lynx]

See here

[molten pebble]

I mean, the probability of N being between 11-14

[civic lynx]

N is a free variable

Do you mean i?

[molten pebble]

ah

I mean

the average

[molten pebble]

this

[civic lynx]

oh

The avg so that should be very high

bcuz expected value is 14.7

And by LLN avgs should go to this value

For large N

[molten pebble]

how do you know that ?

[civic lynx]

You can prove it mathematically

Look up law of large numbers

[molten pebble]

also, did you try to reuse the Random object ?

[civic lynx]

You mean rand?

(Im kinda new to java)

[molten pebble]

yes

[civic lynx]

Yeah

I used it repeatedly to get random values between 1 to 6

[molten pebble]

I indeed get 14.7

[molten pebble]

each time you create a new one, can you try to reuse the same one please

[civic lynx]

Yeah thats 6x(1+1/2+1/3+1/4+…+1/6)

[civic lynx]

How can i do that ?

I’m calling the same random object repeatedly

rand

So no new random objects are getting created

[molten pebble]

yea

you shouldn't do that

but

[civic lynx]

O

Wouldn’t it give me new random number everytime tho?

I mean it does give me new integer everytime

[molten pebble]

int CC(RandomGenerator rand, int y) {
        Set<Integer> T = new TreeSet<Integer>();
        //Coupon Collector
        int i = 1;
        int j = 1;
        T.add(1 + rand.nextInt(y));
        do {
              int x = 1 + rand.nextInt(y);
              if (T.contains(x)) {
              }
              else {
                    T.add(x);
                    j++;
              }
              i++;
        }
        while (j != y);
        return i;
  }
void run(int k, int N, RandomGenerator rand) {
  long sum = 0;
  for (int m = 0; m < N; m++) {
    sum += CC(rand, k);
  };
  System.out.print(sum/(double)N);
}
run(6, 100000)

anyway

@civic lynx what's the problem already ? because personally I find 14.7

[civic lynx]

I too get 14.7 , the problem is for smaller values of N

Run this repeatedly for N= 10

You get avgs in range 10 to 16

and irl you get avgs in range 17 to 25

Of course they will match for large values of N (that is in limit)

its the small values that are problematic

[molten pebble]

but then, what is the probability for small values of n ?

[civic lynx]

N is free variable bro

N is the number of time you wanna do this process of finding out the number of tosses to see all sides of die

[molten pebble]

i am talking about avg

[civic lynx]

Yeah

It is avg over the N values of T

[molten pebble]

what is the probability of having avg between 11 and 14 for small values of n ?

[civic lynx]

lets see hmm

[molten pebble]

Because random doesn't lie

[civic lynx]

All hail random()

[molten pebble]

I tried different random algo and they seem to all give of the same results

[civic lynx]

OOOF

hm

so maybe the rolls i did were skewed ?

But i rolled the online dies

[molten pebble]

or maybe your formula is wrong

[civic lynx]

Which formula

[molten pebble]

14.7

[civic lynx]

Nah thats correct

Its a well known problem

coupon collector

[molten pebble]

it's only 14.7 for large n right ?

[civic lynx]

Yeah

but we have the following

Do we have latex here

$test$

[molten pebble]

== we have this thing

maybe

[civic lynx]

[molten pebble]

how does it work

idk

the thing with this problem

[civic lynx]

Lemme just photo

[molten pebble]

is that it tends to 14.7 right

but

it's not your regular average

because it doesn't do 14 then 18 then 12 then etc

[civic lynx]

so the probability is only considerably small when

c>13

[molten pebble]

it's rather that most results are under 14.7 and only when you take them all, it gives something like 14, 13, 12.5, 14.7, 16.8, etc and in the long run, it will average with 14.7 but there are little average that are more than 14.7

maybe you could try to do a median

[civic lynx]

But the usual avgs should work

LLN holds

[molten pebble]

what is c ?

[civic lynx]

Error

T is the avg

Or rather

Value of time T

But huh we want avg

Just a sec

Yeah so there N to

So we have

Let AVG be the avg random variable

[molten pebble]

alright

it works

[civic lynx]

epsilon our error

[molten pebble]

you can use this

[civic lynx]

Yeah so the thing you asked

AVG_10 being avg when N=10

So yeah its quite less

And it is confirmed by the irl tosses i did

Avgs lied between 17 and 25

So there must mistake in code or the random()

[molten pebble]

but why, didn't you say that the avg should be 14.7 ?

[civic lynx]

Yeah but for large N

Oh wait i mean

oh fuck

Its the probability of AVG_10 being greater than 18.7 or less than 10.7

which is less than 0.37

huh

so the only thing faulty was the tosses i did IRL

[molten pebble]

can you try this ? I changed a few things

[civic lynx]

Okay

[molten pebble]

lol

[civic lynx]

https://tenor.com/view/thanos-reality-is-often-disappointing-marvel-mcu-gif-16316172

So true

[civic lynx]

Getting buncho ; erros

Probably I'm missing some stuff

But yeah I don't think we need to dig this anu further

random() works

🙏

[spark pendant]

@civic lynx

Your question has been closed due to inactivity.

If it was not resolved yet, feel free to just post a message below
to reopen it, or create a new thread.

Note that usually the reason for nobody calling back is that your
question may have been not well asked and hence no one felt confident
enough answering.

When you reopen the thread, try to use your time to improve the quality
of the question by elaborating, providing details, context, all relevant code
snippets, any errors you are getting, concrete examples and perhaps also some
screenshots. Share your attempt, explain the expected results and compare
them to the current results.

Also try to make the information easily accessible by sharing code
or assignment descriptions directly on Discord, not behind a link or
PDF-file; provide some guidance for long code snippets and ensure
the code is well formatted and has syntax highlighting. Kindly read through
https://stackoverflow.com/help/how-to-ask for more.

With enough info, someone knows the answer for sure 👍

[civic lynx]

Okay brotha

[odd rivet]

You reopened the thread because u have written a new message