I'm working on a "claw game" sort of thing where t...

# getting-started

[Colton Idle]

I'm working on a "claw game" sort of thing where the randomization is dictated by my backend. The backend returns between a 0 and 10 and this is the percentage of "loser every time" to "winner every time". (0 loser, and 10 winner) and everything else between is just a percent. So ~5 is roughly 50% win/lose rate. I guess it'd be easier if the server used 0-9 instead of 0-10, but yeah.

What's the best way to take an "int" and come up with a winPercentageRate based off of that?

[Vishnu Shrikar]

your range is 0-10 so... just divide by 10? Or 10.0 if you want more accuracy... Then multiply by 100 and you have your percentage

[Colton Idle]

Once i have the percentage... is there a stdlib method to get a boolean based on a random probability?

[Vishnu Shrikar]

Random.nextInt(100)

[Vishnu Shrikar]

Use that and just check where the result falls in relation to ur percent

[Francesc]

Random.nextInt(100) gives you an Int in the [0..99] range (both included), it excludes 100. If you want to map 0-10, then you'd have to use Random.nextInt(101)

[Vishnu Shrikar]

Yes thats right, whoops, i forgot the range thing

[Vishnu Shrikar]

Wait I just read the docs, the ranges are inclusive

[Vishnu Shrikar]

So the original 100 would be correct

[Casey Brooks]

Something like Random.nextInt(100) < randomWinPercentageAsInt should do the job. See example here https://pl.kotl.in/ImlV8xKkY

[Francesc]

/**
     * Gets the next random non-negative `Int` from the random number generator less than the specified [until] bound.
     *
     * Generates an `Int` random value uniformly distributed between `0` (inclusive) and the specified [until] bound (exclusive).
     *
     * @param until must be positive.
     *
     * @throws IllegalArgumentException if [until] is negative or zero.
     *
     * @sample samples.random.Randoms.nextIntFromUntil
     */
    public open fun nextInt(until: Int): Int = nextInt(0, until)

the upper limit is excluded

[Klitos Kyriacou]

You actually do need nextInt(100). You want a value from 0 to 99, and then test it like this:

val won = Random.nextInt(100) < winPercentageRate

If your rate is 50, then Random.nextInt(100) returns a value between 0 and 99. There will be 50 such values between 0 and 49, and 50 such values between 50 and 99, giving an equal chance.

[Francesc]

enum class CatchResult {
    Win,
    Loss,
}

fun winOrLose(val input: Int /* 0-10 */): CatchResult {
    val winFraction = input.toDouble() / 10.0
    val random = Random.nextInt(101)
    val loses = random > (winFraction * 100.0).toInt()
    return if (loses) CatchResult.Loss else CatchResult.Win
}

[Francesc]

you can actually simplify all this

fun winOrLose(input: Int /* 0-10 */): CatchResult {
    val random = Random.nextInt(11)
    val loses = random > input
    return if (loses) CatchResult.Loss else CatchResult.Win
}

[Klitos Kyriacou]

val random = Random.nextInt(11)

val loses = random > input

No, if input = 10 it should win every time, but if random is 10 it will lose.

You really need nextInt(10) so between 0 and 9.

[Francesc]

random is at most 10, so the > is always false for input 10

[Francesc]

however, it should be
val loses = input == 0 || random > input

[Vishnu Shrikar]

i think its random.nextInt(10) >= input

[Klitos Kyriacou]

> random is at most 10, so the > is always false for input 10
You're right, I had got it the wrong way around.
If input is 0, it should lose, but if random is 0 it wins.

>

val loses = input == 0 || random > input

Introducing "special cases" is generally suspicious that something is wrong. Here, the probabilities are not evenly distributed.
It should be random.nextInt(10) >= input as Vishnu said.

[Colton Idle]

Glad it wasn't just me that was getting caught up in the details. Maybe i need more coffee. lol. Thank you all!

@Casey Brooks your example was amazing since it printed all of the rates and I could verify that things were working as expected. thank you for teaching!

[Vishnu Shrikar]

If you are young remember coffeee stunts your growth 🙂 but sometimes the simplest things are the hardest, 🙂

[Colton Idle]

Yeah, I ended up writing something that had like 20 lines of code... and I was like. wait. there has to be a better way. then i started going down the route of looking up a probability stdlib in java/kotlin then i just ended up posting here. Glad that the solution was a one liner.