Are dice true random?

[Schnorble Schnacker]

At the start of the game it felt normal, few crits, few crit misses, normal sort of odds when hitting and missing.

Now I'm missing attack rolls when it's 4 or lower a hell of a lot. Always 4 or lower. If it's between 5 or higher it feels like normal dice rolls.

Edit: For reference im not accusing the game of being "rigged" and although everybody goes "it's confirmation bias" i always track data on this before questioning it.

I crit as often as i should (1 in 20 or 5% if you prefer percentages)

Yet going by the log of my last few fights since posting i roll a 1 12%(ish) of the time (thats 1 in 8 roughly)

Again as checking this post I go through the first round where the needed rolls are between 2 and 4. Out of the 12 rolls 2 landed a hit. Second round with higher rolls needed to hit and 9 hits land

I guess it's just ridiculous luck (or rather bad luck) that is occurring

[Drake]

Originally posted by Ryanatix:

At the start of the game it felt normal, few crits, few crit misses, normal sort of odds when hitting and missing.

Now I'm missing attack rolls when it's 4 or lower a hell of a lot. Always 4 or lower. If it's between 5 or higher it feels like normal dice rolls.

Confirmation bias. Random doesn't mean entropic distribution.

A true random series of a d20 could give you the infinite decimals of pi. You would have to be extremely unlucky or lucky, but it can definitely happen.

It's been checked several times. The game uses the standard random function of unity engine. There is no transformation made to the roll. While there is a seed and if you have all the parameters known before hand, you could predict the next roll in theory (which is in itself very difficult). But the game doesn't adjusts the rolls in anyway depending on your playthrough.

You just have bad luck.

[reidj062]

You can roll 3 nat 1s in a roll in a regular tabletop game. I speak from experience. Additionally...yeah rolling a four or lower generally is a miss. It also depends on what sort of bonuses and penalties you have.

[EvilDonut]

Originally posted by reidj062:

You can roll 3 nat 1s in a roll in a regular tabletop game. I speak from experience. Additionally...yeah rolling a four or lower generally is a miss. It also depends on what sort of bonuses and penalties you have.

Yes, but i NEVER seen, in reality, anyone rolling 1 TEN TIMES IN A ROW unless the dice were rigged.
Happened to me yesterday.

[Razer]

natural 1s are more telling and annoying so you tend to see them more. You hardly notice the immense amount of 20s you roll. Is it rigged? Sometimes it feels like it is, but then there's always that 5% chance. So no it's not rigged. And 10 misses in a row is also possible. In fact 100 misses in a row is also possible. You shouldn't be looking for average performance in a game of chance.

It's especially unforgiving in the early game where an enemy rolling a 20 once may cause a TPK because of inflated enemy stats.

[EvilDonut]

It's not a matter of noticing it, it's a matter of what's in the log. Ten rolls, ten 1s.

[Razer]

Originally posted by EvilDonut:

It's not a matter of noticing it, it's a matter of what's in the log. Ten rolls, ten 1s.

Yep. Possible. Not rigged. Even you get 100 1s. Still possible.

[EvilDonut]

Lol the fact that it's physically possible for it to happen doesn't mean it should happen with the frequency to which it happens in the game. It's obvious there's some script somewhere that periodically goes rogue.

[Nosferatu]

Originally posted by Ryanatix:

At the start of the game it felt normal, few crits, few crit misses, normal sort of odds when hitting and missing.

Now I'm missing attack rolls when it's 4 or lower a hell of a lot. Always 4 or lower. If it's between 5 or higher it feels like normal dice rolls.

Edit: For reference im not accusing the game of being "rigged" and although everybody goes "it's confirmation bias" i always track data on this before questioning it.

I crit as often as i should (1 in 20 or 5% if you prefer percentages)

Yet going by the log of my last few fights since posting i roll a 1 12%(ish) of the time (thats 1 in 8 roughly)

Again as checking this post I go through the first round where the needed rolls are between 2 and 4. Out of the 12 rolls 2 landed a hit. Second round with higher rolls needed to hit and 9 hits land

I guess it's just ridiculous luck (or rather bad luck) that is occurring

It's random, or at least I've never seen any convincing evidence indicating otherwise. Sometimes bad streaks happen, obviously, but I've also had fights where my party just can't seem to stop critting. I've finished the game once and am in the early phase of my second playthrough, and my rolls overall have been about what you'd expect.

[CascadingDragon]

Originally posted by EvilDonut:

Lol the fact that it's physically possible for it to happen doesn't mean it should happen with the frequency to which it happens in the game. It's obvious there's some script somewhere that periodically goes rogue.

I feel like you don't understand what "Random" means in RNG.

[EvilDonut]

Originally posted by Boring Killer Player:

Originally posted by EvilDonut:

Lol the fact that it's physically possible for it to happen doesn't mean it should happen with the frequency to which it happens in the game. It's obvious there's some script somewhere that periodically goes rogue.

I feel like you don't understand what "Random" means in RNG.

And i feel like you don't understand that if something that has a microscopic chance of happening and most likely would never happen in real life starts happening again and again it means there's something broken in the way chances are being calculated.

[Deo]

Originally posted by Ryanatix:

I guess it's just ridiculous luck (or rather bad luck) that is occurring

Dices are not random. Enemies have higher chance to roll high on d20. It's easy to test.

Make Witch, take her Protective Luck. Find an enemy. Make tank with AC that is enemy's AB + 18-19. So enemy will need 18-19 to hit you. Make sure that your tank will need 18-19 to hit enemy's AC as well.

Then fight it with your tank and Protective Luck from Witch. The amount of times enemy will roll double 18 or double 19 is very high. Happens around round 7-8 usually. The statistical chance to roll double 18 or double 19 is absurdly small.

As long as your tank get hit, make a mark somewhere and reload and do the fight again. Only make mark if your tank will get hit on 18,18 or 19,19. Rolls like 18,20 or 19,20 are statistically higher than doublers. You will see then, that at least in 4 fights out of 8 enemy will roll double 18 or double 19.

Then you can test that again, but you will be casting Protective Luck on enemy (it's possible). The chances of your controlled character to roll double 18 or double 19 will be absurdly low, like 1 fight out of 12. Rolls 18,20 and 19,20 do not count for that.

Do that when you are bored, and see for yourself.

[CascadingDragon]

As mention earlier, we literally have access to the code, and it's just regular RNG. I'm sorry that you are extremely unlucky, but that doesn't change what the truth.

[ADDB]

There is no script.

Here[docs.unity3d.com] is an explanation of how RNG is implemented in Unity, and it’s based on this paper about Xorshift-RNGs[www.jstatsoft.org]

All the game does is use Unity Random. No weight; no checks; no better rolls for enemies or anything.

You have to understand that in RNG, 1K rolls if nothing. Even 10K isn’t much. RNGs are (more or less) equally distributed over a very very long amount of numbers. The longest period of Xorshift128 is (afaik) 2^128-1. Which means you’d need ~10^38 dice throws until the series starts repeating; so if you want to show that the RNG is actually weighted you’d need to do more than "but the last 10 dice I my log".

[Deo]

Originally posted by ADDB:

All the game does is use Unity Random.

Prove it.

[ADDB]

Originally posted by Deo:

Originally posted by ADDB:

All the game does is use Unity Random.

Prove it.

As you can easily see yourself if you use a Decompiler:

Anything that throws a dice in the game calls

private void Roll() {
    this.m_Result = this.m_PreRolledResult ?? RulebookEvent.Dice.D(this.DiceFormula);
    this.RollHistory = new List<int> { this.m_Result };
    for (int i = Math.Abs(this.m_RerollAmount); i > 0; i--) {
        this.Reroll();
    }
}

which calls

public static int D(DiceFormula formula) {
    return RulebookEvent.Dice.D(formula.Rolls, formula.Dice);
}

which calls

private static int D(int rolls, DiceType dice) {
    int num = 0;
    while (rolls-- > 0) {
        int num2 = global::UnityEngine.Random.Range(1, dice.Sides() + 1);
	num += num2;
    }
    return num;
}

which uses the earlier described UnityEngine.Random