In which you realize your dice have air bubbles

Stupid dice trick, probably useless:

Say you're stuck on a desert island and all you've got are a bag of crappy dice, like Chessex aftermarket extras, and a copy of the D&D Rules Cyclopedia. You know the dice are shitty, so how do you make them unshitty?

Roll two of them, and add the results together, rolling back around if it goes over the maximum number. So like a d6 is 2d6, and if you roll a total of 10, then thats (10 - 6 =) 4.
This will make the result close to properly random even if both dice are very badly loaded, due to averaging the distribution.

Exactly. But finding/formulating such a material would be the challenge, especially doing so subtly enough that the distributed weight isn't outright noticeable when someone holds it.

>Say you're stuck on a desert island and all you've got are a bag of crappy dice
I though you were going to tell us how to start a fire with dice or something useful :/

Nah, you're gonna die on the island, but at least you can pretend to be an elf while you do it.

Now if you had a copy of Avalon Hill's Outdoor Survival, you might be okay.

>you're gonna die on the island

I play with wooden dice.

That's bullshit, I've tried that before and the dice just sink.

You need to add more epsom salt. If you have enough, the dice will float.

>2017
>playing with dice

nignoggidism.

Just use dice rollers on your phone fucking retards

>he doesn't manipulate the random seed on his phone to roll high

BULLSHIT. I HAVE DEPLETED URANIUM DICE AND NO MATTER HOW MUCH SALT I ADD THEY STILL SINK. STOP LYING.

>dice rollers on your phone

>on your phone

underrated

Won't that actually not have a normal distribution at all?

Work the math if you like, it drastically reduces bias, even if both dice are loaded, by redistributing the biased results around to other numbers, evening out the disparities.

Alternately some guy on Veeky Forums wrote a program to test it by brute force and found it worked to his surprise. I should have asked him for source code.

This is like a silly mathematician trick, since it's kinda clunky for actual use when you could just use a better die.

>In which you lean what a refractive index is.

Daily reminder you are a gatekeeping shitlord if you don't allow dice rollers in your game. Because normalfucks can spend 3000 dollars on their smart phones but asking them to spend 2.50 on dice is too much, while pirating the books to make sure no game company ever makes a dime again.

>In which we settle this dispute with laser rifles

It works modulo the die size, but not modulo another number. So for 2d10, you would want to subtract 10 if the total is over 10.

anydice.com/program/cc06

That's true, you can count it back up yourself
1 and 7 have 16.66 %; 2 and 8 have 16.65%; 3 and 9 have 16.66%; 4 and 10 have a chance of 16.66%; 5 and 11 have 16.66% and lastly 6 and 12 have a chance of 16.65%

>I can find someone on Reddit saying it, so wrong
>What do you mean that's not an argument?
And you're seriously implying no one on this site pirates books?

That's what I said.

>add the results together, rolling back around if it goes over the maximum number.

Maximum number being the die size, 10 in the case of a d10.The d6 was just the example part.

You're doing it wrong. A value of 7 on 2d6 is 1, 8 is 2, 9 is 3. You have to add the dice together and roll the number over if it exceeds the max, which is 6 for a d6. Doing that you get a normal distribution again.

Some dice float, some sink. I had some red jewel transparent dice and they floated. I figured they would more likely be balanced. The solid red ones sunk. I have a suspicion that they love the number 16.

I did it that way?
1 0%
2 2.77 %
3 5.5 %
4 8.3 %
5 11.11 %
6 13.88 %
7 16.66 % -> 1
8 13.88 % -> 2
9 11.11 % -> 3
10 8.33 % -> 4
11 5.55 % -> 5
12 2.77 % -> 6

Die died

>have 100 different d20s
>after every roll a player has to use a different one
>dice that have been used go into a different tray
>once reach the 100 dice put them all back in the original tray

Now I don`t have to care about some unbalanced die because even if there's a dice in the pile that will always roll high/low, it's gonna be random whether it's picked or not.

I have nothing in particular against dice rollers, (I am playing star wars funky dice edition so I have to use the dice rollers until I get my own set of funky dice) but I find them to be more annoying to use than just rolling real dice.

Now add them together:

1 0+16.66 = 16.66
2 13.88+2.77 = 16.65
3 11.11+5.5 = 16.61
4 8.33+8.3 = 16.63
5 5.55+11.11 = 16.66
6 2.77+13.88 = 16.65

Apart from small rounding errors, it's flat.

I tried to do this, and my die dissolved

>even if both dice are loaded, by redistributing the biased results around to other numbers, evening out the disparities.
No. It works by reducing the impact of any one die. Weighted of otherwise. The extreme case of this technique would be to roll a infinite number of dice and use the mode of the set as the result.

Well you could probably shave the dice for plastic curls to use as a waterproof firestarter, but they'd have to be super thin. You'd be better off using coconut husks. The fibers burn well and you can use the shell to make some sandals for your tender neet feet

>using soap dice

>a dice
RETARD ALERT

po-tay-to, po-tah-to.

I literally used the correct term in the same sentence m8.

You didn't pay attention, I was saying the method works generally.
I even added up the correct number, you just immediately thought I disagreed and didn't bother reading.

Ah, I got confused because your second quote here was the guy saying "but that won't have a normal distribution at all" and then you said "That's true" like you were agreeing with him, and your dice chart image shows the 2-12 curved distribution, and so the fact that you then agreed with me didn't register because it didn't match what I thought you were saying above.
Ya threw me a curve, get it?

No. They are not the same thing. Incorrectly explaining things is pretty much like being wrong though.

>le ebin droll

Pardon me, but by adding the numbers you push the bias away from the original number, spreading it out. If you graph the results, it's a visible flattening. If my explanation wasn't precise enough for you, I apologize, but I don't think I was wrong.

It's not a troll, it works. Test it out and see, it's not that hard to do if you can code a simple script.

>And you're seriously implying no one on this site pirates books?

This is a site of the strongest moral fiber.

Same, but mostly because they look prettier. They look like magic crystals

ADD MORE SALT, YOU DIRTY FAGGOT!

Here's a shitty python code test of this I half-assed cobbled together before bed.

pastebin.com/04EMMqeE

Results of a run through of 6000 dice, where each side should show up roughly 1000 times. (Pretend it says like "100.0% of expected results")

Good die rolled : [1048, 1045, 960, 1036, 938, 972]
Bad die rolled : [848, 859, 885, 920, 817, 1670]
Added result : [968, 1018, 974, 1045, 988, 1006]

Run it yourself and see. It works.

Replaced the good die with a second, identical bad one:

First bad die rolled : [863, 846, 833, 814, 900, 1743]
Second bad die rolled : [895, 829, 872, 815, 833, 1755]
Added result : [1009, 968, 955, 955, 984, 1128]

It still works, though it's less close to perfect. It's still far better than either bad die. Run a few more times:

Added result : [1003, 921, 984, 1006, 991, 1094]
Added result : [984, 967, 963, 1023, 1009, 1053]
Added result : [998, 986, 947, 953, 1007, 1108]
Added result : [1001, 986, 1001, 947, 994, 1070]

You can see the load is still there on the six, but most of it has been spread out onto the remaining numbers. It's gone from 170.0% of the correct number of sixes down to 105-112%, a drastic reduction even with two loaded dice.

I think the take-away from all this is two-fold:

1. No one cares - except possibly people who roll dice underwater for some reason, and
2. A potential 0.0001% deviation from theoretical perfection is perfectly fine for all but the most profoundly autistic half-fey dragonkin catboi.

Play on!

Gatekeeping isn't a bad thing when you're keeping out the undesirables.

One can still easily detect this. Use a tall cylinder full of water - like a clear glass vase for example - and drop the dice in. Weighted dice are fairly easy to spot as they slowly rotate to the same general side while settling to the bottom.

Isn't this somewhat suspect since PC's don't really generate totally random numbers? Hence the term 'Psuedo-Random' number generation...

even the crappiest dice when I test them are not that off from random unless someone has tried to load them.

>desert island and rules cyclopedia
sounds like fun

...

I ADDED SO MUCH SALT THE DIE MADE A CRATER IN MY WATER AND NOW IT WON'T SHIFT TO ANY SIDE.

Daily reminder that random number generators aren't completely random and follow an algorithm to generate numbers that feel random.

Nothing is actually random, but a good pseudo-random number generator is sufficient approximation.

Using dice is significantly more random than a random number generator because it's unpredictable and unless the dice have an unbalanced weight ratio the outcome is unpredictable. Random Number Generators use algorithms which can be predicted and exploited.

Depends on the algorithm.

Barring a coding error like debian had for a while, the random numbers from linux calls are rather hard to predict as they are generated by entropy

I need to find that site with the guy who built a machine to roll dice and record rhe results, he ran it for 144,000 rolls. Granted they were small d6'es, but he found out that the material removed to make the dimples for the colors unbalanced them enough to make them roll a disproportionately high amount of 1's and low amount of 6'es.

Dimples for the numbers*

dakkadakka.com/wiki/en/That's_How_I_Roll_-_A_Scientific_Analysis_of_Dice

add more salt. Yeah it is a lot of salt you are going to waste so don't waste table salt. I tried it and got all my dice to float and I could swear the glass was 50% salt by the time they finally floated. Don't be cheap on the salt.

Epsom salt will work better than table salt.

nope. quite the opposite. With Dice I can learn the right way to roll it or pick my lucky (unbalanced) die to always roll good. I can also nudge my dice and ask to roll again or a million other tricks that just wont work in a dice roller. I hate the rolls in my dice roller compared to what I can get away with using actual dice.

Wish I had thought of that at the time.

A good random number generator is about as good as a die.
Bad dice will lean towards producing certain numbers more than others, whereas most bad RNGs will instead produce recognizable patterns, or in some really bad cases, repeatable results.
The linux system RNG is excellent, so your android phone has a good one built in. However, there's no way to tell if an app writer has used the system's RNG or amateurishly coded his own (which is a bad idea, but you'd be surprised what some coders do) without testing the app thoroughly. I wouldn't trust a dice roller that doesn't provide source code or at least some other method of verification, myself. (Plus I think there's no substitute for the feel and clatter of real dice.)

Add enough salt to the water to make it dense enough for the die to float, spin it and check that way. Easier to repeat the test.

What brands of dice are best? I had been buying chessex, but this thread seems to think they're shit. What are better brands?

Chessex is fine, just be aware that they sell both good ones and cheaper factory seconds. (The "Pound o' Dice" bag is the latter) If you don't mind picking out the ones that roll funny, the latter is a way to get some super cheap dice.

Daily reminder that there's no such thing as truly random in the physical world, and any such beliefs only stem from inadequacies of understanding

Daily reminder that there websites in which you can roll dice with HARDWARE random number generators. They're as random as dice and don't use seed which can be manipulated.

Well aren't you the most enlightened motherfucker in the room

Is there a good way to check the balance of metal dice? I doubt the salt thing would work

inb4 >>using table wreckers

There's a mathematical test you can use that will tell you with some certainty, with like 30 rolls of a d6.
deltasdnd.blogspot.com.es/2009/02/testing-balanced-die.html

...

> the undesirables.

No such thing, normie.

Just take a bit out.

Please, unless you are opening a casino, never ever take dice seriously.

funny enough i just found out my d6 is bugged at about 997 out of 1000 significance.
yeah i rolled it like 360 times.

it rolls 5 slightly more than half the amount it should(37 instead of 60), and absolutely adores 1 and 6.
this shit makes little sense because the 1 face is less carved than the 5's so it should come up less, and a single bubble can't alone explain the anomaly since the most rolled faces( 1 and 6) are opposites.there might be multiple factors in play.