20-sided smart dice over the rainbow: Ars tests $39 LED-filled Pixels Dice

sneezeparty

Seniorius Lurkius
41
Subscriptor
Upvote
73 (73 / 0)

Wheels Of Confusion

Ars Legatus Legionis
75,421
Subscriptor
Can you roll it 1000 times or so and post the results of the rolls? I'm curious about the balance.
It's probably crap, but the dirty little secret of TTRPGs is that all gaming dice are pretty crap. Even brands like Game Science that tout their exacting specifications in the name of balance produce randomly biased dice. The most fair dice are those made for casinos, which take into account the relative densities of both the die body itself and the putty used to fill in the pips. Gaming dice don't typically go that far.

There's a rigorous lab test here. It also seems to indicate that the float test isn't a valid one.
The upside is that their unfairness usually only manifests after many thousands of rolls, not a few hundred or even one thousand.

Conclusions
1. All dice are unfair because they cannot be precisely manufactured with uniform density and dimensional correctness. However, the number of rolls required to detect the unfairness of a casino die manufactured to a few ten thousandths of an inch tolerance is very high.
2. It is very difficult to manufacture D20s that will test fair in 3000 rolls. This includes both plastic and metal dice.
3. Though our sample was small, there appears to be some merit to measuring the diameters of D20 dice; a die with a maximum diameter difference of less than 0.12 mm has a good chance of testing fair in 3000 rolls.
4. Float testing dice may tell if the die is unbalanced but it will not tell which side or sides will roll with higher or lower probabilities.
5. The manner in which a die is rolled makes a large difference in the face probabilities. It varies with surface and the manner in which it is rolled, that is, dice cup to felt, drop from a height to a hard surface, or rolling in a dice tower.
6. Rolling D20 dice 100 times and using the obtained chi square values to judge relative fairness is statistically unjustifiable.
7. A D20 die must be highly unfair to be detected in a few hundred rolls.
8. The exact probability distributions for unfair dice are not asymptotically a chi square distribution.
 
Upvote
153 (153 / 0)
Post content hidden for low score. Show…

Wheels Of Confusion

Ars Legatus Legionis
75,421
Subscriptor
they'll eventually come in other popular polyhedral flavors (6-sided, 10-sided, etc.).

Guessing that's a typo, if not what use is there for a die that isn't a regular polyhedron?
Easy with the pitchfork, Plato.
As long as it's symmetrical it can be used. D10s (and the double-digit version, the percentile die) are not uncommon.

twitpic.png
 
Upvote
143 (143 / 0)

DRJlaw

Ars Praefectus
5,722
Subscriptor
they'll eventually come in other popular polyhedral flavors (6-sided, 10-sided, etc.).

Guessing that's a typo, if not what use is there for a die that isn't a regular polyhedron?

I didn't see the adjective "regular" anywhere in that sentence. That's not an inherent requirement of a polyhedron.
 
Upvote
68 (68 / 0)

Eastman

Seniorius Lurkius
44
Subscriptor++
Can you roll it 1000 times or so and post the results of the rolls? I'm curious about the balance.
It's probably crap, but the dirty little secret of TTRPGs is that all gaming dice are pretty crap. Even brands like Game Science that tout their exacting specifications in the name of balance produce randomly biased dice. The most fair dice are those made for casinos, which take into account the relative densities of both the die body itself and the putty used to fill in the pips. Gaming dice don't typically go that far.

There's a rigorous lab test here. It also seems to indicate that the float test isn't a valid one.
The upside is that their unfairness usually only manifests after many thousands of rolls, not a few hundred or even one thousand.

Conclusions
1. All dice are unfair because they cannot be precisely manufactured with uniform density and dimensional correctness. However, the number of rolls required to detect the unfairness of a casino die manufactured to a few ten thousandths of an inch tolerance is very high.
2. It is very difficult to manufacture D20s that will test fair in 3000 rolls. This includes both plastic and metal dice.
3. Though our sample was small, there appears to be some merit to measuring the diameters of D20 dice; a die with a maximum diameter difference of less than 0.12 mm has a good chance of testing fair in 3000 rolls.
4. Float testing dice may tell if the die is unbalanced but it will not tell which side or sides will roll with higher or lower probabilities.
5. The manner in which a die is rolled makes a large difference in the face probabilities. It varies with surface and the manner in which it is rolled, that is, dice cup to felt, drop from a height to a hard surface, or rolling in a dice tower.
6. Rolling D20 dice 100 times and using the obtained chi square values to judge relative fairness is statistically unjustifiable.
7. A D20 die must be highly unfair to be detected in a few hundred rolls.
8. The exact probability distributions for unfair dice are not asymptotically a chi square distribution.
For regular tabletop use I don't think it's very important with balanced dice. I don't even mind playing with cheap or home made novelty dice that are heavily biased one way or another if it's just with my friends because there's a good chance you'd never know if you weren't keeping a record of the rolls.

Part of the fun when we were younger was having "cursed" dice that for one reason or another got a bad rep after a particularly bad run of 1's or 20's, depending on which end of the outcome you were. "No, give me that other die, I'm not rolling with this one again!" :D
 
Upvote
78 (80 / -2)

Wheels Of Confusion

Ars Legatus Legionis
75,421
Subscriptor
"No, give me that other die, I'm not rolling with this one again!" :D
And when someone does that, you absolutely know which way that roll is about to go :D :D
Or if they preface it with "Don't fuck me, Gil!" :eng101:


But seriously, aside from EGREGIOUSLY unfair dice it doesn't happen often enough to matter. People like to look for mechanisms behind random chance, or create little rituals that they instinctively feel will govern the outcome one way or another, when simply tossing a die from a cup or shaken hand is usually more than enough to get "good enough" results.
For some people I think it's just a natural expression of our desire to believe there are ways to control randomness because randomness is uncomfortable, rather than thrilling, for them. For others it's probably part of an inherently superstitious outlook. It's a powerful reflexive belief for many, and I've even found myself automatically inventing rationalizations or imaginary rules to put a system around something that's actually just chance. So I try not to be superstitious because it brings bad luck.
 
Upvote
88 (90 / -2)
20+ years of playing D&D I will pass and just use my plain old fancy looking resin dice.

$40 would purchase me a new board game rather than a gimmick.

If you can afford it, go for it they already have 20K backers. Stuff like this never excites me because the shine eventually goes away or it breaks in some way.
 
Upvote
28 (35 / -7)
Upvote
2 (5 / -3)

BasiliskPie

Smack-Fu Master, in training
64
Subscriptor++
But... why N-sided die? Once you've put the electronics inside, why not just
- use a 6-sided die - of no particular "balance" (who cares if one side is more likely "up")
- allow each side to display "any" number (e.g. 0-99)
- permit the Owner to set the die to a range (e.g. 1-31)
[or even multi-die ranges (e.g. modelling the sum of 2 seven-sided dice plus a 13-sided one)]
- use the electronics to randomize an answer within that range
- and display the answer on all sides

With this you get The Universal Die, true (okay, Pseudo-) randomness, and the Joy (?) of throwing the die (or dice if you can afford several).

Yeah, 'haven't played those games for decades....
 
Upvote
24 (36 / -12)
But... why N-sided die? Once you've put the electronics inside, why not just
- use a 6-sided die - of no particular "balance" (who cares if one side is more likely "up")
- allow each side to display "any" number (e.g. 0-99)
- permit the Owner to set the die to a range (e.g. 1-31)
[or even multi-die ranges (e.g. modelling the sum of 2 seven-sided dice plus a 13-sided one)]
- use the electronics to randomize an answer within that range
- and display the answer on all sides

With this you get The Universal Die, true (okay, Pseudo-) randomness, and the Joy (?) of throwing the die (or dice if you can afford several).

Yeah, 'haven't played those games for decades....

At that point, just use a phone app that lets you shake your phone.
 
Upvote
98 (101 / -3)

Ozy

Ars Tribunus Angusticlavius
7,449
But... why N-sided die? Once you've put the electronics inside, why not just
- use a 6-sided die - of no particular "balance" (who cares if one side is more likely "up")
- allow each side to display "any" number (e.g. 0-99)
- permit the Owner to set the die to a range (e.g. 1-31)
[or even multi-die ranges (e.g. modelling the sum of 2 seven-sided dice plus a 13-sided one)]
- use the electronics to randomize an answer within that range
- and display the answer on all sides

With this you get The Universal Die, true (okay, Pseudo-) randomness, and the Joy (?) of throwing the die (or dice if you can afford several).

Yeah, 'haven't played those games for decades....
Very clever!
 
Upvote
-1 (10 / -11)
Can you roll it 1000 times or so and post the results of the rolls? I'm curious about the balance.
It's probably crap, but the dirty little secret of TTRPGs is that all gaming dice are pretty crap. Even brands like Game Science that tout their exacting specifications in the name of balance produce randomly biased dice. The most fair dice are those made for casinos, which take into account the relative densities of both the die body itself and the putty used to fill in the pips. Gaming dice don't typically go that far.

There's a rigorous lab test here. It also seems to indicate that the float test isn't a valid one.
The upside is that their unfairness usually only manifests after many thousands of rolls, not a few hundred or even one thousand.

Conclusions
1. All dice are unfair because they cannot be precisely manufactured with uniform density and dimensional correctness. However, the number of rolls required to detect the unfairness of a casino die manufactured to a few ten thousandths of an inch tolerance is very high.
2. It is very difficult to manufacture D20s that will test fair in 3000 rolls. This includes both plastic and metal dice.
3. Though our sample was small, there appears to be some merit to measuring the diameters of D20 dice; a die with a maximum diameter difference of less than 0.12 mm has a good chance of testing fair in 3000 rolls.
4. Float testing dice may tell if the die is unbalanced but it will not tell which side or sides will roll with higher or lower probabilities.
5. The manner in which a die is rolled makes a large difference in the face probabilities. It varies with surface and the manner in which it is rolled, that is, dice cup to felt, drop from a height to a hard surface, or rolling in a dice tower.
6. Rolling D20 dice 100 times and using the obtained chi square values to judge relative fairness is statistically unjustifiable.
7. A D20 die must be highly unfair to be detected in a few hundred rolls.
8. The exact probability distributions for unfair dice are not asymptotically a chi square distribution.

BS! Everyone knows that how often dice roll the desired number is correlated to how well the dice’s guardian maintains their dice shrine.
 
Upvote
64 (64 / 0)

esque

Wise, Aged Ars Veteran
112
The article unfortunately missed to mention some of the potential use cases for these dice, beyond the mere geek factor.

Visually impaired people can roll dice and perceive the results themselves rather than have other tell them what they rolled. Different colors and flash patterns for people who still have some eyesight, the app speaking the result for others.

The creator is also working with different companies to get support for the pixel dice into online services like Roll20 or D&D Beyond. You will be able to play with your friend on the internet and still be able to roll physical dice and everybody will see the result instantly.

In the year until the Pixels ship the community will certainly have a few more creative ideas.
 
Upvote
95 (95 / 0)

jbee

Ars Centurion
225
One thing I really would like to know, especially with the price of these dice: Is the battery replaceable? After charging them regularly for 1-2 years, I can see those batteries dying. Are you then out of luck, or can you open up a die and replace the battery? I do not expect some kind of no-soldering-required replacement, but if you cannot even open up the dice because they are glued shut, that would be annoying.
 
Upvote
36 (36 / 0)

...m...

Ars Scholae Palatinae
1,137
One thing I really would like to know, especially with the price of these dice: Is the battery replaceable? After charging them regularly for 1-2 years, I can see those batteries dying. Are you then out of luck, or can you open up a die and replace the battery?
...sounds like they're inductive-charging wirelessly-programmable cores cast in resin, so no, but the loads are probably intermittent and light-enough to be very gentle on the battery life...
 
Upvote
38 (38 / 0)

Nop666

Ars Praefectus
3,862
Subscriptor++
One thing I really would like to know, especially with the price of these dice: Is the battery replaceable? After charging them regularly for 1-2 years, I can see those batteries dying. Are you then out of luck, or can you open up a die and replace the battery? I do not expect some kind of no-soldering-required replacement, but if you cannot even open up the dice because they are glued shut, that would be annoying.
I would guess at a lifetime of about 1000 charging cycles before the battery capacity would drop below a usable charge per gaming session. So probably a fair number of years, depending on how often you play.
TL;DR: I don't think it'd be a problem for most users.
Edit: Why on earth would anyone downvote this comment? I mean, go ahead & do so if that floats your boat - it's no skin off my nose, but why?
 
Upvote
37 (47 / -10)

graylshaped

Ars Legatus Legionis
67,723
Subscriptor++
One thing I really would like to know, especially with the price of these dice: Is the battery replaceable? After charging them regularly for 1-2 years, I can see those batteries dying. Are you then out of luck, or can you open up a die and replace the battery? I do not expect some kind of no-soldering-required replacement, but if you cannot even open up the dice because they are glued shut, that would be annoying.

Per the FAQ on the kickstarter, the dice are sealed resin encasing the internals and the batteries are not replaceable. The battery life on a single charge will depend on how fancy one gets with the animation. I appreciated Sam's heads up that the charge decays when not in use, and hope that is addressed.
 
Upvote
32 (32 / 0)

sheepless

Ars Tribunus Militum
2,101
But... why N-sided die? Once you've put the electronics inside, why not just
- use a 6-sided die - of no particular "balance" (who cares if one side is more likely "up")
- allow each side to display "any" number (e.g. 0-99)
- permit the Owner to set the die to a range (e.g. 1-31)
[or even multi-die ranges (e.g. modelling the sum of 2 seven-sided dice plus a 13-sided one)]
- use the electronics to randomize an answer within that range
- and display the answer on all sides

With this you get The Universal Die, true (okay, Pseudo-) randomness, and the Joy (?) of throwing the die (or dice if you can afford several).

Yeah, 'haven't played those games for decades....
I remain unconvinced about the Pixels, but at least their failure mode is "inert lumps which still function perfectly well as dumb dice".
 
Upvote
90 (90 / 0)

jbee

Ars Centurion
225
One thing I really would like to know, especially with the price of these dice: Is the battery replaceable? After charging them regularly for 1-2 years, I can see those batteries dying. Are you then out of luck, or can you open up a die and replace the battery? I do not expect some kind of no-soldering-required replacement, but if you cannot even open up the dice because they are glued shut, that would be annoying.

Per the FAQ on the kickstarter, the dice are sealed resin encasing the internals and the batteries are not replaceable. The battery life on a single charge will depend on how fancy one gets with the animation. I appreciated Sam's heads up that the charge decays when not in use, and hope that is addressed.

OK - that's a bit annoying, but pretty much what I expected. I just hope that the batteries do not suffer too much when the dice are sitting around for a couple weeks with a completely dead battery, should you ever have to skip the weekly gaming session due to another lockdown or whatever....
 
Upvote
9 (9 / 0)
Post content hidden for low score. Show…