4 ms·

Chi-squared test. When I was in 7th grade, I wrote a science paper on the chi-squared test and applied it to a bunch of D&D dice. I think it is a much better wa

Chi-squared test. When I was in 7th grade, I wrote a science paper on the chi-squared test and applied it to a bunch of D&D dice. I think it is a much better way to assess dice accuracy than stacking dice in this manner.

slaymaker1907 5y ago
The issue with statistical testing is that you need to have the patience to roll die many times for useful results, at least for d20s (which are the dice where bias is both most likely and often most impactful).
- NavinF 5y ago
  That’s a job for machines: https://hackaday.com/2019/06/20/automated-dice-tester-uses-machine-vision-to-ensure-a-fair-game/ https://hackaday.com/2019/06/20/automated-dice-tester-uses-m...
shagie 5y ago
Having a chart with the numbers is there... COMPANY MAX 1-20 2-19 3-18 4-17 5-16 6-15 7-14 8-13 9-12 10-11 Chessex 0.014 0.014 0.010 0.006 0.006 0.006 0.006 0.010 0.009 0.008 0.007 The problem is that chart is kind of boring and it doesn't say anything about the manufacturing process. It says that the dice aren't fair for some reason. The image with the dice stacked shows clearly that the manufacturing process and quality control are flawed. From that it isn't a "the dice isn't fair" but rather "the way that the dice are made will make it so that every dice is more likely to not be fair" It is one thing to test the fairness of one dice or ten dice. For a D20 you'd need at least 100 rolls per dice to determine if that dice is fair. But it could just be a bad dice and other ones are fair. Now the question is "how many dice do you need to roll to demonstrate that the dice maker can't make fair dice? On the other hand, if we can take 20 random dice from a maker and stack them one way and measure it and then stack it another way and measure it and show that the dice themselves aren't clearly regular polyhedra, that visual demonstration is much more striking than a table of numbers.
- TedDoesntTalk 5y ago
  Stacking only reveals physical malformations. It won't reveal weight imbalances, will it? e.g. one part of the die weighs more than the other for some reason.
  - shagie 5y ago
    Stacking reveals physical inconsistencies in the manufacturing process. Weight imbalances would be a characteristic of one dice (and possibly not another). With sufficient rolls of one dice, you could say "yes, this dice is fair" (or not). But one dice isn't a representative indication of "if you buy a dice from X, how much attention do they pay to making dice consistently?" The claim of GameScience is that all their dice are consistently made and equally fair. If you buy a dice from Chessex, it may be fairish, it may not - it depends on how it cooled.

slaymaker1907 5y ago

The issue with statistical testing is that you need to have the patience to roll die many times for useful results, at least for d20s (which are the dice where bias is both most likely and often most impactful).

NavinF 5y ago
That’s a job for machines: https://hackaday.com/2019/06/20/automated-dice-tester-uses-machine-vision-to-ensure-a-fair-game/ https://hackaday.com/2019/06/20/automated-dice-tester-uses-m...

shagie 5y ago

Having a chart with the numbers is there... COMPANY MAX 1-20 2-19 3-18 4-17 5-16 6-15 7-14 8-13 9-12 10-11 Chessex 0.014 0.014 0.010 0.006 0.006 0.006 0.006 0.010 0.009 0.008 0.007 The problem is that chart is kind of boring and it doesn't say anything about the manufacturing process. It says that the dice aren't fair for some reason. The image with the dice stacked shows clearly that the manufacturing process and quality control are flawed. From that it isn't a "the dice isn't fair" but rather "the way that the dice are made will make it so that every dice is more likely to not be fair" It is one thing to test the fairness of one dice or ten dice. For a D20 you'd need at least 100 rolls per dice to determine if that dice is fair. But it could just be a bad dice and other ones are fair. Now the question is "how many dice do you need to roll to demonstrate that the dice maker can't make fair dice? On the other hand, if we can take 20 random dice from a maker and stack them one way and measure it and then stack it another way and measure it and show that the dice themselves aren't clearly regular polyhedra, that visual demonstration is much more striking than a table of numbers.

TedDoesntTalk 5y ago
Stacking only reveals physical malformations. It won't reveal weight imbalances, will it? e.g. one part of the die weighs more than the other for some reason.
- shagie 5y ago
  Stacking reveals physical inconsistencies in the manufacturing process. Weight imbalances would be a characteristic of one dice (and possibly not another). With sufficient rolls of one dice, you could say "yes, this dice is fair" (or not). But one dice isn't a representative indication of "if you buy a dice from X, how much attention do they pay to making dice consistently?" The claim of GameScience is that all their dice are consistently made and equally fair. If you buy a dice from Chessex, it may be fairish, it may not - it depends on how it cooled.