When it comes to evaluating the fairness of dice, the most common method cited by hobbyists is the saltwater float test. While this method might provide a rough indication of balance for lightweight plastic dice, it simply doesn’t work for heavy weighted metal dice like Rule Dice's Serpent’s Dagger Dice. Due to their density and weight, metal dice will typically sink regardless of any internal imbalance, rendering the saltwater test ineffective.
That’s why we took a different approach. To properly assess the fairness of our Serpent’s Dagger D20, we rolled it 1,000 times—yes, one thousand physical, manual rolls—and recorded every result, takes over 1 hour. The entire process was documented in a continuous video, which you can watch in our link below. The goal? To evaluate whether this die rolls randomly and evenly across all 20 faces.
Additionally, since we last shared our test video and data for the Serpent’s Dagger Dice D20, we’ve received a variety of feedback. Some of you specifically asked for a test of the Serpent’s Dagger D6 dice — so we listened! We rolled the D6 a total of 1,000 times, and after completing the test, we’ve now added the D6 data to the dataset as well.
Analyzing the Rolls with Bayesian Statistics
To analyze the results, we used a powerful statistical tool developed by evjrob called the Bayesian D&D Dice Analysis. This method relies on Bayesian inference, which allows us to evaluate how likely the observed distribution of dice rolls is under the assumption that the die is fair (i.e., all faces have an equal probability of appearing).
Bayesian analysis doesn’t just test whether results differ from the expected; it quantifies the uncertainty and provides a posterior distribution, giving us a probability range for how biased or unbiased a die might be. In our case, the test showed that the Serpent’s Dagger D20 & D6 can be considered fair.
D20:
D6:
Verifying with John Ellmore's Die Fairness Tester
To further validate the findings, we input the same roll data into another independent tool: John Ellmore’s Die Fairness Tester. This tool runs statistical tests including the Chi-Square goodness-of-fit test to determine how evenly results are distributed. Once again, our D20 & D6 passed the test—showing no significant bias.
D20:
D6:
Below is a statistical chart from our 1,000-roll test:
As the chart illustrates, the distribution of results across all 20 faces falls within expected variance for a fair die.
When Saltwater Fails, Roll for Real
Saltwater float tests may be a quick DIY method, but when you're dealing with precision dice, especially metal ones, real roll testing is the gold standard. If you're ever uncertain about the fairness of a die, perform a high-volume roll test and analyze the data with reliable statistical tools like evjrob's Bayesian Dice Analysis or John Ellmore's Die Fairness Tester.
Final Verdict
After 1,000 consecutive rolls and two statistical tests, we’re proud to say that the Serpent’s Dagger D20 & D6 passes the fairness test. If you also want to test the fairness of your dice, you can refer to the testing method we used. Although it will take more time, the result will make you feel at ease and the data will be more reliable.
Tools Used:
Watch the Full Test Video of D20: https://www.youtube.com/watch?v=cdrE6H_5emM
Watch the Full Test Video of D6: https://www.youtube.com/watch?v=95cGGeNYWPw