NAF block dice fairness test

The dice of Zeus always fall luckily. Sophocles

Dice are funny things. If you play games a lot, especially wargames, you get use to how fickle they can be. In Blood Bowl, "Nuffle" the god of luck, has cost me more game than I care to remember by punishing me for daring to roll dice. Some people have mad rituals around dice and how they roll them, change them after they roll low, blow on them for luck etc. For me a dice is a random implement and so superstition has no place at my table. I do though, often wonder, if any dice are truly random, especially custom dice.


So I decided to test a set of NAF Blood Bowl dice by rolling them a thousand times and seeing how far they deviated from what would be expected by randomness.


A simple test but one that should highlight any potential problems with the fairness of the dice.


I decided to set the measure of statistical significance
at 5% (p-value >= 0.05). That would mean that the actual dice outcomes would have to deviate by 5% or more from the expected outcomes for the test to indicate a potential problem. I figured 5% over 1000 rolls should be safe enough. 


Test results

Outcomes	Total	Expected	% diff
Pow	171	166.67	2.60
Pow!	160	166.67	-4.00
Skull	162	166.67	-2.80
Pow/Skull	163	166.67	-2.20
Push	344	333.33	3.20
Check	1000	1000	0.00

Conclusions

The dice all stayed within the 5% threshhold and so from this test, I am pretty sure, the NAF Blood Bowl block dice are safe to use.

Rolling more dice would provide a greater degree of certanty, so the results would be safer after 10,000 or even 100,000 rolls but even at 1,000 the results seem pretty sound to me.