Comments on: Design for the D5

By: LS

LS — Tue, 12 Mar 2013 20:19:58 +0000

That’s disappointing. Your numbers and methodology are solid. The d5 is not random. Likely the same is true for the d7, though I imagine the rest of the Zocchi dice are okay.

I’d be interested to see the same test performed with multiple dice. I’ve heard that Game Science Dice were of higher quality in the 80s and early 90s than they are now. But getting your hands on those would probably be difficult.

By: Bill J

Bill J — Tue, 12 Mar 2013 20:13:23 +0000

After 1,000 rolls down a die tower into a catch box, here are the results. I think it is safe to say, this is a non-random outcome.

Triangle Side 1 = 273 (27.3%)
Triangle Side 5 = 269 (26.9%)
Square Side 4 = 158 (15.8%)
Square Side 3 = 155 (15.5%)
Square Side 2 = 145 (14.5%)

Asymptotic results and statistically significant at the .05 level
Mean of die rolls = 3.0050
Standard Deviation = 1.572

By: Bill J

Bill J — Tue, 12 Mar 2013 00:12:35 +0000

I will go up to 1,000. This will give me a confidence level of .01%. But the tentative results are not promising.

By: LS

LS — Mon, 11 Mar 2013 23:55:14 +0000

I’ve often suspected that this was the case. It’s unfortunate. Perhaps making the sides a little wider might help.

250 seems like kind of a small sample size. Are your tests ongoing?

By: Bill J

Bill J — Mon, 11 Mar 2013 23:49:12 +0000

Viable in the sense that when you roll it, you don’t know for certain what it will land on. However, it will produce results that are non-random, favoring the triangular sides (1) and (5). Occasionally, it will generate 2 through 4 results, but the results will be skewed towards the (1) and the (5).

I have done a study rolling the d5 through a dice tower 250 times. The results suggest it a U-shaped curve highlighting very high and very low results. Triangle 1=69. Triangle 5=71. Square 2=39. Square 3=32. Square 4=39. Using a computer program called SPSS, the results suggest that this distribution is not a random one.