The obvious thing to do is then conduct a physical test on the die, for example trying to spin it on an axis with the 20 face near the top of the axis. An obvi

The obvious thing to do is then conduct a physical test on the die, for example trying to spin it on an axis with the 20 face near the top of the axis. An obviously loaded die will not be mass-symmetrical and will not spin well. In general this is called an independent test for systematic bias and it's something often left out of statistical arguments.