I have six six-sided dice. Five of their faces are blank and identical. One face on each die contains the number 1,2,3,4,5, or 6. Suppose I roll them together, the output from this random event is an unordered set for example: {B,B,3,1,B,6} (blank side results, for our purposes, are indistinguishable.)
What is a simple way to use this result like an ordinary, fair six-sided die?
My first thought was to keep rolling until the only non-blank results are permutations of
{A,B,B,B,B,B}
{A,A,B,B,B,B}
{A,A,A,B,B,B}
{A,A,A,A,B,B}
{A,A,A,A,A,B}
{A,A,A,A,A,A}
where A is some repeated number 1-6 and B is blank.
This would work but it is not efficient. Obviously the many {B,B,B,B,B,B} results can’t be used (probably?) but, what about mixed results? For any result where one number is repeated more than the rest I could take the most repeated number as the result... but how can I interpret {1,2,1,2,4,B,6}?
Basically, I’m looking for a simple function that a human* could remember easily that maps the results of this random event to the numbers 1-6 with equal frequency.
*A table is another solution, but I would like fewer than 5 or 6 dead simple rules that use all of the results that are not {B,B,B,B,B,B}. If there is a way to use {B,B,B,B,B,B} that is even better.
