Here is something that may help you understand a little better in terms of genetics.
Assume that the deaf gene is represented by two homozygous recessive
dd (two d's combined). Each case (1,2,3,4) represent different children. To acquire deafness,
dd must be present after crossing over.
Thus, there is a 50% of obtaining the dd loci of 4 possibilities in this particular example, I think (not sure, forgot about this) the number of times the chromatids crossover is not a set amount. You could very well end up with a different case number than expected every time there is an offspring, which is the random part. These are just the basic biology representations that assume both chromatids are different like black and white. Actual genetics gets deeper than what I've described, by factoring the type of deafness, relatedness of the individuals, as well as their state of requiring homozygosity, etc.