Mutation rates for STRs are predictions of how likely a mutation is to happen over a length of time. They should not be considered set or predictable like the ticking of a clock. Individual mutations are unpredictable like the number of times a year one forgets the car keys. They can be plotted over time though to create a backwards dating system with a reliability range.

For the first 37 markers the Q Project uses the mutation rates from chandler_2006.

Please reference that work for confidence intervals and other details.

STR