We modeled balancing selection assuming that the relative fitness of the βA/βA homozygote was 1, the relative fitness of the βS/βS homozygote was 0, and the relative fitness of the βA/βS heterozygote was 1+s. On the basis of the 504 continental Africans from the 1000 Genomes Project, we estimated that s=0.158, which was in agreement with previous estimates.14 Next, by assuming a single initial copy and an effective population size Ne=25,542, we modeled random genetic drift plus balancing selection to estimate how many generations it would take for an equilibrium frequency of 12.0% to be reached. We found that the mutant allele was lost 74.6% of the time and, conditional on reaching equilibrium, reached a frequency of 12.0% after an average of 87 (95% confidence interval [68,124]) generations, or approximately 2,400 (95% confidence interval from 1,900 to 3,500) years. We stress that this value is not the age of the sickle mutation nor the age since the onset of balancing selection but rather the time to reach a frequency of 12.0%. To determine the fate of the mutant allele in the absence of heterozygote advantage, we repeated the simulation while assuming s=0. We found that the mutant allele was lost after an average of 12 generations (95% confidence interval [1,92]), with a median of two generations.