Rather than using a binning technique and treating alleles as discrete variables, the allele distribution is considered to be continuous. This approach provides a better fit to the true situation. The variation between two comparable bands is assessed using previously determined experimental parameters. Another advantage lies in the "quality of the match." A poor match, in which there are greater displacements between the bands in the suspect sample and the stain, will be given a lower numerator value and consequently, a lower likelihood ratio. A good match, in which there are low or no displacements between the bands, will be given a higher likelihood ratio. In its simplest form, this method involves arithmetic with four values read from smoothed experimental graphs. 4 figures, 5 references, and 1 appendix (Author abstract modified)