The aim of this paper is to show an original method of calculating individual lexical richness. This method leads to a non-linear optimization. A randomized algorithm, Simulated Annealing, is used in order to carry out the optimization. This procedure has allowed us to represent a function from which we obtain a reliable pattern for lexical estimation. Furthermore, this method is compatible with other traditional procedures used for the estimation of lexical richness. Therefore, in this work we have taken an alternative and more general approach: we wish to calculate the virtual richness of the individual’s vocabulary. In order to validate the new model that calculates lexical richness, we carried out a pilot study based on the subjects’ sociolinguistic patterns that govern their lexical richness. We have explored the lexical variation of the Spanish system that occurs during the oral exchanges of 86 speakers born in Malaga.