The Breeder Genetic Algorithm (BGA) is based on the equation for the response to selection, which involves the realized heritability, selection intensity, and the standard deviation of fitness. Estimating these parameters is crucial for predicting the mean fitness of the population. The paper discusses the challenges in estimating the variance, particularly in sexual recombination, and explores several modifications of sexual recombination methods. The first method introduced is gene pool recombination (GPR), which uses marginal distribution algorithms. The paper also examines more sophisticated methods that estimate the distribution of promising points. The analysis of uniform crossover for two loci is provided, showing that it can be thought of as Mendelian recombination for haploid organisms. The paper concludes with a discussion on the problem of estimating distributions and the application of the conditional distribution algorithm to optimization problems.The Breeder Genetic Algorithm (BGA) is based on the equation for the response to selection, which involves the realized heritability, selection intensity, and the standard deviation of fitness. Estimating these parameters is crucial for predicting the mean fitness of the population. The paper discusses the challenges in estimating the variance, particularly in sexual recombination, and explores several modifications of sexual recombination methods. The first method introduced is gene pool recombination (GPR), which uses marginal distribution algorithms. The paper also examines more sophisticated methods that estimate the distribution of promising points. The analysis of uniform crossover for two loci is provided, showing that it can be thought of as Mendelian recombination for haploid organisms. The paper concludes with a discussion on the problem of estimating distributions and the application of the conditional distribution algorithm to optimization problems.