The heritability of specific phenotypical traits relevant for physical performance has been extensively investigated and discussed by experts from various research fields. By deciphering the complete human DNA sequence, the human genome project has provided impressive insights into the genomic landscape. The hope that this information would reveal the origin of phenotypical traits relevant for physical performance or disease risks has proven overly optimistic, and it is still premature to refer to a 'post-genomic' era of biological science. Linking genomic regions with functions, phenotypical traits and variation in disease risk is now a major experimental bottleneck. The recent deluge of genome-wide association studies (GWAS) generates extensive lists of sequence variants and genes potentially linked to phenotypical traits, but functional insight is at best sparse. The focus of this review is on the complex mechanisms that modulate gene expression. A large fraction of these mechanisms is integrated into the field of epigenetics, mainly DNA methylation and histone modifications, which lead to persistent effects on the availability of DNA for transcription. With the exceptions of genomic imprinting and very rare cases of epigenetic inheritance, epigenetic modifications are not inherited transgenerationally. Along with their susceptibility to external influences, epigenetic patterns are highly specific to the individual and may represent pivotal control centers predisposing towards higher or lower physical performance capacities. In that context, we specifically review how epigenetics combined with classical genetics could broaden our knowledge of genotype-phenotype interactions. We discuss some of the shortcomings of GWAS and explain how epigenetic influences can mask the outcome of quantitative genetic studies. We consider epigenetic influences, such as genomic imprinting and epigenetic inheritance, as well as the life-long variability of epigenetic modification patterns and their potential impact on phenotype with special emphasis on traits related to physical performance. We suggest that epigenetic effects may also play a considerable role in the determination of athletic potential and these effects will need to be studied using more sophisticated quantitative genetic models. In the future, epigenetic status and its potential influence on athletic performance will have to be considered, explored and validated using well controlled model systems before we can begin to extrapolate new findings to complex and heterogeneous human populations. A combination of the fields of genomics, epigenomics and transcriptomics along with improved bioinformatics tools and precise phenotyping, as well as a precise classification of the test populations is required for future research to better understand the inter-relations of exercise physiology, performance traits and also susceptibility towards diseases. Only this combined input can provide the overall outlook necessary to decode the molecular foundation of physical performance.