The food niche partitioning of omnivorous fish is commonly estimated on the basis of the alimentary tract content (ATC). However, since omnivorous fish utilise different ecological formations, data relating to ATC are very noisy, since an identified ATC comprises remains that can be determined according to the species level, determined only according to general food categories (i.e. higher taxonomic levels) as well as amounts of fragmented and digested remains that cannot be determined taxonomically. Thus, a variety of scales of precision can be applied during work on ATC. Up until now, there has been no evidence as to whether and how precision in ATC estimation can affect the results. This study aims at assessing how three different options of the same database influence the effectiveness and concurrency of indexes commonly used to describe the food niche of fish. The options include: (1) only general (higher than species level) food categories; (2) categories of different levels of generality; and (3) only detailed (species level) food categories. The study shows that the use of detailed (species level) food categories only, with the exclusion of general food categories, is the best method to recognise food niche partitioning and competitive interactions among fish. The food categories estimated in detail were cladocerans, and the possibility to use cladocerans as specific markers to find similarities in fish diets is discussed.