Abstract: The aim of the research is the development of a relevant and accurate instrument to analyse statistical characteristics of speech acts of different kinds. The object of the research is English apology speech formulas. The subject of the research is the statistical characteristics of the aforesaid formulas. The authors have analyzed English apology speech formulas and formalized an apology cluster, as a result, they have discovered the basic formulas apologies can be built on. In this article the authors are focusing on the development of accurate instruments that is so needed in modern linguistic studies. In the course of their research the authors have used both linguistic and mathematical methods. To recognize apologies, the authors apply a classifier that is based on reference vectors. The novelty of the research is caused by the fact that valid instruments to define particular speech acts in texts have not been yet created. As a result, the authors achieve 94% of the accuracy of their classification which is a good indicator. The results received can be used for statistical analysis of other speech formulas besides apologies. In the future the authors suggest to extend groups of respondents in order to increase accuracy of classification.