The paper is devoted to the topic of strategic grading, which is a term that describes situations in which a judge acquires his preferred result by giving a dishonest opinion. The initial issues concern searching for the possibility of supplementing the rules of
determining the verdict of the Scottish jury in a way that prevents jurors from manipulating it and strategic grading in the context of a system of artifi cial majority used in Polish courts of criminal justice. As it is shown, the Polish system is immune to the strategic behaviour of the judges. Article puts forward a new model of grading systems which generalizes the model constructed by Balinski & Laraki, and is similar to the classical model of Moulin. The new model is based on the assumption that for any given grades
one should be able to determine how close they are. An utterly new defi nition of strategic grading is given, and the article explains why it is needed. The main goal of the article is to give a characterization of grade aggregation functions which, in the new model, are immune to the strategic grading. On the basis of the proof by Balinski and Laraki, it is shown that order functions meet this criterion. Moreover, it is proven that under some additional assumptions they are the only type of such functions.