The well known argument for realism is that the explanation of the success of science is possible only on grounds of one's acceptance of realist stance. The argument in question is the so-called inference to the best explanation or 'no miracles' argument. The aim of this work is to analyze this argument as well as the most important antirealist objections to it. We will show that although these objections carry certain weight, nonetheless they fail to present an essential obstacle for the realist position.