Problems with fitting to the power-law distribution