Influence versus utility in the evaluation of voting rules: a new look at the Penrose formula