At their best, polls are magical.

When they reveal the future before it occurs, they’re as compelling as crystal balls. And when they’re wrong, they consume careers, reputations, and fortunes.

Polls are like bets—their risk is part of their attraction. They tantalize us with their potential to push the present into the past like time machines. If we get just them right.

Polling would be so simple if voters knew what they wanted and never changed their minds. But that’s not the human condition. Instead, our opinions evolve. Determining what people think right now is the science of polling.

To me, anticipating and measuring how public opinion will change as its evolving is the art of polling. That’s what makes the wording of the questions so critical. The more insightful the questions, the better the poll should be.

Take political polls, where a candidate’s goal is always to win. A poll that shows who’s ahead in a race is helpful, but limited.

-Andy Hiller is a Guest Political Blogger for The Bernett Group. He reported on government and analyzed politics on television news in Boston for nearly forty years.