My late paternal grandfather was quite a character. I could probably write volumes on him. He had a wacky set of catch phrases -- one of which was to say of someone who made a mistake, "Once stupid, always stupid."

We can't know yet if the Democratic Party is brilliant or deluded in choosing Obama. He is fairly inexperienced as a U.S. senator, but he is a spellbinding orator, too. He offers the change his supporters crave, but neither he nor they can say exactly what that means.

But I know this: For most of us who are pundits and columnists, our forecast a year or so ago that Hillary Clinton was a lock-cinch guarantee to win the Democratic nomination was a misjudgment that recalls my grandfather's maxim, too.

Just to show that sometimes writers will admit to how far off they are, let's just look at how inaccurate I was about the Democratic nomination. In January 2007, I wrote, "Like it or don't like it, but trust me: Hillary will win the nomination Barack Obama is too green behind the gills to be ready for the most-prime prime time of all, a presidential campaign the Democratic nomination? Take it to the bank. It's Hillary."

Hold on a second. The blood is still draining from my face.

I wasn't alone, thank goodness. Consider this little number from one of the world's most respected news magazines, The Economist, from October 2007. "Mrs. Clinton is not only the frontrunner. She is well on her way to becoming the prohibitive frontrunner."

But a funny thing happened on the way to the finish line. Several things, in fact.

First, the Democratic National Committee refused to allow the Florida primary vote to count until it was too late to do Clinton any good. So instead of a New Hampshire primary win followed by a potential big win in Florida, for Hillary it was instead a loss in South Carolina, followed by more losses in states mostly unsympathetic to her.

Second, a Des Moines Register poll that appeared just days before the Iowa caucus created a wild scenario that had hordes of young voters stampeding one another on the way to vote for Obama.

Those projections didn't materialize. But they did something better for Obama: The publicity about the poll persuaded many caucus participants to cast their second ballots for Obama when their first choice didn't get the required 15 percent of the vote on the first ballot.

The next day after the Des Moines poll, our polls showed a massive shift to Obama as the "second choice." The poll became self-fulfilling.

Finally, there was the mysterious movement of party leaders, one by one -- from Kennedys to onetime Clinton loyalists -- who at remarkably regular intervals declared the race to belong to Obama and gave him their support.

There's a lesson here for both Obama and McCain boosters: Predictions don't work this year. Continued...