Ten months ago, I gave readers of this column an early preview of the race for the Democratic presidential nomination.

Keep an eye on the new faces, I sagely advised: Sens. Kamala Harris of California, Cory Booker of New Jersey, and Sherrod Brown of Ohio, plus former Rep. Beto O’Rourke of Texas.

Sorry about that. Despite a fawning cover story in Vanity Fair, O’Rourke flamed out fast. Harris staged an impressive launch, but then fell to earth. Brown never entered the race. Only Booker is still running, and his campaign is on life support.

Next time I recommend a hot technology stock or a soon-to-be-famous restaurant, ignore the tip.

At the end of each year, I look back at my columns to see what I got wrong —and what, if anything, I got right.

It’s a humbling exercise but a useful one. Unlike politicians, journalists are duty-bound to admit their errors in public —not just the small ones, like misspelled names, but big ones, too. You can’t learn from mistakes if you don’t acknowledge them.

And there’s one easy way to get things wrong in a presidential campaign: by trying to forecast the outcome.

It’s almost impossible to resist the temptation. Answering “beats me” never impresses anyone, even though it’s often the most accurate response.

We surround our forecasts with caveats —”the race is wide open,” “early polls don’t mean much” —and give the customers what they want: not predictions, exactly, but our best guesses as to where the campaign is going.

But primary campaigns are notoriously unpredictable. At this point in the 2008 campaign, Rudolph W. Giuliani led polls for the Republican nomination. Four years later, Newt Gingrich led the GOP polls. Neither won.

In early 2019, to my conventional-wisdom-addicted brain, Joe Biden appeared to be a commanding candidate if he survived controversies over his past positions on school desegregation, crime laws and the invasion of Iraq.

But Biden’s real problem wasn’t his historical baggage; it was his tendency to sound like a 20th century candidate lost in a 21st century debate.

I recently wrote that Biden still tops national polls because so many Democrats think he’s the most electable of their candidates.

But this time I hedged, noting that the perception could change overnight if Biden is shellacked in Iowa and New Hampshire, the first caucus and primary states.

I didn’t see Pete Buttigieg coming. The 37-year-old gay mayor of a small city? Inconceivable, I thought. Iowa voters may shortly prove me wrong.

I did see Elizabeth Warren coming. Her focus on plans to make the economy work better for the middle class was effective, I wrote.

Then Warren stumbled on healthcare. When she belatedly offered a plan, it proposed a government-run health insurance system, but only after a long transition period.

That seemed smart, I wrote. It’s not clear that voters agree.

To be fair, I did get some things right.

I figured out that the controversies over Biden’s verbal gaffes were really a polite proxy for questions about his age. He’ll be 78 on Inauguration Day; is he up to the job?

I noted that most Democratic voters aren’t Bernie Sanders-style socialists, and that the progressive “litmus tests” that dominated early months of the campaign —”Medicare for all,” the Green New Deal, and abolishing the Immigration and Customs Enforcement agency —weren’t a sure path to winning primaries.

And I got some things right on the other big political story of the year: the impeachment of President Donald Trump.

I wrote last February that House Speaker Nancy Pelosi (D-Calif.) and her Democrats were heading in that direction.

“Call this phase ‘pre-impeachment,’” I wrote. “Pelosi and her committee chairs, all Democrats, are doing what they need to do to make impeaching Trump possible.”

They were mostly just waiting for an impeachable offense to come along. Once it did, in the form of Trump’s self-incriminating telephone call with the president of Ukraine, readers should not have been surprised that the Democrats moved as quickly as they did.

It’s easy to forecast the likely verdict of an impeachment trial in the Senate. With 53 Republicans on the jury, Trump is in little danger of conviction.

The tougher question is what impact impeachment will have on the presidential election. Trump says he thinks it will help him. I wrote that it could help the Democrats instead.

David Axelrod, President Barack Obama’s former adviser, thinks it won’t have much impact at all —that by November, voters will have largely forgotten a trial that happened in January. He’s probably right.

Campaign reporting is at its best when we investigate the candidates, their histories and their proposals, and when we talk to voters to learn what they think. We’re at our worst when we make forecasts without warning how fallible we are.

In 2020, I hope you’ll return to this column for the reporting, the analysis and the occasional insight. Just don’t ask me who’s going to win.