As Election Day approached in 2006, 2008 and 2010, I was reasonably confident about the kind of an election we would have, even if I wasn’t sure about the exact outcome. Four years ago at this time, for example, we all had a pretty good idea who the next president would be.

But this cycle, I am less confident about everything, including who will win.

I can easily imagine Democrats netting a Senate seat or Republicans netting three. I can see President Barack Obama winning a second term by carrying Ohio, Virginia and Colorado, or challenger Mitt Romney winning all three states and the White House. And while I expect Republicans to hold the House quite comfortably, I can see the GOP picking up a seat or two or Democrats netting 10 seats.

With the lack of a strong partisan wave favoring one party or the other, it’s unclear how undecided voters will break. This uncertainty is compounded by the fact that too many polls from normally reliable pollsters are contradictory. 

Sure, averaging all polls (as some do) is one way to figure out where the voters really are, but I prefer an alternative approach: Look at the best polls and talk with the best pollsters to see what is likely to happen. This year, I have more questions than answers.

The ABC News/Washington Post tracking poll has consistently shown a tight national presidential race (though with Romney often ahead by 1 point), and the most recent NBC News/Wall Street Journal poll, which is now 10 days old (and ancient history), was a dead heat. Those kinds of numbers are not ideal for a sitting president, who can’t count on breaking even among the late deciders. On the other hand, polling is an inexact science, so a race this tight is too close to call.

Key state polls don’t clarify the situation. Obama seems to have an easier path to 270 electoral votes than does Romney, but tight contests in a handful of states leave enough doubt to raise questions about the electoral vote.

Is it possible for the challenger to win the popular vote and Obama to win the electoral vote? Of course.

President George W. Bush beat John Kerry by 2.4 points in 2004 but still lost Wisconsin and New Hampshire, two of this year’s swing states. In addition, Romney looks weaker than Bush was in two other swing states, Ohio and Nevada, and a third state that Bush carried very narrowly, Iowa, is uncertain. 

Bush won nationally but carried Ohio by just 2.1 points. This cycle, Obama’s early start and initial successes in the state, in part because of his auto bailout message, has improved the president’s prospects in the Buckeye State. And the growing Hispanic vote in Nevada, which went for Bush by 2.6 points, makes the state more difficult for Romney than it was for Bush eight years ago.

So an electoral split, with Obama losing the popular vote but winning the electoral vote, is possible. But obviously, since a split verdict requires just the perfect division of votes in key states, it’s always an unusual outcome.

One additional question: Could Romney win states such as Pennsylvania, Minnesota and Michigan, which seem to be tightening in late polls? I can’t rule it out, but it is still unlikely. Those three states were among the nine closest contests in 2004 (Kerry won Pennsylvania with 50.1 percent, Minnesota with 51.1 percent and Michigan with 51.2 percent), so while it wouldn’t be unusual if they ended up being tight, they are not likely to flip from one party to the other.

The House certainly will remain in GOP hands, and Democrats never positioned themselves to have an even-money chance to net 25 seats. Republicans did a good job taking districts off the table with redistricting, and divided control in Washington and national polarization have helped potentially vulnerable Republicans to hang on. 

Democratic retirements and vulnerable incumbents allowed the National Republican Congressional Committee to stay on offense even after the party won more than 60 seats two years ago. Democratic recruiting wasn’t terrible, except in Pennsylvania, but the party might well have a disappointing night in New York, where Democrats once hoped to net at least four seats but could end up with a net gain of zero. A disappointing election night for the party in Illinois could mean a disastrous election nationally.

Who will gain seats and how many? Your guess is as good as mine. I’m expecting anything between no net change and a Democratic gain of 10 seats. A net GOP gain is unlikely but not impossible.

The fight for the Senate probably depends on the top of the ticket. It isn’t that people don’t know how to split their tickets. They do. But partisan polarization is strong, and Romney’s or Obama’s strength in Virginia, Wisconsin or Nevada could tilt those races. The same holds for states as diverse as North Dakota, Massachusetts and Montana.

As many others have noted, it’s easier to see Democrats losing two seats or fewer (possibly even gaining a seat) than it is to see the GOP netting four seats or more. But if the presidential race breaks one way or the other over the final weekend, the Senate could move as well. So while Democrats are more likely than not to retain the Senate next week, I’d advise caution of anyone making predictions.

Finally, I am expecting surprises, whether a presumed safe incumbent going down to defeat or a heavily favored nominee in an open seat losing. There are some strange numbers out there, and a few of them could well be true. Plus there are races that have attracted little polling but which are ripe for an upset or two.

Stuart Rothenberg is editor of the Rothenberg Political Report.