Saturday saw the release of an Iowa poll that showed Kamala Harris up 3 points -- a shocking result in a state where she has done no campaigning, is historically unpopular, and has trended Republican for the last several election cycles.

The poll has breathed new life into Democratic hopes, as one prediction market now says that Harris will win (the others say Trump will win), and Democrats wonder if polls showing Trump doing well (including in Iowa) are wrong.

What is actually going on?

We begin with a couple of basic principles. First, a poll is a snapshot of the electorate, but once in a while it will get the electorate wrong. Typical polls report their results with a 95% confidence interval -- but it is possible, in rare cases, that they simply collected an unrepresentative sample. That happened recently to Rasmussen, which has shown Trump leading nationally for months. One night, it showed Harris leading, then went right back to a Trump lead.

Second, polls can be manipulated by changing various assumptions about the nature of the electorate. These assumptions are used to process the raw data. I doubt that was the case with this poll, so close to the election, but who knows.

Third, it really is possible that Kamala Harris voters are being under-sampled elsewhere -- that she is bringing in a host of new female voters. This has been the Democratic hope for months: that abortion will put her over the top.

Fourth, though unlikely, is that Trump's comedian, with his Puerto Rico joke, turned off enough voters to give Harris a massive surge. I doubt this happened (I had trouble remembering the whole thing, just one week ago), but maybe.

The truth is that we don't know what is going to happen. Vote-by-mail has added to the confusion, because the polls are coming out while voters in many states are voting, or have voted, already, so they may be skewed somehow.

Personally, I begin with the assumption that Republicans never win close races, except by surprise. This is because Democrats control the process of voting (Republicans treat it as a neutral exercise), so they can turn out votes where they need to.

When Republicans win, it is because they have won in places Democrats did not anticipate (see the "blue wall" in 2016), or because Democrats are demoralized (which happens, on occasion) and don't bother fighting anymore.

Next, there is also the problem of bias. Republicans believed a massive victory was coming in 2022, and it was wiped out. Why? Partly because abortion was a big issue, but also because Republicans have stopped trusting nearly ALL media.

The fact that the media do lie about major things, and often, has meant that the conservative world has turned to partisan sources for news. The Republican confidence in recent days might reflect a more airtight bubble, not reality.

Both sides seem to think their candidate has the edge. I would still rather be Trump, for a variety of reasons. I think he has campaigned like a winner and she has campaigned like a loser. But the "Hitler Heil Mary" may also be working.

We can't know. I do know that if Harris wins, Republicans will not accept the result. Nor should they: the media have been biased, the Democrat candidate won zero primary votes; and even the judicial system has been weaponized.

The best result for the country is one in which Trump wins a landslide and we can all move on, seeing as he is termed out in 2029. The worst result is one in which the outlier Iowa poll is correct and we are back at each other's throats.

Perhaps the likeliest result is the most entertaining one: Harris wins everywhere Biden won, except Michigan and Pennsylvania, where she loses because of the Middle East; and then the House decides the race on January 6, 2025.