At the conclusion of the third debate, President Bush is in much better shape than the polls would suggest. This is not based on wishful thinking, outdated historical models, or a Magic Eight Ball. It is based on the clear fact that for the past two years, the major pollsters have been completely blindsided by the Republican Party’s vastly enhanced turnout efforts.

In 2000, the situation was reversed. Mr. Bush galloped into Election Day with a lead of four or five points in most polls. But a variety of factors knocked him down in the final days. First, the last-minute disclosure of a decades-old driving-under-the-influence charge caused many voters to question the honesty and reliability of this relatively new figure on the national scene. Karl Rove has lamented that as many as 4 million evangelical voters stayed home on Election Day that year.

Meanwhile, the Democratic Party’s two most powerful organizers, African-American churches and labor unions, excelled that year. Stiff white guy Al Gore got a better turnout from African-Americans in November 2000 (54.1%) than the “first black president,” Bill Clinton in 1996 (50.6%). Voter turnout from union households was 26%, which showed an improvement from 23% in 1996, saving Pennsylvania and Michigan for the vice president.

Karl Rove and other high-ranking GOP strategists realized that they came within 537 votes of losing Florida and the presidency because the Democrats were running rings around them in their get-out-the-vote efforts. So they invested enormous resources of time, energy, and money in developing methods to make sure likely GOP voters were motivated to vote and got their butts to the polls.

The major pollsters, however, took the 2000 results as the new standard and assumed that Democrats would retain their turnout advantage. And they paid the price for that assumption. Consult the results of the final polls in the key gubernatorial and senatorial races of 2002 on RealClearPolitics.com, and then look at the final results of the elections.

While pollsters thought that the Democrats would have another solid turnout effort, they completely missed the GOP turnout machine, and underestimated the performance of GOP candidates by anywhere from 4 to 13 points.

Pollster John Zogby – who is consistently hyped for “most accurately predicting how the 2000 presidential election would play out” – deserves particular ridicule for his wild overestimation of Democrats’ chances in 2002. In Minnesota, Mr. Zogby missed by 9 points. In Colorado, he was off by 10. In Georgia, he was off by 7. In Texas, he was off by 8. In all of these cases, he was underestimating the Republican.

In fact, Mr. Zogby “called” 29% of the polls wrong in 2002, while the average for everyone else was about 13% percent. Have you seen that little detail mentioned in any recent articles about Zogby’s state or nationwide polls?

The four gubernatorial races in 2003 were the same – in the three that the Republicans won, the news polls beforehand underestimated their performance by 3 to 9 points. In the case of Governor Schwarzenegger’s victory in the recall, the Los Angeles Times put Mr. Schwarzenegger’s lead at 8 points. Actual result? Der Governator wins by 17. Hasta la vista, baby.

Now, turnout in presidential races differs from turnout in off-year elections like 2002 and 2003. And perhaps the black community will turn out in droves for – stifling laugh – windsurfing Mr. Kerry, who has openly proclaimed he would like to be the second black president.

But there’s little doubt that the GOP turnout effort will be miles ahead of 2000, and it’s darn unlikely that pollsters have started taking that into account. So consider Mr. Kerry’s numbers in these polls inflated by at least a point or two.

The other reason for Bush team confidence is the Electoral College. Tear your head away from the updated-daily Rasmussen and Washington Post tracking polls that seem to swing a point or two each day, randomly untied to daily events. Take a good look at the map.

Presume that Mr. Kerry wins all of Mr. Gore’s states. He still has to get seven more electoral votes because of changes in population since the 2000 election. This means that winning New Hampshire (four votes) or Nevada (five) alone won’t do it.

Meanwhile, at least four Blue States look like real trouble for Mr. Kerry – Minnesota (10) and Wisconsin (another 10). If you want to win America’s Dairy land, you don’t call the state’s most important almost-religious site “Lambert Field.” Iowa (seven votes) and New Mexico (five) also are looking like possible pickups for Mr. Bush.

This list doesn’t count the disaster scenarios for Mr. Kerry if he loses Pennsylvania (21 electoral votes), where his lead has been at one or two points in the four most recent polls, Michigan (17 votes), or New Jersey (15 votes). If Mr. Kerry loses any of those, we can all go to bed early on Election Night.

Where are Mr. Kerry and Senator Edwards spending their time in these crucial final weeks?

Here is Mr. Kerry’s recent schedule: Nevada, Arizona for the debate, New Mexico, Missouri for the debate, Iowa, New Hampshire, Iowa. Mr. Edwards: Oregon, Iowa, California (to appear on the “Tonight Show”), New Jersey.

That’s an inordinate amount of time to spend in Blue States. The only non-debate-site Red States on that list are New Hampshire and Nevada, and neither one is enough to put Mr. Kerry over the top. Meanwhile, adding Mr. Edwards to the ticket has put not one Southern state into play, and the Kerry camp has abandoned Missouri.

Mr. Kerry absolutely needs to pick off one of Mr. Bush’s key big states – Florida, where three of the four most recent polls had Mr. Bush up outside the margin of error, or Ohio, where Mr. Kerry has some resurgence, although his leads are inside the margin of error. Even then, team Kerry has to successfully defend eight Blue States that look shaky. Possible, but difficult.

The election isn’t over. But Mr. Kerry faces a far more uphill climb than his fans in the press would have you believe.

Mr. Geraghty writes the Kerry Spot for National Review Online.