To talk to Democrats these days is to witness a strange mix of enormous anticipated relief and near nervous collapse. They haven’t cared about an election this much for a long time. They went through the interminable agony of 2000 and then the crushing whiplash of election day in 2004, when they thought they’d won and realised, as the night went on, that they’d lost.

To put it bluntly: they simply cannot believe the polls right now. The numbers are too good for a Democrat running for the presidency from scratch. They just don’t feel right. A clear lead for a nonincumbent black guy from Chicago against a war hero and a Barbie doll? Shurely shome mishtake. But in fact there is no obvious mishtake.

The cold reality – measured as these things have been measured in the past – is undeniable. If you put every poll into the blender, if you include all the psepho-logical data and run the model as many times as you can, the statistical likelihood of Barack Obama winning the election, if it were held today, is well over 90%.

And that’s deeply unnerving. There is, in fact, very little recent precedent for such a comfortable victory. The last Democrat who won the White House as a nonincumbent was Bill Clinton, and he ran in a strange three-way race in 1992 and won only 43% of the vote, which is roughly what John McCain is getting now. Jimmy Carter was the last one before that, and his margin of victory in 1976 was 2%, in the wake of Watergate. Lyndon Johnson inherited the office from a martyred Kennedy, and Kennedy won it by a mere sliver in 1960. So the idea that a nonincumbent Democrat could win the White House by a landslide is indeed strange. It hasn’t happened in my lifetime. In fact, it hasn’t happened since Franklin D Roosevelt.

So are we dreaming? Experience suggests that the race will indeed tighten. The Bradley effect – named after a black candidate for California governor who was ahead in the polls but lost – implies that many voters simply lie to pollsters when it comes to black candidates because they don’t want to appear racist (even if they’re not). There’s a great deal of academic debate about this, but it would not surprise me at all if older white voters in Ohio, West Virginia, Missouri and western Pennsylvania lurched towards the old white guy at the very end of the race.

The mood among some of the attendees at Sarah Palin rallies is indeed reminiscent of the angry crowds that once followed the segregationist George Wallace. And Palin’s own attempts to label Obama as an alien, a friend of terrorists and a traitor to his own country have been as ugly as they have been crude. “It sure would be nice if, just once, Barack Obama said he wants America to win,” she declared last week. Just because it’s disgusting doesn’t mean it can’t work.

At the same time, it seems to me possible that the gap between Obama and McCain could grow some more. Here’s why. The fury at the economic crisis is deep and wide, but there is nothing the Republicans can do to stop the tsunami of discontent. Older Americans who lean Republican have seen their retirement savings go up in smoke in a month. Although both parties were involved in the deregulation of banking, the Republicans have been in charge for much of recent history. It feels like the Tories in 1997 all over again.

Obama has almost twice the financial resources of McCain, because of his astonishing money machine. By every reasonable measure, he won all three head-to-head debates. McCain’s running mate is a national joke who cannot even hold a press conference, let alone run a country in a pinch. And the polls tend to miss a lot of voters who have swapped land lines for mobile phones. These people – disproportionately young and often minority – skew towards Obama. None of this is good news for McCain.

Two further, critical concerns lie behind the possibility of a bigger win. The first is race again. Americans are so used to fearing racism that they may have missed the positive power of a nonwhite candidate. Obama is doing very well with Latinos, and new black-voter registrations have soared. Take Georgia, a state McCain will almost certainly win easily. The number of eligible first-time voters has gone up by 9% since 2004. But within that number there’s been a 27% increase in new voter registration for African-Americans, and a 13.7% decrease for whites. Now replay that in much more promising states for Obama, such as Virginia and North Carolina, and you see why he’s favoured to win in both. Race may depress the white vote for Obama in some areas. But it may significantly increase his vote in others.

But the principal reason to bet on Obama is organisation. His primary campaign was brilliant, not just because his message was so powerful: it was because he worked out how to make the machinery of politics work for him. He mastered the art of the caucus; he used the internet as a way to find voters and then got them to the primaries and caucuses. He beat the Clintons largely because he out-hustled them on the ground.

Remember 2004? The reason Karl Rove and George Bush won, despite losing in many polls, was because they found and organised a new bloc of white, evangelical voters and brought them to the ballot box. This time, it’s Obama who has found the new voters under the radar: the young, the disengaged, various minorities and a large phalanx of middle-class white Americans. Of all the unprecedented money he has raised, he has poured an equally unprecedented amount into his ground game. Staff and volunteers are right now working phones relentlessly, walking streets, knocking on doors, constructing peer-to-peer networks and focusing almost mani-cally on turning out their voters on election day and before. Obama has even managed to insert ads for early voting into video games.

Voting has already begun in many states, by postal vote and old-fashioned ballot boxes. The candidate who has the superior organisation will add at least a couple of points to his eventual margin. I know of no objective observer who doesn’t believe that Obama’s ground game is much, much better than McCain’s right now.

Of course, the wonder of democracy is that we won’t know anything for sure until November 4. Obama weakened towards the end of his epic battle with the Clintons; he may weaken again. The smears may stick. The unease may grow. The fear of the new may eclipse the horror of the recent past.

But in the end, with two weeks to go, you can only really go by the polls. They point to somewhere between a Kennedy-style squeaker for Obama – and an FDR-style landslide. Yes, this could be 1932 again.

Times Recommends

• Care in old age: we won’t pay, the State can’t
• Viewers’ choice: our talentless Everyman
• Memo to the poor: go forth and don’t multiply