• Is the Betfair Exchange election winner market wrong and could it flip?

• Late polls favour Harris, and importantly those polls are where it matters

• Trump too reliant on high male turnout

• Follow our 2024 US Election live blog here

• Get the latest data including current prices and volume of bets in our daily update

• View Betfair Exchange politics markets as you have never seen before at Betfair Predicts

In just over 48 hours, polling stations will begin to close and soon after, results pour in. The marathon US election cycle has been, as ever, packed with drama and fluctuations. Having played them well with bets on Kamala Harris from odds of 84-1 down, I cashed out my outright position with a 208 unit profit two weeks ago. The plan was to have a clean book heading into the in-play bonanza. Then last night, as alluded to in my Live Blog post earlier, I simply had to go back in on Harris.

First, the overview. My default thinking over the past four years has been that it would be much harder for any Democrat to repeat the 81.3M votes which gave Joe Biden a narrow electoral college victory, simply on the grounds of being the incumbent. That's without a far from perfect record in office or embarrassing late switch of candidates.

On the other side, I haven't shifted from a nine-year belief that Trump is the single worst candidate for winning converts and the centre ground which determine elections. That MAGA is fundamentally a losing brand. That a less divisive, conventional Republican would always fare better. An election they should win, is far from certain.

Voters less fixed than previously assumed

The balancing effect of those two opinions is a race which is too close to call, and that is what the polls say. One might assume that there won't be many switchers and that the result will revolve around turnout. But in fact, polls have shown plenty of churn. Minorities shifting slightly to Trump, seniors to Harris. A bigger gender gap. Counter-intuitively, Trump's huge rural lead slipping back.

Those small shifts affect states differently. Previously in the blog, I mused that Harris may require a smaller popular vote lead to win the electoral college. If she loses votes in populous California and New York as expected, she won't lose those states and the converse may be true in some of the whiter, Northern or Mid-West states. Wider electoral results strongly imply that her 'Blue Wall' - Pennsylvania, Michigan and Wisconsin - has become more blue in recent years.

Bombshell Iowa poll offers strong clues

Win those three states and Harris is President, bar a massive upset beyond the swing states. For several days, the polling and early voting data coming from the Blue Wall has made positive reading for Harris. Then arguably the most respected pollster in the USA produced numbers that not only reinforce that trend, but raise the prospect that polls could be wrong on a massive scale in Harris's favour.

IOWA POLL by Des Moines/Selzer

🔵 Kamala Harris 47%
🔴 Donald Trump 44%

-- umichvoter 🏳️‍🌈 (@umichvoter) November 2, 2024

To that Selzer poll, I am certain it is an outlier. Trump trails by 4% with 9% either third party or undecided. I still think he'll win Iowa by around four points. But even the swing on my margin - 2% as opposed to 6% - would have a decisive effect if part of a broader trend. Moreover, there is a longstanding theory that Iowa trends correlate closely with the Blue Wall states.

In particular, Wisconsin. This has been my worry for Harris. It was the closest of the three in 2020. Its population is more rural, white and has a higher share of non-college educated voters than Pennsylvania. The groups among whom Trump fares best. The swing in 2020 was 0.8% in Iowa compared to 0.65% in Wisconsin, 0.85% in Pennsylvania and 1.5% in Michigan.

12% of Iowans are black or Latino, compared to 15% in Wisconsin. 37% are rural, compared to 28%. 28% are evangelical, compared to 22%. Wisconsin is higher educated by 1.5%. All things equal in today's politics, I would rate Iowa about eight points more Republican. So even my projection of a 4% Trump win in Iowa (eight points more in his favour than the Selzer margin), Harris wins Wisconsin by a comfortable 4%.

As laid out in my election bets of the day series, I already had PA and MI in her column. Add WI and Nebraska's 2nd district, and Harris is up to 270 before getting to the other states. What of them?

North Carolina rated best Harris chance of the rest

I remain pro-Trump in Nevada, based on Jon Ralston's normally reliable analysis of the early voting. I remain hopeful of an upset in Arizona, but not confident. In these two states, the churn works against Harris, assuming her growth among whites is balanced by decline among Latinos.

The NV early voting blog is updated!

GOP still in driver's seat w/40,000-plus ballot lead. But Clark County mail deluge began overnight and if it continues, could get the Dems back in game.

Charts, models and down-ballot looks.

Here for you, folks!https://t.co/Pl43XBMcPA

-- Jon Ralston (@RalstonReports) November 3, 2024

North Carolina and Georgia are really hard to call. I couldn't back Trump at short odds-on, but do marginally prefer him in the latter - still on the red side of purple on the basis of the gubernatorial race. But North Carolina, fueled by a massive turnout in Charlotte, could well go to Harris.

On the matter of polls, throughout the election I have been almost exclusively reporting the high grade polls, based on the Fivethirtyeight pollster rankings. Who is the best? I pick YouGov, who are ranked fourth but have earned that rating on the basis of a far higher number of polls (twice as many as any in the top-ten). They also have a great record in MRPs so here's their final numbers.

YouGov's Final MRP Model based on interviews with 57,784 registered voters between October 25-31

PRESIDENT
Michigan - 🔵 Harris 50-46%
Nevada - 🔵 Harris 50-48%
Wisconsin - 🔵 Harris 49-77%
Pennsylvania - 🔵 Harris 49-48%
North Carolina - 🔵 Harris 49-48%
Arizona - 🔴 Trump...

-- InteractivePolls (@IAPolls2022) November 3, 2024

Harris gains from most reliable voter groups

The importance of turnout cannot be overstated. The danger is less enthusiasm to turn out for Harris, the incumbent, than Trump the challenger, compared to when roles were reversed in 2020. The Nevada early vote pointed that way. But returning to the critical Blue Wall, here the news looks good for Harris.

According to Selzer, older women were particularly driving the swing in Iowa. This fits with the early vote evidence from the Blue Wall, discussed earlier here. Seniors turnout at higher rates, as do women.

The Trump campaign has centred almost singularly on younger men. He swerved the mainstream, legacy media gigs in favour of podcasts and his target audience. Wisely, in my view. This is a group who, with higher turnout, could blindside the polls in Trump's direction, or potentially compensate for his loss among other groups.

I do expect Trump to perform a little better in urban Democratic strongholds. He starts from a very low base, has definitely made progress among younger men and minorities, and turnout on the Dem side may well fall. Last week's Monmouth poll of Pennsylvania nailed it with the sub-headline "Trump's prospects hinge on low-propensity voter turnout".

Not exactly the headline one wants to read when backing an odds-on favourite. We should be able to get an early idea about male turnout to inform whether Trump's strategy pays off.

Ultimately, what the correct odds would reflect is an aggregate of electoral college permutations. Which candidate has the more realistic paths to victory. Trump has more paths, but none as plausible as Harris defending the Blue Wall. Harris is ahead in the polls for all three states, with similar early voting trends, regardless of Selzer's stunning poll and the implications. The market is wrong. She should be favourite. My prediction is Harris wins the electoral college by 286-252.

Now see the latest US predictions from Betfair predicts