Cold-calling America - my 12 year-old son tries polling

Who will win? My son has been conducting his own telephone polls

I haven't even made it to America yet but my head is full of its electoral maths after two days working out Donald Trump's potential routes to power on the BBC's giant election touchscreen. We have a calculator button which is great fun and highly useful for the innumerate among us. It imagines the way a candidate can get to the White House once they've lost a big swing state along the way.

What's already clear is that Clinton - on current poll projections - has many ways of getting there. Let's call it a five-lane highway. Trump's path is pretty narrow - more of a country lane - and almost automatically fails without either Florida and Ohio/Pennsylvania. It's tight.

On election day, the first state-wide polls shut at 19:00 Eastern Time, and we will wait for results then from Indiana, Kentucky, Georgia, South Carolina, Vermont and Virginia. The safest seats should come pretty quickly - based on a mixture of real ballots and exit polls. Places like Virginia - where it's closer - will take much longer. It was never actually called last time around.

Don't be surprised if Trump seems to quickly take a lead in terms of his electoral votes tally - the "deep red" states like Kentucky and South Carolina should come pretty quickly. Anything that doesn't come soon means - possibly - it wasn't as predictable as we thought. And that's where the races get exciting.

This stuff can become slightly contagious. Last night, I got home to find my 12-year-old had spent his penultimate day of half term - and my phone bill - cold calling the US to do his own telephone polls. We posted the findings on Twitter, where many of you asked to know more about his choice of states and approach.

So here's what I gleaned of his methodology over breakfast:

He and his friend Dom went to a website called 50states.com

It appears to contain a lot of American phone numbers. Data nirvana.

They borrowed Dom's mobile phone. A Nokia brick. And looked up the state dialling codes.

They thought up funny fictional people (Pete Griffin from Family Guy) and hunted down real people with the same name (in Rhode Island).

They put on (appalling - I've heard them) American accents and told those whom they called that they were students (not technically a lie in the broadest sense).

They woke people up in Oregon as they forgot the time difference.

It seems even when disturbed in their beds, Americans were extremely friendly or at least receptive to phoney phone pollsters and no-one slammed the phone down.

Questions:

They asked them their profession, ethnicity, marital status and level of education. They promised them "aminomity" - which I'm informed is akin to anonymity if you're speaking in a cod American accent.

They yielded some interesting new demographics: ie the white-female-married-lumberjack for Hillary (I guess that's what happens if you call Oregon, I guess not many pollsters do).

They seemed to focus more on safe states than swing states for reasons best known to themselves (but may have something to do with the aforementioned quest for Funny Fictional Characters in Real Life Places).

I offer up some of their data below:

So? The big question - who comes out on top? Well, truth be told, my son Milo admitted he thought his polls were skewed as he hadn't found enough, "Sandistas or powerful Republicans". Nope, I didn't entirely understand it either. But I'm not getting involved.

And Dom, I imagine, is running away as fast as he can before his Nokia brick takes another long distance battering...

If you're reading this and you're American and you answered the phone, thank you. If you live in Oregon and never got back to sleep, I apologise. You were still a tremendous help.

Emily Maitlis is presenting BBC Newsnight's coverage on the US presidential election. You can follow her on Twitter here, and watch more of her reports here.