Monday, 3 November 2008

Just this once, might Our Genial Host be wrong?

Only a madman and a fool bets against Mike Smithson. I should know - I've been burned a couple of times, daring to vote against his advice.

My worst political gamble of the year was Crewe and Nantwich - I thought Labour would hold (this was before Tamsin Dunwoody was announced), and I lost £40 to Peter from Putney, Peter the Punter, David Herdson, and Double Carpet on the back of it. Mike won a holiday.

By-elections not being my specialty, I was skeptical that Labour would actually lose their deposit at Henley. Mike insisted that this was one of the surest things he had bet on this year, and his money was heavily where his mouth was, taking good odds on Labour not reaching 5%. They came fifth, and lost their deposit - I didn't lose money, but I learnt a lesson. Mike won enought to go on holiday again.

So, two cosmic laws we know to be true - Morus shouldn't be allowed to bet independently on by-elections (I've followed Mike's advice on Glasgow East and Glenrothes), and Mike Smithson rarely makes the wrong call on a binary question.

But the impetuousness of youth gives me hope that I might win one bet against Mike this time around - the US Presidential race turnout market. Shadsy at Ladbrokes was (I believe) offering 5/6 on above 60%, and 5/6 on below 60% when the market opened. My money has always been on this being a high-turnout year, but I am adamant that turnout in the US will not reach 60%. Mike Smithson, to my horror, is fairly confident that it will.

The reason I am not yet as fearful as I should be is that the market is based, not on the proportion of Registered Voters who turn-out to vote, but on the percentage of the Voting Age Population. If only 80% of the VAP register to vote, then turnout amongst registered voters would have to exceed 75% for the 60% of VAP threshold to be met. This might seem a perverse way of calculating, but it is well-recognised against firm, federally-agreed data (the constitutionally mandated decennial census ensure's that such mechanisms are in place), whereas the number of registered voters (who either did or didn't vote) is only held at a state level, and with less scrutiny. It is more common, and less arguable, to settle disputes using VAP than the slightly less centralised '% registered voters' measure.

Turnout last crossed the 60% mark in 1968, when Richard Nixon finally won the Presidency against Hubert Humphrey. It had reached 63% in 1960 when nixon lost to a young Jack Kennedy. Even 2004's record number (absolute number that is) of voters for each candidate didn't cross the 60% rubicon, and I remain skeptical that a race that no longer seems tight will actually energise enough *Republicans* to see record turnout yet again.

I have never seen a GOTV operation as slick as Obama's, and the enthusiasm and fundraising are testament to one of the best campaigns perhaps ever waged. We might well see a record number of absolute voters, certainly for a Democratic candidate. but a big part of GW Bush's re-election was on the back of the millions of Rovian footsoldiers from the churches and community centres of the Bible Belt. I would be surprised if they turned-out in equal or greater numbers for John McCain, or even for Sarah Palin on the bottom of the ticket.

So in spite of the adamant cosmic law, I still have faith that I might not lose this one. It's not a UK by-election, which at least means I'm playing on more familiar turf, but that was what I thought about the Democratic VP market too. If I lose this one, you can be sure that I will never again, as long as I live, bet against Mike Smithson.

Morus

31 comments:

edmundintokyo said...

I take the point that if the Republicans aren't keeping the turnout up at their end, it might be hard to set a historical record.

On the other hand, how are they calculating the Voting Age Population? Is it coming from the census? 'Cos if they're underestimating the number of voting age residents, it might be easier to reach a higher proportion of the number they come up with than Nixon/Kennedy.

(Thinking of this analysis here:
http://www.theonion.com/content/node/28578
)

Dan said...

I think it's a fair bet Morus - the other one that I think Mike is calling wrong is the Obama landslide. The default position of US elections is of two big blocks and a few swing states in between almost regardless of campaigns and candidates - so I think Obama will win, but it will be in the 270-350 range, so there's a lot of value on the Betfair electoral college votes market.

Morus said...

Dan, I wouldn't swear to this, but this is how I think they work out VAP in election years.

Although the Constitutionally-mandated Census is every ten years (and sampling is illegal), the Federal Gvts OMB and the Census Bureau update it every year based on immigration stats, extrapolation from previous surveys and updates from states and municipalities. The Decennial figure has to be done manually at a cost of $billions, but they have an arguably more accurate statistical answer each year.

Based on this year's figure (they may do this more than once a year) they would calculate total votes over VAP (according to that year's sample figure).

Other thing to mention is counting all votes may take some time (if only posted, not arrived, by Election Day they are still valid, esp Military votes). If this market is betwee 59-60% before final result is announced, it may take a few weeks to resolve as Sec of States in each State delay final number of votes case info.

Peter the Punter said...

Tut, tut, Morus, will you never learn?

If you want an even tenner, you can have it with me. You can pay me tonite. Should soon be obvious this is going to be an epic turnout.

shadsy said...

Good Luck Morus!
We took that market down a couple of weeks ago. I believe that the media's favoured measure is %age of registered voters and, as you point out, that will inevitably be a fair bit higher than the VAP number we specified. I expect you are right that our measure will be under, but the reported measure wil be over 60%. No matter how carefully we word our rules, that is a recipe for customer service woe. I think intrade use the VAP as the criteria too.

Chri A said...

Quiet here, iinit?

Benedict White said...

Morus, I on the other hand have never lost a bet to Mike Smithson and say the race is still in play! (that said I think Obama McCain is something like 5/1 on V 5/1 against)

Anonymous said...

First!

Richard Howell said...

Main Site Down?

Ave it 08 said...

Test.......

Chris A said...

Has channel 1 crashed then?

Stickers said...

Main site down....

Anonymous said...

Channel 1 takes a dive!

grrr, what's my googleblogger user/pass......

ukpaul said...

Checking in, just in case.

Sean Thomas said...

It's down!

No, I'm talking about the £. Just dipped to an all-time low against the euro.

Mike L said...

Yes, Main site down.

Ted said...

Seems to be down - even before the main event starts

Philippe Magnan said...

This format sucks!

James Burdett said...

Aargh. Was bound to happen!

Sean Thomas said...

Why is the bandwidth so sucky?

edmundintokyo said...

Anyone got a URL for a live election feed over the internet showing people voting, pundits with charts pretending they know something, etc?

Chris A said...

Jeremy Vine is at least in a sober suit - maybe he's learned a lesson

tpfkar said...

How long till this one goes down then!

Baskerville said...

With talk of record turnouts, what are the chances that Mac will win more votes than Dubya and still lose?

chris said...

Wooo Hooo - the BBC coverage is great. It's a bad night for the tories.

Morus said...

NEW THREAD

Ted said...

edmundintokyo
www.cnn.com/live

Benedict White said...

Baskerville "With talk of record turnouts, what are the chances that Mac will win more votes than Dubya and still lose?"

large. I think that Tony Blair won on less votes than Kinnock lost on, possibly in 1997 certainly in 2001.

Mike L said...

Main site back up.

Anthony said...

Edmundintoyko -

CNN lives here - http://edition.cnn.com/video/live/live_asx.html?stream=stream1

edmundintokyo said...

OK, found one:
http://www.huffingtonpost.com/2008/10/30/election-results-electora_n_139361.html

CNN is giving me "not available in your geographical area", and I don't think their proprietary plugin works on the Linux.