For everyone else, this site will soon automatically take you across to the new one.

Latest post at the new site: US Election – The Rational Realignment.

]]> http://possumcomitatus.wordpress.com/2008/09/13/last-post-at-old-address-for-rss-feed-readers/feed/ 2 possumcomitatus http://possumcomitatus.wordpress.com/2008/09/11/thunderblogs-are-gooooooooooooooooo/ http://possumcomitatus.wordpress.com/2008/09/11/thunderblogs-are-gooooooooooooooooo/#comments Thu, 11 Sep 2008 08:41:01 +0000 Possum Comitatus http://possumcomitatus.wordpress.com/?p=1284 ]]> I’ve always wanted to say that

Click through.

Any comments made will disappear into the ether – never to be seen again, so it might be wise to hold them off for a little bit.

To start with we’ll run through a chart dump of all the polling stuff fit to print:

Next up, we’ll throw this into our Pollytrack, where we also have a new EMC and Morgan Face to Face to throw into the mix with it.

That little data graphic tells us which polls make up the current Pollytrack, the weight they have based on sample size and all the other interesting bits worth having a squiz at. If you look at the Pollytrack lines of the Primary and TPP votes, there is a definite movement back to the ALP that started around the end of July.

If we look at our All Polls series with the loess regression run through it, it becomes even more pronounced.

Rudd’s “honeymoon” ended around mid April, with the ALP primary and the TPP slowly eroding from mid April through until the end of July. Yet since that bottoming out of the ALP vote, the last 5 weeks or so have clearly shown increasing ALP vote momentum with the Coalition primary vote tanking. The Coalition have effectively wiped out 3 months of vote recovery in the last 5 weeks  – with that trend, particularly in their primary vote, this is a Very Bad Thing.

I’ve banged on about this before, where public political opinion exhibits a thing called hysteresis – where the longer a party stays at a given magnitude of vote support, the harder it becomes for that party to move away from that level of support. This hysteresis was how Crean’s leadership of the ALP had an enormous negative impact on their long term vote, requiring both Latham and Rudd to lift the ALP primary vote out of the levels Crean Labor left it in. Ignore the popular narrative on Mark Latham and just follow the data over here to see what I mean on this.

The Coalition cannot afford to go backwards in voting intention, for the larger the drop in their vote share, historically at least, the longer it will take to recover.

In an ideal world, Nelson peaked at the beginning of August and should have been removed then or sometime since. But the Coalition struggles on, frozen over the actions of Peter Costello and have been throwing away their previous gains in voteshare because of the prancing around of a man that isnt even popular with the public.

We now have the data to back up what we’ve expected for a long time on this – Essential Research ran a question in their polling released yesterday that put Rudd and Costello head-to-head on which would be the Better Prime Minister. Rudd romps it home with a two to one margin of 53/27.

That is the Liberal Party savior?  The Liberal Party are letting their voteshare gains be unwound over the petulant actions of a bloke that pulls in numbers like that, numbers that when compared to anyone but Nelson, Crean and Downer are complete shockers?

Yet to make it worse, Costello is already a known quantity – he faces the same problem Hillary Clinton faced in the US where everyone already has an opinion, mostly polarised, and it sticks a great whopping vote ceiling on top of you as a result.

The longer the Libs wait to take out Nelson, the further their vote will fall, the larger will be the hysteresis that sets in, making it a longer period it will take to get back to being competitive. Crean left the ALP taking 2 terms because the ALP dithered. How long will the Libs continue to dither – especially over a bloke that is only popular in the minds of a few hack journos and delusioned party apparatchik’s ?

Elsewhere: Pollbludger and Larvatus Prodeo

First up, our usual weekly update:

On the week by week changes in the projected Electoral College votes given by the Intrade state markets, as well as the expected ECV value (where the probability of the Democrats winning each state is multiplied by the electoral votes of that state, then all 50 contests summed) we get:

Just click to expand charts

The current win probabilities by State, as well as the change over the last week in those State probabilities have seen a surge to the Dems over the 7 days:

Moving on to the 100,000 trial simulation results – the State markets have diverged sharply from the headline “Democrat as President” market. For new US readers, you might want to go and have a look at the bottom of the US Election Page to see the methodology of the simulation, as well as exactly what all these charts mean if you’re interested.

For the raw sim results we have:

This is currently showing that the State markets give the probability of a Democrat victory at 62.7%, yet the headline market is showing a probability of a “Democrat as President” of 57.3% as can be seen below.

That is quite a sharp divergence. Our sim market has been a leading indicator thus far of the headline market, but we’ve never had a divergence this large before – so it will be interesting to see if the State markets are overcooking themselves, or the headline market is running with the newstraders over the convention and Palin hype.

Our simulated Electoral College votes and Electoral Map come in as:

But let’s have a closer look at what went on in the markets and polls over the last week, by looking at 5 days worth of movements for the last 7 days.

The yellow area represents the change from the weekly measurement, to the almost daily measurement of the markets. The Palin announcement sent the Democrat Electoral College vote projections sharply upwards in the State markets. Yet if we look at the probabilities and run them with the daily Gallup tracking poll results (where the Gallup data is dated on the day it was in play in the press rather than when the poll was actually taken) we get:

The headline “Democrat as President” market didn’t much care for the Democrat Convention, didn’t believe the polling bounce yet has continued to move strongly and consistently to the Republicans throughout last week.

The Dem polling bounce has now finally deflated, yet the State markets have priced the Palin announcement as a big negative for Republican chances of victory, moving nearly instantly on the news of Palin’s win in the Republican Veepstakes. The States pulled back slightly on the 4th September, but then moved again toward the Democrats yesterday.

We can see it clearly in the behaviour of the Electoral College vote allocations as well:

The States along with their electoral votes moved towards the Democrats, but the headline Intrade probability moved sharply Republican.

We have disequilibrium – you’d think something will have to give over the next week or so since arbitrage opportunities like this aren’t sustainable.

Basically – the headline market thinks the Palin choice and the Republican Convention were great for Republican chances of victory in November, and didn’t think the Democrat convention was worth a hill of beans in terms of generating electoral support.

But the States, the actual Electoral College vote market thought exactly the opposite, that the Palin announcement and the Republican Convention were negatives and still are.

One is wrong – the question is which?

It really is spiffy – all you need is two fingers to take the test and a spare 5 minutes. Having just done it myself, and thinking about the timing issues involved in the test framework, the results will be fascinating when they’re finalised.

So hop to it – Andrew, Alison and the expansion of human knowledge need you to click this link pronto and do your bit for furthering social science research in this country. Because if you won’t, who will?

Vote Estimates and Better Premier Rating

Satisfaction Ratings

Net Satisfaction Ratings

The Newspoll taken over the 10th to 14th of August didnt publish satisfaction ratings for the new leader of the Opposition, yet the current values are actually on the chart – they’re just clustered around Carpenter’s ratings.

Net Satisfaction for Carpenter at the moment is -6%, or -6.7% once undecideds are removed.

For the WA Opposition, net satisfaction is sitting on -3%, or -3.6% once undecideds are removed.

While the polls say it’s tight, the betting markets all have Labor around $1.30 and the Coalition sitting between $2.30 and $3.00.

Since I know about the intricacies of the WA State electorate slightly less than I know about 17th century sponge cake making – WA homeboy William Bowe at Pollbludger is the one to read.

Down in the Lyne by election, Oakeshott will win on the primary vote alone and over in Mayo – one of the more peculiar seats in the country, they too are voting for a new member, and it’s hard to imagine it would be anyone else other than the Liberal candidate Jamie Briggs.

]]> http://possumcomitatus.wordpress.com/2008/09/06/wa-election/feed/ 0 possumcomitatus wavotes wabp wapremsats waopsats wanetsats http://possumcomitatus.wordpress.com/2008/09/05/mccain-gets-impalined/ http://possumcomitatus.wordpress.com/2008/09/05/mccain-gets-impalined/#comments Fri, 05 Sep 2008 04:49:29 +0000 Possum Comitatus http://possumcomitatus.wordpress.com/?p=1212 ]]> This was me in Crikey today.

With the Democrat convention and the Palin choice making for an event ridden week in US politics, the big question remaining is “Did it make a difference?” – And that’s a question we can probably answer, yet the answer is one we may not be expecting.

The Gallup Daily Tracking polls since the beginning of August tell us all we need to know about the way the “Convention Bounce” played out for the Democrats, where the period of the convention itself is shaded in the chart below.

But the problem with the bounce in polling that usually occurs with political conventions in the US is that it tends to only be a bounce, and one that quickly washes out of the system. Yet the complicating factor here is the Palin effect, where the announcement of Palin as the McCain Veep choice seems to have benefited both the Republicans and Democrats in the polling, making us question whether the current Democrat polling levels are solid because of a public reaction against Palin, or merely a post convention balloon that has yet to deflate because of all the Republican Convention hoo har.

Thankfully we have other tools at our disposal here for digging deeper into the way the politics are playing out – the markets. My esteemed colleague Richard Farmer takes the odd dig at me for being a betting market skeptic. It’s not that I’m a skeptic, but rather, if we are going to look at the markets, we need to do it properly – preferably with large amounts of number crunching and plenty of spiffy charts.

If we look at the chart above, but this time replace the McCain Gallup tracking with the current Intrade “Democrat as President” market (that can be read from the right, below), we find something interesting.

As the polls for Obama went north, the markets became cynical and went south, yet reacted strongly for the Democrats with the Palin announcement, suggesting that the Palin effect is floating the Democrat vote and preventing any post convention deflation in Democrat public support that might ordinarily have been expected.

Yet even this headline Intrade market doesn’t tell the full story, in fact, it’s not a very good metric to use at all – if we want to get to the real guts of public expectations in the market, we need to look at how things play out collectively in the individual State Intrade markets – after all, the US election is effectively the combined result of 50 separate, but interdependent electoral contests. Electoral College votes are the only game in town.

For a few months now I’ve been running a weekly 100,000 trial simulation based on the State Intrade markets that not only adjusts for the fact that US States aren’t independent events, but also adjusts for the dodgy nature of the long probability tails of political markets – you can see more about the simulation and the methodology soon here at Crikey, but for the moment, here are the results from the overnight Intrade numbers.

That last chart tells us the probability of the Democrats winning at least any number of Electoral College Votes – simply choose a number from the bottom, trace it vertically until the bar ends, then trace horizontally to the left to find the probability of the Democrats winning at least that many ECV’s

So far this year, the simulation results have been a leading indicator of the headline “Democrat as President” market, and are far more attuned and sensitive to political events as they unfold. Redoing the simulation again using the last two days worth of Intrade data, we can see clearly that the Democrat Convention dragged the markets down, but Palin is quickly becoming McCain’s nightmare.

Firstly, we need to have a squiz at the comparisons between the Democrat as President headline market probability and the simulation probability using State market data – the highlighted region on the right of the following charts shows where the data moves from its usual weekly period, to a daily period starting last Sunday:

Since Sunday, the combined State markets have moved strongly to the Democrats, lifting the simulated probability of a Democrat victory from 58.7% to 64.9%. Yet the week before, the Democrat Convention actually dragged the Democrats down. The headline “Democrat as President” market probability on the other hand, well, it’s still having a nap – but if history is to go by, that will change soon and catch up with the simulation.

If move on to the Electoral College Votes that are currently allocated to the Democrats in the State markets (remembering folks, it takes 270 to claim the Presidency), as well as the simulated Electoral College vote allocation we get:

Again, the Democrat Convention drove the markets down, but the Palin announcement has lifted the Electoral College vote numbers for the Democrats up from 293 on Sunday, to 311 today, with the simulation results all measuring a large, broad, State by State movement toward the Democrats every day since the Palin announcement. 26 States have actually moved to the Democrats since Sunday, with Missouri, Pennsylvania, Colorado, Nevada, New Mexico, Wisconsin, North Carolina, Montana and Florida all moving by more than 5%.

More importantly for the Democrats though has been the way the Palin announcement has stalled the two month Republican recovery in public opinion and market expectations, giving momentum back to Obama and returning the Republicans to the electoral position they were experiencing at the beginning of August.

It wasn’t the Democrat Convention that has become the Republican’s problem – in 5 days, the McCain campaign has been Impalined.

UPDATE – 7th September:

The last few days worth of Intrade market data have been updated with some serious movements in the markets and polls. Click through to see.

——————————————————————————–

In other news, one of the best Qld political journos, Dennis Atkins, writes a regular column in the Courier Mail called “Party Games”. Dennis has become courageous enough to turn it into a blog, and not merely a faux blog. Worth checking out.

And finally, for something completely unusual:

WIN A DOUBLE PASS TO A PREVIEW OF THE MODERN INTERNATIONAL DEAD

Win a double pass to a preview of The Modern International Dead. Simply email info@griffintheatre.com.au and quote ‘POSSUMS POLLYTICS WIN PREVIEW TICKET’ to go into the draw. Valid for previews only and subject to availability.

The Modern International Dead

by Damien Millar

They say time changes things, but you actually have to change them yourself. Andy Warhol

From the winner of the 2007 Griffin Award and stage adaptor of Emergency Sex (and Other Desperate Measures), comes an inspiring new work for the stage.

Told with humour and courage, this enthralling kaleidoscope of personal stories features a stellar cast including Belinda McClory , Ian Meadows and Colin Moody.

Against a backdrop of miraculous visions and terrible repercussions in Rwanda, secret facilities in Iraq, landmines in Cambodia and political cleansing in East Timor comes the extraordinary stories of some very unique Australians setting out to bring relief and assistance to a troubled world. But when politics impedes progress, and reality shatters aspiration, the only consistency is compromise.

So what keeps these individuals risking their lives and going back time and time again? And when duty is done, what does it actually mean to do good in the world?

Directed by Chris Mead (Imago), this is witness theatre at its most compelling. This is The Modern International Dead.

Cast Belinda McClory , Ian Meadows & Colin Moody

Director Chris Mead / Designer Genevieve Blanchett  / Lighting Designer Bernie Tan / Composer/Sound Designer Nick Wishart

Season: 5 September – 11 October

Times: Mon 6.30pm, Tues – Sat 8pm, Sat Matinee 2 pm

Venue: SBW Stables Theatre

TICKETS CAN BE PURCHASED THROUGH MCA-TIX. CALL 1300 306 776 or go to http://www.griffintheatre.com.au

So, if you like your theatre political and independent, and happen to be in Sydney – then it might be worth checking out. You may even get a free ticket out of it.

Oh, yeah – Iemma got the boot.

The problem for NSW Labor is will anyone in NSW actually care?

If we look at both the headline “Democrat as President” market, and our probability of a Democrat as President derived from our State market simulation we get:

So far, our hypothesis of the State market simulations being a leading indicator of the headline market is looking pretty good – with the simulation probability crossing the headline probability a few weeks ago after each metric shot off in conflicting directions, leading to the headline market then playing catchup.

The electoral college votes of both the simulation as well as the current State market ECV allocation have been slowly heading down as well over the last few months.

The Dems seem to have lost their mojo in the win expectations stakes since late July – though so saying, you’d rather be Obama than McCain at the moment. Something worth having a squiz at is that first chart, but with the Gallup tracking polls added to the mix:

The markets and the polls seem to be moving together as we’d expect, except for the last week. It will be interesting to see whether the market moves to follow the polls, or whether the polls behave as the market expects and that Democrat bounce is all air that quickly washes out of the system.

In other news, many apologies for not being around lately – I’m building the new site which should go live in a few weeks. It’s a lot cleaner, information is far more accessible and I even built my first spiffy toy. Many thanks for all the suggestions, quite a few will be taken on board, if not straight away then certainly over time.

As always, the rest of the Intrade data on the US Election can be found over on its dedicated page.

Today we unveil a new Pollytrack methodology to plug these numbers into.

Previously we were using just phone polls, but over the last few weeks I’ve been experimenting around with different polling aggregations and comparing them to election results to find a more accurate way to do it (and a big thanks goes out to a few folk that gave me access to unpublished polling data to help with the calibration of this thing).

Instead of just using phone polls, we are going to do the obvious and use every piece of polling data we can find to give us rolling super samples calculated weekly. Each Pollytrack calculation is made up of a rolling all pollster average weighted by sample size, that operates over a 5 week maximum window, and where each pollster has only their latest poll contributing to the super sample at any given time.

As a pollster releases a new poll, their old poll result drops out of the rolling window, and gets replaced by their new polling data.

The 5 week window works best as ACNielsen (a monthly pollster) occasionally releases polls 5 weeks apart rather than the usual 4, simply because of timing issues. Using a 5 week window allows us to always have an ACN poll in the Pollytrack mix.

The current Pollytrack mix is made up of Newspoll, ACNielsen, Morgan face-to-face and Essential Research (EMC). Morgan phone polls will also be added when they pop up, but the last phone poll from Morgan was taken over 5 weeks ago, so it drops out of our Pollytrack window.

The beauty of this is that it gives us super samples. Our current sample is 6039, but it has been as high as 7450 in early July, while the lowest it’s been all year was 4158 in early May. That gives us minimum MoEs on this baby of between 1.1 and 1.5%.

On the primary and TPP Pollytrack charts, the minimum MoEs have also been added to the chart itself, giving us a much better handle on the current state of play that includes pure sampling error uncertainty.

Currently, Pollytrack has the primary vote on 45.7/37.2 to the ALP, with the two party preferred running at 56.6/43.4 to Labor. The minimum MoE for this week’s Pollytrack is plus or minus 1.3%.

That now makes Pollytrack by far and away the most accurate tracking poll in the country bar none. Nothing else comes within a bull’s roar of this baby.

(As always, just click the charts to blow them up)

The estimated seat changes are calculated using an assumption of a uniform swing on the national pendulum.

Essential Media have also been added to our Loess regression charts, which now give us:

Currently our local regression series has the ALP leading 45.6/36.4 on the primary vote for a TPP of 57.0/43.0.

Next up, we’ll take a squiz at the Newspoll qualitative metrics.

Finally a squiz at how all the pollsters have been traveling, including the new EMC polls we’ve added to the mix. There were a few numbers out of place in this series previously for the Morgan polls – I had a bit of a brainfart, but that’s all been fixed up.

Things might get a little bit all over the shop round these parts for the next few weeks as I prepare for the site move, so I might run a day or so late on stuff.

But good psephing to all, I’ve got to scurry off and do the latest US Intrade data that should have been completed yesterday.

Elsewhere: Pollbludger and Larvatus Prodeo